/***********************************************************************/ /* Document : Oracle 8i,9i,10g queries, information, and tips */ /* Doc. Versie : 59 */ /* File : oracle9i10g.txt */ /* Date : 30-06-2008 */ /* Content : Just a series of handy DBA queries. */ /* Compiled by : Albert van der Sel */ /***********************************************************************/ CONTENTS: 0. Common data dictionary queries for sessions, locks, perfoRMANce etc.. 1. DATA DICTIONARY QUERIES m.b.t. files, tablespaces, logs: 2. NOTES ON PERFORMANCE: 3. Data dictonary queries m.b.t perfoRMANce: 4. IMP and EXP, 10g IMPDB and EXPDB, and SQL*Loader Examples 5. Add, Move AND Size Datafiles,logfiles, create objects etc..: 6. Install Oracle 92 on Solaris: 7. install Oracle 9i on Linux: 8. Install Oracle 9.2.0.2 on OpenVMS: 9. Install Oracle 9.2.0.1 on AIX 9. Installation Oracle 8i - 9i: 10. CONSTRAINTS: 11. DBMS_JOB and scheduled Jobs: 12. Net8,9,10 / SQLNet: 13. Datadictionary queries Rollback segments: 14. Data dictionary queries m.b.t. security, permissions: 15. INIT.ORA parameters: 16. Snapshots: 17. Triggers: 19. BACKUP RECOVERY, TROUBLESHOOTING: 20. TRACING: 21. Overig: 22. DBA% and v$ views 23 TUNING: 24 RMAN: 25. UPGRADE AND MIGRATION 26. Some info on Rdb: 27. Some info on IFS 28. Some info on 9iAS rel. 2 29 - 35 9iAS configurations and troubleshooting 30. BLOBS 31. BLOCK CORRUPTION 32. iSQL*Plus and EM 10g 33. ADDM 34. ASM and 10g RAC 35. CDC and Streams 36. X$ Tables ============================================================================================ 0. QUICK INFO/VIEWS ON SESSIONS, LOCKS, AND UNDO/ROLLBACK INFORMATION IN A SINGLE INSTANCE: =========================================================================================== SINGLE INSTANCE QUERIES: ======================== -- --------------------------- -- 0.1 QUICK VIEW ON SESSIONS: -- --------------------------- SELECT substr(username, 1, 10), osuser, sql_address, to_char(logon_time, 'DD-MM-YYYY;HH24:MI'), sid, serial#, command, substr(program, 1, 30), substr(machine, 1, 30), substr(terminal, 1, 30) FROM v$session; SELECT sql_text, rows_processed from v$sqlarea where address='' -- ------------------------ -- 0.2 QUICK VIEW ON LOCKS: (use the sys.obj$ to find ID1:) -- ------------------------ First, lets take a look at some important dictionary views with respect to locks: SQL> desc v$lock; Name Null? Type ----------------------------- -------- -------------------- ADDR RAW(8) KADDR RAW(8) SID NUMBER TYPE VARCHAR2(2) ID1 NUMBER ID2 NUMBER LMODE NUMBER REQUEST NUMBER CTIME NUMBER BLOCK NUMBER This view stores all information relating to locks in the database. The interesting columns in this view are sid (identifying the session holding or aquiring the lock), type, and the lmode/request pair. Important possible values of type are TM (DML or Table Lock), TX (Transaction), MR (Media Recovery), ST (Disk Space Transaction). Exactly one of the lmode, request pair is either 0 or 1 while the other indicates the lock mode. If lmode is not 0 or 1, then the session has aquired the lock, while it waits to aquire the lock if request is other than 0 or 1. The possible values for lmode and request are: 1: null, 2: Row Share (SS), 3: Row Exclusive (SX), 4: Share (S), 5: Share Row Exclusive (SSX) and 6: Exclusive(X) If the lock type is TM, the column id1 is the object's id and the name of the object can then be queried like so: select name from sys.obj$ where obj# = id1 A lock type of JI indicates that a materialized view is being SQL> desc v$locked_object; Name Null? Type ----------------------------- -------- -------------------- XIDUSN NUMBER XIDSLOT NUMBER XIDSQN NUMBER OBJECT_ID NUMBER SESSION_ID NUMBER ORACLE_USERNAME VARCHAR2(30) OS_USER_NAME VARCHAR2(30) PROCESS VARCHAR2(12) LOCKED_MODE NUMBER SQL> desc dba_waiters; Name Null? Type ----------------------------- -------- -------------------- WAITING_SESSION NUMBER HOLDING_SESSION NUMBER LOCK_TYPE VARCHAR2(26) MODE_HELD VARCHAR2(40) MODE_REQUESTED VARCHAR2(40) LOCK_ID1 NUMBER LOCK_ID2 NUMBER SQL> desc v$transaction; Name Null? Type ----------------------------- -------- -------------------- ADDR RAW(8) XIDUSN NUMBER XIDSLOT NUMBER XIDSQN NUMBER UBAFIL NUMBER UBABLK NUMBER UBASQN NUMBER UBAREC NUMBER STATUS VARCHAR2(16) START_TIME VARCHAR2(20) START_SCNB NUMBER START_SCNW NUMBER START_UEXT NUMBER START_UBAFIL NUMBER START_UBABLK NUMBER START_UBASQN NUMBER START_UBAREC NUMBER SES_ADDR RAW(8) FLAG NUMBER SPACE VARCHAR2(3) RECURSIVE VARCHAR2(3) NOUNDO VARCHAR2(3) PTX VARCHAR2(3) NAME VARCHAR2(256) PRV_XIDUSN NUMBER PRV_XIDSLT NUMBER PRV_XIDSQN NUMBER PTX_XIDUSN NUMBER PTX_XIDSLT NUMBER PTX_XIDSQN NUMBER DSCN-B NUMBER DSCN-W NUMBER USED_UBLK NUMBER USED_UREC NUMBER LOG_IO NUMBER PHY_IO NUMBER CR_GET NUMBER CR_CHANGE NUMBER START_DATE DATE DSCN_BASE NUMBER DSCN_WRAP NUMBER START_SCN NUMBER DEPENDENT_SCN NUMBER XID RAW(8) PRV_XID RAW(8) PTX_XID RAW(8) Queries you can use in investigating locks: =========================================== SELECT XIDUSN,OBJECT_ID,SESSION_ID,ORACLE_USERNAME,OS_USER_NAME,PROCESS from v$locked_object; SELECT d.OBJECT_ID, substr(OBJECT_NAME,1,20), l.SESSION_ID, l.ORACLE_USERNAME, l.LOCKED_MODE from v$locked_object l, dba_objects d where d.OBJECT_ID=l.OBJECT_ID; SELECT ADDR, KADDR, SID, TYPE, ID1, ID2, LMODE, BLOCK from v$lock; SELECT a.sid, a.saddr, b.ses_addr, a.username, b.xidusn, b.used_urec, b.used_ublk FROM v$session a, v$transaction b WHERE a.saddr = b.ses_addr; SELECT s.sid, l.lmode, l.block, substr(s.username, 1, 10), substr(s.schemaname, 1, 10), substr(s.osuser, 1, 10), substr(s.program, 1, 30), s.command FROM v$session s, v$lock l WHERE s.sid=l.sid; SELECT p.spid, s.sid, p.addr,s.paddr,substr(s.username, 1, 10), substr(s.schemaname, 1, 10), s.command,substr(s.osuser, 1, 10), substr(s.machine, 1, 10) FROM v$session s, v$process p WHERE s.paddr=p.addr SELECT sid, serial#, command,substr(username, 1, 10), osuser, sql_address,LOCKWAIT, to_char(logon_time, 'DD-MM-YYYY;HH24:MI'), substr(program, 1, 30) FROM v$session; SELECT sid, serial#, username, LOCKWAIT from v$session; SELECT v.SID, v.BLOCK_GETS, v.BLOCK_CHANGES, w.USERNAME, w.OSUSER, w.TERMINAL FROM v$sess_io v, V$session w WHERE v.SID=w.SID ORDER BY v.SID; SELECT * from dba_waiters; SELECT waiting_session, holding_session, lock_type, mode_held FROM dba_waiters; SELECT p.spid unix_spid, s.sid sid, p.addr, s.paddr, substr(s.username, 1, 10) username, substr(s.schemaname, 1, 10) schemaname, s.command command, substr(s.osuser, 1, 10) osuser, substr(s.machine, 1, 25) machine FROM v$session s, v$process p WHERE s.paddr=p.addr ORDER BY p.spid; Usage of v$session_longops: =========================== SQL> desc v$session_longops; SID NUMBER Session identifier SERIAL# NUMBER Session serial number OPNAME VARCHAR2(64) Brief description of the operation TARGET VARCHAR2(64) The object on which the operation is carried out TARGET_DESC VARCHAR2(32) Description of the target SOFAR NUMBER The units of work done so far TOTALWORK NUMBER The total units of work UNITS VARCHAR2(32) The units of measurement START_TIME DATE The starting time of operation LAST_UPDATE_TIME DATE Time when statistics last updated TIMESTAMP DATE Timestamp TIME_REMAINING NUMBER Estimate (in seconds) of time remaining for the operation to complete ELAPSED_SECONDS NUMBER The number of elapsed seconds from the start of operations CONTEXT NUMBER Context MESSAGE VARCHAR2(512) Statistics summary message USERNAME VARCHAR2(30) User ID of the user performing the operation SQL_ADDRESS RAW(4 | 8) Used with the value of the SQL_HASH_VALUE column to identify the SQL statement associated with the operation SQL_HASH_VALUE NUMBER Used with the value of the SQL_ADDRESS column to identify the SQL statement associated with the operation SQL_ID VARCHAR2(13) SQL identifier of the SQL statement associated with the operation QCSID NUMBER Session identifier of the parallel coordinator This view displays the status of various operations that run for longer than 6 seconds (in absolute time). These operations currently include many backup and recovery functions, statistics gathering, and query execution, and more operations are added for every Oracle release. To monitor query execution progress, you must be using the cost-based optimizer and you must: Set the TIMED_STATISTICS or SQL_TRACE parameter to true Gather statistics for your objects with the ANALYZE statement or the DBMS_STATS package You can add information to this view about application-specific long-running operations by using the DBMS_APPLICATION_INFO.SET_SESSION_LONGOPS procedure. Select 'long', to_char (l.sid), to_char (l.serial#), to_char(l.sofar), to_char(l.totalwork), to_char(l.start_time, 'DD-Mon-YYYY HH24:MI:SS' ), to_char ( l.last_update_time , 'DD-Mon-YYYY HH24:MI:SS'), to_char(l.time_remaining), to_char(l.elapsed_seconds), l.opname,l.target,l.target_desc,l.message,s.username,s.osuser,s.lockwait from v$session_longops l, v$session s where l.sid = s.sid and l.serial# = s.serial#; Select 'long', to_char (l.sid), to_char (l.serial#), to_char(l.sofar), to_char(l.totalwork), to_char(l.start_time, 'DD-Mon-YYYY HH24:MI:SS' ), to_char ( l.last_update_time , 'DD-Mon-YYYY HH24:MI:SS'), s.username,s.osuser,s.lockwait from v$session_longops l, v$session s where l.sid = s.sid and l.serial# = s.serial#; select substr(username,1,15),target,to_char(start_time, 'DD-Mon-YYYY HH24:MI:SS' ), SOFAR,substr(MESSAGE,1,70) from v$session_longops; select USERNAME, to_char(start_time, 'DD-Mon-YYYY HH24:MI:SS' ),substr(message,1,90),to_char(time_remaining) from v$session_longops; 9i and 10G note: ================ Oracle has a view inside the Oracle data buffers. The view is called v$bh, and while v$bh was originally developed for Oracle Parallel Server (OPS), the v$bh view can be used to show the number of data blocks in the data buffer for every object type in the database. The following query is especially exciting because you can now see what objects are consuming the data buffer caches. In Oracle9i, you can use this information to segregate tables to separate RAM buffers with different blocksizes. Here is a sample query that shows data buffer utilization for individual objects in the database. Note that this script uses an Oracle9i scalar sub-query, and will not work in pre-Oracle9i systems unless you comment-out column c3. column c0 heading 'Owner' format a15 column c1 heading 'Object|Name' format a30 column c2 heading 'Number|of|Buffers' format 999,999 column c3 heading 'Percentage|of Data|Buffer' format 999,999,999 select owner c0, object_name c1, count(1) c2, (count(1)/(select count(*) from v$bh)) *100 c3 from dba_objects o, v$bh bh where o.object_id = bh.objd and o.owner not in ('SYS','SYSTEM','AURORA$JIS$UTILITY$') group by owner, object_name order by count(1) desc ; -- ----------------------------- -- 0.3 QUICK VIEW ON TEMP USAGE: -- ----------------------------- select total_extents, used_extents, total_extents, current_users, tablespace_name from v$sort_segment; select username, user, sqladdr, extents, tablespace from v$sort_usage; SELECT b.tablespace, ROUND(((b.blocks*p.value)/1024/1024),2), a.sid||','||a.serial# SID_SERIAL, a.username, a.program FROM sys.v_$session a, sys.v_$sort_usage b, sys.v_$parameter p WHERE p.name = 'db_block_size' AND a.saddr = b.session_addr ORDER BY b.tablespace, b.blocks; -- -------------------------------- -- 0.4 QUICK VIEW ON UNDO/ROLLBACK: -- -------------------------------- SELECT substr(username, 1, 10), substr(terminal, 1, 10), substr(osuser, 1, 10), t.start_time, r.name, t.used_ublk "ROLLB BLKS", log_io, phy_io FROM sys.v_$transaction t, sys.v_$rollname r, sys.v_$session s WHERE t.xidusn = r.usn AND t.ses_addr = s.saddr; SELECT substr(n.name, 1, 10), s.writes, s.gets, s.waits, s.wraps, s.extents, s.status, s.optsize, s.rssize FROM V$ROLLNAME n, V$ROLLSTAT s WHERE n.usn=s.usn; SELECT substr(r.name, 1, 10) "RBS", s.sid, s.serial#, s.taddr, t.addr, substr(s.username, 1, 10) "USER", t.status, t.cr_get, t.phy_io, t.used_ublk, t.noundo, substr(s.program, 1, 15) "COMMAND" FROM sys.v_$session s, sys.v_$transaction t, sys.v_$rollname r WHERE t.addr = s.taddr AND t.xidusn = r.usn ORDER BY t.cr_get, t.phy_io; SELECT substr(segment_name, 1, 20), substr(tablespace_name, 1, 20), status, INITIAL_EXTENT, NEXT_EXTENT, MIN_EXTENTS, MAX_EXTENTS, PCT_INCREASE FROM DBA_ROLLBACK_SEGS; select 'FREE',count(*) from sys.fet$ union select 'USED',count(*) from sys.uet$; -- Quick view active transactions SELECT NAME, XACTS "ACTIVE TRANSACTIONS" FROM V$ROLLNAME, V$ROLLSTAT WHERE V$ROLLNAME.USN = V$ROLLSTAT.USN; SELECT to_char(BEGIN_TIME, 'DD-MM-YYYY;HH24:MI'), to_char(END_TIME, 'DD-MM-YYYY;HH24:MI'), UNDOTSN, UNDOBLKS, TXNCOUNT, MAXCONCURRENCY AS "MAXCON" FROM V$UNDOSTAT WHERE trunc(BEGIN_TIME)=trunc(SYSDATE); select TO_CHAR(MIN(Begin_Time),'DD-MON-YYYY HH24:MI:SS') "Begin Time", TO_CHAR(MAX(End_Time),'DD-MON-YYYY HH24:MI:SS') "End Time", SUM(Undoblks) "Total Undo Blocks Used", SUM(Txncount) "Total Num Trans Executed", MAX(Maxquerylen) "Longest Query(in secs)", MAX(Maxconcurrency) "Highest Concurrent TrCount", SUM(Ssolderrcnt), SUM(Nospaceerrcnt) from V$UNDOSTAT; SELECT used_urec FROM v$session s, v$transaction t WHERE s.audsid=sys_context('userenv', 'sessionid') and s.taddr = t.addr; (used_urec = Used Undo records) SELECT a.sid, a.username, b.xidusn, b.used_urec, b.used_ublk FROM v$session a, v$transaction b WHERE a.saddr = b.ses_addr; SELECT v.SID, v.BLOCK_GETS, v.BLOCK_CHANGES, w.USERNAME, w.OSUSER, w.TERMINAL FROM v$sess_io v, V$session w WHERE v.SID=w.SID ORDER BY v.SID; -- -------------------------------- -- 0.5 SOME EXPLANATIONS: -- -------------------------------- -- explanation of "COMMAND": 1: CREATE TABLE 2: INSERT 3: SELECT 4: CREATE CLUSTER 5: ALTER CLUSTER 6: UPDATE 7: DELETE 8: DROP CLUSTER 9: CREATE INDEX 10: DROP INDEX 11: ALTER INDEX 12: DROP TABLE 13: CREATE SEQUENCE 14: ALTER SEQUENCE 15: ALTER TABLE 16: DROP SEQUENCE 17: GRANT 18: REVOKE 19: CREATE SYNONYM 20: DROP SYNONYM 21: CREATE VIEW 22: DROP VIEW 23: VALIDATE INDEX 24: CREATE PROCEDURE 25: ALTER PROCEDURE 26: LOCK TABLE 27: NO OPERATION 28: RENAME 29: COMMENT 30: AUDIT 31: NOAUDIT 32: CREATE DATABASE LINK 33: DROP DATABASE LINK 34: CREATE DATABASE 35: ALTER DATABASE 36: CREATE ROLLBACK SEGMENT 37: ALTER ROLLBACK SEGMENT 38: DROP ROLLBACK SEGMENT 39: CREATE TABLESPACE 40: ALTER TABLESPACE 41: DROP TABLESPACE 42: ALTER SESSION 43: ALTER USE 44: COMMIT 45: ROLLBACK 46: SAVEPOINT 47: PL/SQL EXECUTE 48: SET TRANSACTION 49: ALTER SYSTEM SWITCH LOG 50: EXPLAIN 51: CREATE USER 25: CREATE ROLE 53: DROP USER 54: DROP ROLE 55: SET ROLE 56: CREATE SCHEMA 57: CREATE CONTROL FILE 58: ALTER TRACING 59: CREATE TRIGGER 60: ALTER TRIGGER 61: DROP TRIGGER 62: ANALYZE TABLE 63: ANALYZE INDEX 64: ANALYZE CLUSTER 65: CREATE PROFILE 66: DROP PROFILE 67: ALTER PROFILE 68: DROP PROCEDURE 69: DROP PROCEDURE 70: ALTER RESOURCE COST 71: CREATE SNAPSHOT LOG 72: ALTER SNAPSHOT LOG 73: DROP SNAPSHOT LOG 74: CREATE SNAPSHOT 75: ALTER SNAPSHOT 76: DROP SNAPSHOT 79: ALTER ROLE 85: TRUNCATE TABLE 86: TRUNCATE COUSTER 88: ALTER VIEW 91: CREATE FUNCTION 92: ALTER FUNCTION 93: DROP FUNCTION 94: CREATE PACKAGE 95: ALTER PACKAGE 96: DROP PACKAGE 97: CREATE PACKAGE BODY 98: ALTER PACKAGE BODY 99: DROP PACKAGE BODY -- explanation of locks: Locks: 0, 'None', /* Mon Lock equivalent */ 1, 'Null', /* N */ 2, 'Row-S (SS)', /* L */ 3, 'Row-X (SX)', /* R */ 4, 'Share', /* S */ 5, 'S/Row-X (SRX)', /* C */ 6, 'Exclusive', /* X */ to_char(b.lmode) TX: enqueu, waiting TM: DDL on object MR: Media Recovery A TX lock is acquired when a transaction initiates its first change and is held until the transaction does a COMMIT or ROLLBACK. It is used mainly as a queuing mechanism so that other sessions can wait for the transaction to complete. TM Per table locks are acquired during the execution of a transaction when referencing a table with a DML statement so that the object is not dropped or altered during the execution of the transaction, if and only if the dml_locks parameter is non-zero. LOCKS: locks op user objects, zoals tables en rows LATCH: locks op system objects, zoals shared data structures in memory en data dictionary rows LOCKS - shared of exclusive LATCH - altijd exclusive UL= user locks, geplaats door programmatuur m.b.v. bijvoorbeeld DBMS_LOCK package DML LOCKS: data manipulatie: table lock, row lock DDL LOCKS: preserves de struktuur van object (geen simulane DML, DDL statements) DML locks: row lock (TX): voor rows (insert, update, delete) row lock plus table lock: row lock, maar ook voorkomen DDL statements table lock (TM): automatisch bij insert, update, delete, ter voorkoming DDL op table table lock: S: share lock RS: row share RSX: row share exlusive RX: row exclusive X: exclusive (ANDere tansacties kunnen alleen SELECT..) in V$LOCK lmode column: 0, None 1, Null (NULL) 2, Row-S (SS) 3, Row-X (SX) 4, Share (S) 5, S/Row-X (SSX) 6, Exclusive (X) Internal Implementation of Oracle Locks (Enqueue) Oracle server uses locks to provide concurrent access to shared resources whereas it uses latches to provide exclusive and short-term access to memory structures inside the SGA. Latches also prevent more than one process to execute the same piece of code, which other process might be executing. Latch is also a simple lock, which provides serialize and only exclusive access to the memory area in SGA. Oracle doesn’t use latches to provide shared access to resources because it will increase CPU usage. Latches are used for big memory structure and allow operations required for locking the sub structures. Shared resources can be tables, transactions, redo threads, etc. Enqueue can be local or global. If it is a single instance then enqueues will be local to that instance. There are global enqueus also like ST enqueue, which is held before any space transaction can be occurred on any tablespace in RAC. ST enqueues are held only for dictionary-managed tablespaces. These oracle locks are generally known as Enqueue, because whenever there is a session request for a lock on any shared resource structure, it's lock data structure is queued to one of the linked list attached to that resource structure (Resource structure is discussed later). Before proceeding further with this topic, here is little brief about Oracle locks. Oracle locks can be applied to compound and simple objects like tables and the cache buffer. Locks can be held in different modes like shared, excusive, null, sub-shared, sub-exclusive and shared sub-exclusive. Depending on the type of object, different modes are applied. Foe example, a compound object like a table with rows, all above mentioned modes could be applicable whereas for simple objects only the first three will be applicable. These lock modes don’t have any importance of their own but the importance is how they are being used by the subsystem. These lock modes (compatibility between locks) define how the session will get a lock on that object. -- Explanation of Waits: SQL> desc v$system_event; Name ------------------------ EVENT TOTAL_WAITS TOTAL_TIMEOUTS TIME_WAITED AVERAGE_WAIT TIME_WAITED_MICRO v$system_event This view displays the count (total_waits) of all wait events since startup of the instance. If timed_statistics is set to true, the sum of the wait times for all events are also displayed in the column time_waited. The unit of time_waited is one hundreth of a second. Since 10g, an additional column (time_waited_micro) measures wait times in millionth of a second. total_waits where event='buffer busy waits' is equal the sum of count in v$waitstat. v$enqueue_stat can be used to break down waits on the enqueue wait event. While this view totals all events in an instance, v$session select event, total_waits, time_waited from v$system_event where event like '%file%' Order by total_waits desc; column c1 heading 'Event|Name' format a30 column c2 heading 'Total|Waits' format 999,999,999 column c3 heading 'Seconds|Waiting' format 999,999 column c4 heading 'Total|Timeouts' format 999,999,999 column c5 heading 'Average|Wait|(in secs)' format 99.999 ttitle 'System-wide Wait Analysis|for current wait events' select event c1, total_waits c2, time_waited / 100 c3, total_timeouts c4, average_wait /100 c5 from sys.v_$system_event where event not in ( 'dispatcher timer', 'lock element cleanup', 'Null event', 'parallel query dequeue wait', 'parallel query idle wait - Slaves', 'pipe get', 'PL/SQL lock timer', 'pmon timer', 'rdbms ipc message', 'slave wait', 'smon timer', 'SQL*Net break/reset to client', 'SQL*Net message from client', 'SQL*Net message to client', 'SQL*Net more data to client', 'virtual circuit status', 'WMON goes to sleep' ) AND event not like 'DFS%' and event not like '%done%' and event not like '%Idle%' AND event not like 'KXFX%' order by c2 desc ; Create table beg_system_event as select * from v$system_event Run workload through system or user task Create table end_system_event as select * from v$system_event Issue SQL to determine true wait events drop table beg_system_event; drop table end_system_event; SELECT b.event, (e.total_waits - b.total_waits) total_waits, (e.total_timeouts - b.total_timeouts) total_timeouts, (e.time_waited - b.time_waited) time_waited FROM beg_system_event b, end_system_event e WHERE b.event = e.event; Cumulative info, after startup: ------------------------------- SELECT * FROM v$system_event WHERE event = 'enqueue'; SELECT * FROM v$sysstat WHERE class=4; select c.name,a.addr,a.gets,a.misses,a.sleeps, a.immediate_gets,a.immediate_misses,a.wait_time, b.pid from v$latch a, v$latchholder b, v$latchname c where a.addr = b.laddr(+) and a.latch# = c.latch# order by a.latch#; -- --------------------------------------------------------------- -- 0.6. QUICK INFO ON HIT RATIO, SHARED POOL etc.. -- --------------------------------------------------------------- -- Hit ratio: SELECT (1-(pr.value/(dbg.value+cg.value)))*100 FROM v$sysstat pr, v$sysstat dbg, v$sysstat cg WHERE pr.name = 'physical reads' AND dbg.name = 'db block gets' AND cg.name = 'consistent gets'; SELECT * FROM V$SGA; -- free memory shared pool: SELECT * FROM v$sgastat WHERE name = 'free memory'; -- hit ratio shared pool: SELECT gethits,gets,gethitratio FROM v$librarycache WHERE namespace = 'SQL AREA'; SELECT SUM(PINS) "EXECUTIONS", SUM(RELOADS) "CACHE MISSES WHILE EXECUTING" FROM V$LIBRARYCACHE; SELECT sum(sharable_mem) FROM v$db_object_cache; -- finding literals in SP: SELECT substr(sql_text,1,50) "SQL", count(*) , sum(executions) "TotExecs" FROM v$sqlarea WHERE executions < 5 GROUP BY substr(sql_text,1,50) HAVING count(*) > 30 ORDER BY 2; -- --------------------------------------- -- 0.7 Quick Table and object information -- --------------------------------------- SELECT distinct substr(t.owner, 1, 25), substr(t.table_name,1,50), substr(t.tablespace_name,1,20), t.chain_cnt, t.logging, s.relative_fno FROM dba_tables t, dba_segments s WHERE t.owner not in ('SYS','SYSTEM', 'OUTLN','DBSNMP','WMSYS','ORDSYS','ORDPLUGINS','MDSYS','CTXSYS','XDB') AND t.table_name=s.segment_name AND s.segment_type='TABLE' AND s.segment_name like 'CI_PAY%'; SELECT substr(segment_name, 1, 30), segment_type, substr(owner, 1, 10), extents, initial_extent, next_extent, max_extents FROM dba_segments WHERE extents > max_extents - 100 AND owner not in ('SYS','SYSTEM'); SELECT segment_name, owner, tablespace_name, extents FROM dba_segments WHERE owner='SALES' -- you use the correct schema here and extents > 700; SELECT owner, substr(object_name, 1, 30), object_type, created, last_ddl_time, status FROM dba_objects where OWNER='RM_LIVE'; WHERE created > SYSDATE-5; SELECT owner, substr(object_name, 1, 30), object_type, created, to_char(last_ddl_time, 'DD-MM-YYYY;HH24:MI'), status FROM dba_objects where OWNER='RM_LIVE' AND last_ddl_time > SYSDATE-5; SELECT owner, substr(object_name, 1, 30), object_type, created, last_ddl_time, status FROM dba_objects WHERE status='INVALID'; Compare 2 owners: ----------------- select table_name from dba_tables where owner='MIS_OWNER' and table_name not in (SELECT table_name from dba_tables where OWNER='MARPAT'); Table and column information: ----------------------------- select substr(table_name, 1, 3) schema , table_name , column_name , substr(data_type,1 ,1) data_type from user_tab_columns where COLUMN_NAME='ENV_ID' where table_name like 'ALG%' or table_name like 'STG%' or table_name like 'ODS%' or table_name like 'DWH%' or table_name like 'MKM%' order by decode(substr(table_name, 1, 3), 'ALG', 10, 'STG', 20, 'ODS', 30, 'DWH', 40, 'MKM', 50, 60) , table_name , column_id Check on existence of JServer: ------------------------------ select count(*) from all_objects where object_name = 'DBMS_JAVA'; should return a count of 3 -- -------------------------------------- -- 0.8 QUICK INFO ON PRODUCT INFORMATION: -- -------------------------------------- ersa SELECT * FROM PRODUCT_COMPONENT_VERSION; SELECT * FROM NLS_DATABASE_PARAMETERS; SELECT * FROM NLS_SESSION_PARAMETERS; SELECT * FROM NLS_INSTANCE_PARAMETERS; SELECT * FROM V$OPTION; SELECT * FROM V$LICENSE; SELECT * FROM V$VERSION; Oracle RDBMS releases: ---------------------- 9.2.0.1 is the terminal release for Oracle 9i. Rel 2. Normally it's patched to 9.2.0.4. As from october patches 9.2.0.5 and little later 9.2.0.6 were available 9.2.0.4 is patch ID 3095277. 9.0.1.4 is the terminal release for Oracle 9i Rel. 1. 8.1.7 is the terminal release for Oracle8i. Additional patchsets exists. 8.0.6 is the terminal release for Oracle8. Additional patchsets exists. 7.3.4 is the terminal release for Oracle7. Additional patchsets exists. IS ORACLE 32BIT or 64BIT? ------------------------- Starting with version 8, Oracle began shipping 64bit versions of it's RDBMS product on UNIX platforms that support 64bit software. IMPORTANT: 64bit Oracle can only be installed on Operating Systems that are 64bit enabled. In general, if Oracle is 64bit, '64bit' will be displayed on the opening banners of Oracle executables such as 'svrmgrl', 'exp' and 'imp'. It will also be displayed in the headers of Oracle trace files. Otherwise if '64bit' is not display at these locations, it can be assumed that Oracle is 32bit. or From the OS level: % cd $ORACLE_HOME/bin % file oracle ...if 64bit, '64bit' will be indicated. To verify the wordsize of a downloaded patchset: ------------------------------------------------ The filename of the downloaded patchset usually dictates which version and wordsize of Oracle it should be applied against. For instance: p1882450_8172_SOLARIS64.zip is the 8.1.7.2 patchset for 64bit Oracle on Solaris. Also refer to the README that is included with the patch or patch set and this Note: Win2k Server Certifications: ---------------------------- OS Product Certified With Version Status Addtl. Info. Components Other Install Issue 2000 10g N/A N/A Certified Yes None None None 2000 9.2 32-bit -Opteron N/A N/A Certified Yes None None None 2000 9.2 N/A N/A Certified Yes None None None 2000 9.0.1 N/A N/A Desupported Yes None N/A N/A 2000 8.1.7 (8i) N/A N/A Desupported Yes None N/A N/A 2000 8.1.6 (8i) N/A N/A Desupported Yes None N/A N/A 2000, Beta 3 8.1.5 (8i) N/A N/A Withdrawn Yes N/A N/A N/A Solaris Server certifications: ------------------------------ Server Certifications OS Product Certified With Version Status Addtl. Info. Components Other Install Issue 9 10g 64-bit N/A N/A Certified Yes None None None 8 10g 64-bit N/A N/A Certified Yes None None None 10 10g 64-bit N/A N/A Projected None N/A N/A N/A 9 9.2 64-bit N/A N/A Certified Yes None None None 8 9.2 64-bit N/A N/A Certified Yes None None None 10 9.2 64-bit N/A N/A Projected None N/A N/A N/A 2.6 9.2 N/A N/A Certified Yes None None None 9 9.2 N/A N/A Certified Yes None None None 8 9.2 N/A N/A Certified Yes None None None 7 9.2 N/A N/A Certified Yes None None None 10 9.2 N/A N/A Projected None N/A N/A N/A 9 9.0.1 64-bit N/A N/A Desupported Yes None N/A N/A 8 9.0.1 64-bit N/A N/A Desupported Yes None N/A N/A 2.6 9.0.1 N/A N/A Desupported Yes None N/A N/A 9 9.0.1 N/A N/A Desupported Yes None N/A N/A 8 9.0.1 N/A N/A Desupported Yes None N/A N/A 7 9.0.1 N/A N/A Desupported Yes None N/A N/A 9 8.1.7 (8i) 64-bit N/A N/A Desupported Yes None N/A N/A 8 8.1.7 (8i) 64-bit N/A N/A Desupported Yes None N/A N/A 2.6 8.1.7 (8i) N/A N/A Desupported Yes None N/A N/A 9 8.1.7 (8i) N/A N/A Desupported Yes None N/A N/A 8 8.1.7 (8i) N/A N/A Desupported Yes None N/A N/A 7 8.1.7 (8i) N/A N/A Desupported Yes None N/A N/A everything below: desupported Oracle clients: --------------- Server Version Client Version 10.1.0 9.2.0 9.0.1 8.1.7 8.1.6 8.1.5 8.0.6 8.0.5 7.3.4 10.1.0 Yes Yes Was Yes #2 No No No No No 9.2.0 Yes Yes Was Yes No No Was No No #1 9.0.1 Was Was Was Was Was No Was No Was 8.1.7 Yes Yes Was Yes Was Was Was Was Was 8.1.6 No No Was Was Was Was Was Was Was 8.1.5 No No No Was Was Was Was Was Was 8.0.6 No Was Was Was Was Was Was Was Was 8.0.5 No No No Was Was Was Was Was Was 7.3.4 No Was Was Was Was Was Was Was Was -- ----------------------------------------------------- -- 0.9 QUICK INFO WITH REGARDS LOGS AND BACKUP RECOVERY: -- ----------------------------------------------------- SELECT * from V$BACKUP; SELECT file#, substr(name, 1, 30), status, checkpoint_change# -- uit controlfile FROM V$DATAFILE; SELECT d.file#, d.status, d.checkpoint_change#, b.status, b.CHANGE#, to_char(b.TIME,'DD-MM-YYYY;HH24:MI'), substr(d.name, 1, 40) FROM V$DATAFILE d, V$BACKUP b WHERE d.file#=b.file#; SELECT file#, substr(name, 1, 30), status, fuzzy, checkpoint_change# -- uit file header FROM V$DATAFILE_HEADER; SELECT first_change#, next_change#, sequence#, archived, substr(name, 1, 40), COMPLETION_TIME, FIRST_CHANGE#, FIRST_TIME FROM V$ARCHIVED_LOG WHERE COMPLETION_TIME > SYSDATE -2; SELECT recid, first_change#, sequence#, next_change# FROM V$LOG_HISTORY; SELECT resetlogs_change#, checkpoint_change#, controlfile_change#, open_resetlogs FROM V$DATABASE; SELECT * FROM V$RECOVER_FILE -- Which file needs recovery -- ---------------------------------------------------------------------------- -- 0.10 QUICK INFO WITH REGARDS TO TABLESPACES, DATAFILES, REDO LOGFILES etc..: -- ----------------------------------------------------------------------------- -- online redo log informatie: V$LOG, V$LOGFILE: SELECT l.group#, l.members, l.status, l.bytes, substr(lf.member, 1, 50) FROM V$LOG l, V$LOGFILE lf WHERE l.group#=lf.group#; SELECT THREAD#, SEQUENCE#, FIRST_CHANGE#, FIRST_TIME, to_char(FIRST_TIME, 'DD-MM-YYYY;HH24:MI') FROM V$LOG_HISTORY; -- WHERE SEQUENCE# SELECT GROUP#, ARCHIVED, STATUS FROM V$LOG; -- tablespace free-used: SELECT Total.name "Tablespace Name", Free_space, (total_space-Free_space) Used_space, total_space FROM (SELECT tablespace_name, sum(bytes/1024/1024) Free_Space FROM sys.dba_free_space GROUP BY tablespace_name ) Free, (SELECT b.name, sum(bytes/1024/1024) TOTAL_SPACE FROM sys.v_$datafile a, sys.v_$tablespace B WHERE a.ts# = b.ts# GROUP BY b.name ) Total WHERE Free.Tablespace_name = Total.name; SELECT substr(file_name, 1, 70), tablespace_name FROM dba_data_files; ---------------------------------------------- -- 0.11 AUDIT Statements: ---------------------------------------------- select v.sql_text, v.FIRST_LOAD_TIME, v.PARSING_SCHEMA_ID, v.DISK_READS, v.ROWS_PROCESSED, v.CPU_TIME, b.username from v$sqlarea v, dba_users b where v.FIRST_LOAD_TIME > '2009-03-08' and v.PARSING_SCHEMA_ID=b.user_id order by v.FIRST_LOAD_TIME ; ----------------------------------------------- -- 0.12 EXAMPLE OF DYNAMIC SQL: ----------------------------------------------- select 'UPDATE '||t.table_name||' SET '||c.column_name||'=REPLACE('||c.column_name||','''',CHR(7));' from user_tab_columns c, user_tables t where c.table_name=t.table_name and t.num_rows>0 and c.DATA_LENGTH>10 and data_type like '%CHAR%' ORDER BY t.table_name desc; create public synonym EMPLOYEE for HARRY.EMPLOYEE; select 'create public synonym '||table_name||' for CISADM.'||table_name||';' from dba_tables where owner='CISADM'; select 'GRANT SELECT, INSERT, UPDATE, DELETE ON '||table_name||' TO CISUSER;' from dba_tables where owner='CISADM'; select 'GRANT SELECT ON '||table_name||' TO CISREAD;' from dba_tables where owner='CISADM'; ======================== 1. NOTES ON PERFORMANCE: ========================= 1.1 POOLS: ========== -- SHARED POOL: -- ------------ A literal SQL statement is considered as one which uses literals in the predicate/s rather than bind variables where the value of the literal is likely to differ between various executions of the statement. Eg 1: SELECT * FROM emp WHERE ename='CLARK'; is used by the application instead of SELECT * FROM emp WHERE ename=:bind1; SQL statement for this article as it can be shared. -- Hard Parse If a new SQL statement is issued which does not exist in the shared pool then this has to be parsed fully. Eg: Oracle has to allocate memory for the statement from the shared pool, check the statement syntactically and semantically etc... This is referred to as a hard parse and is very expensive in both terms of CPU used and in the number of latch gets performed. --Soft Parse If a session issues a SQL statement which is already in the shared pool AND it can use an existing version of that statement then this is known as a 'soft parse'. As far as the application is concerned it has asked to parse the statement. if two statements are textually identical but cannot be shared then these are called 'versions' of the same statement. If Oracle matches to a statement with many versions it has to check each version in turn to see if it is truely identical to the statement currently being parsed. Hence high version counts are best avoided. The best approach to take is that all SQL should be sharable unless it is adhoc or infrequently used SQL where it is important to give CBO as much information as possible in order for it to produce a good execution plan. --Eliminating Literal SQL If you have an existing application it is unlikely that you could eliminate all literal SQL but you should be prepared to eliminate some if it is causing problems. By looking at the V$SQLAREA view it is possible to see which literal statements are good candidates for converting to use bind variables. The following query shows SQL in the SGA where there are a large number of similar statements: SELECT substr(sql_text,1,40) "SQL", count(*) , sum(executions) "TotExecs" FROM v$sqlarea WHERE executions < 5 GROUP BY substr(sql_text,1,40) HAVING count(*) > 30 ORDER BY 2; The values 40,5 and 30 are example values so this query is looking for different statements whose first 40 characters are the same which have only been executed a few times each and there are at least 30 different occurrances in the shared pool. This query uses the idea it is common for literal statements to begin "SELECT col1,col2,col3 FROM table WHERE ..." with the leading portion of each statement being the same. --Avoid Invalidations Some specific orders will change the state of cursors to INVALIDATE. These orders modify directly the context of related objects associated with cursors. That's orders are TRUNCATE, ANALYZE or DBMS_STATS.GATHER_XXX on tables or indexes, grants changes on underlying objects. The associated cursors will stay in the SQLAREA but when it will be reference next time, it should be reloaded and reparsed fully, so the global performance will be impacted. The following query could help us to better identify the concerned cursors: SELECT substr(sql_text, 1, 40) "SQL", invalidations from v$sqlarea order by invalidations DESC; -- CURSOR_SHARING parameter (8.1.6 onwards) is a new parameter introduced in Oracle8.1.6. It should be used with caution in this release. If this parameter is set to FORCE then literals will be replaced by system generated bind variables where possible. For multiple similar statements which differ only in the literals used this allows the cursors to be shared even though the application supplied SQL uses literals. The parameter can be set dynamically at the system or session level thus: ALTER SESSION SET cursor_sharing = FORCE; or ALTER SYSTEM SET cursor_sharing = FORCE; or it can be set in the init.ora file. Note: As the FORCE setting causes system generated bind variables to be used in place of literals, a different execution plan may be chosen by the cost based optimizer (CBO) as it no longer has the literal values available to it when costing the best execution plan. In Oracle9i, it is possible to set CURSOR_SHARING=SIMILAR. SIMILAR causes statements that may differ in some literals, but are otherwise identical, to share a cursor, unless the literals affect either the meaning of the statement or the degree to which the plan is optimized. This enhancement improves the usability of the parameter for situations where FORCE would normally cause a different, undesired execution plan. With CURSOR_SHARING=SIMILAR, Oracle determines which literals are "safe" for substitution with bind variables. This will result in some SQL not being shared in an attempt to provide a more efficient execution plan. -- SESSION_CACHED_CURSORS parameter is a numeric parameter which can be set at instance level or at session level using the command: ALTER SESSION SET session_cached_cursors = NNN; The value NNN determines how many 'cached' cursors there can be in your session. Whenever a statement is parsed Oracle first looks at the statements pointed to by your private session cache - if a sharable version of the statement exists it can be used. This provides a shortcut access to frequently parsed statements that uses less CPU and uses far fewer latch gets than a soft or hard parse. To get placed in the session cache the same statement has to be parsed 3 times within the same cursor - a pointer to the shared cursor is then added to your session cache. If all session cache cursors are in use then the least recently used entry is discarded. If you do not have this parameter set already then it is advisable to set it to a starting value of about 50. The statistics section of the bstat/estat report includes a value for 'session cursor cache hits' which shows if the cursor cache is giving any benefit. The size of the cursor cache can then be increased or decreased as necessary. SESSION_CACHED_CURSORS are particularly useful with Oracle Forms applications when forms are frequently opened and closed. -- SHARED_POOL_RESERVED_SIZE parameter There are quite a few notes explaining already in circulation. The parameter was introduced in Oracle 7.1.5 and provides a means of reserving a portion of the shared pool for large memory allocations. The reserved area comes out of the shared pool itself. From a practical point of view one should set SHARED_POOL_RESERVED_SIZE to about 10% of SHARED_POOL_SIZE unless either the shared pool is very large OR SHARED_POOL_RESERVED_MIN_ALLOC has been set lower than the default value: If the shared pool is very large then 10% may waste a significant amount of memory when a few Mb will suffice. If SHARED_POOL_RESERVED_MIN_ALLOC has been lowered then many space requests may be eligible to be satisfied from this portion of the shared pool and so 10% may be too little. It is easy to monitor the space usage of the reserved area using the which has a column FREE_SPACE. -- SHARED_POOL_RESERVED_MIN_ALLOC parameter In Oracle8i this parameter is hidden. SHARED_POOL_RESERVED_MIN_ALLOC should generally be left at its default value, although in certain cases values of 4100 or 4200 may help relieve some contention on a heavily loaded shared pool. -- SHARED_POOL_SIZE parameter controls the size of the shared pool itself. The size of the shared pool can impact performance. If it is too small then it is likely that sharable information will be flushed from the pool and then later need to be reloaded (rebuilt). If there is heavy use of literal SQL and the shared pool is too large then over time a lot of small chunks of memory can build up on the internal memory freelists causing the shared pool latch to be held for longer which in-turn can impact performance. In this situation a smaller shared pool may perform better than a larger one. This problem is greatly reduced in 8.0.6 and in 8.1.6 onwards due to the enhancement in . NB: The shared pool itself should never be made so large that paging or swapping occur as performance can then decrease by many orders of magnitude. -- _SQLEXEC_PROGRESSION_COST parameter (8.1.5 onwards) This is a hidden parameter which was introduced in Oracle 8.1.5. The parameter is included here as the default setting has caused some problems with SQL sharability. Setting this parameter to 0 can avoid these issues which result in multiple versions statements in the shared pool. Eg: Add the following to the init.ora file # _SQLEXEC_PROGRESSION_COST is set to ZERO to avoid SQL sharing issues # See Note:62143.1 for details _sqlexec_progression_cost=0 Note that a side effect of setting this to '0' is that the V$SESSION_LONGOPS view is not populated by long running queries. -- MTS, Shared Server and XA The multi-threaded server (MTS) adds to the load on the shared pool and can contribute to any problems as the User Global Area (UGA) resides in the shared pool. This is also true of XA sessions in Oracle7 as their UGA is located in the shared pool. (In Oracle8/8i XA sessions do NOT put their UGA in the shared pool). In Oracle8 the Large Pool can be used for MTS reducing its impact on shared pool activity - However memory allocations in the Large Pool still make use of the "shared pool latch". See for a description of the Large Pool. Using dedicated connections rather than MTS causes the UGA to be allocated out of process private memory rather than the shared pool. Private memory allocations do not use the "shared pool latch" and so a switch from MTS to dedicated connections can help reduce contention in some cases. In Oracle9i, MTS was renamed to "Shared Server". For the purposes of the shared pool, the behaviour is essentially the same. Useful SQL for looking at memory and Shared Pool problems --------------------------------------------------------- Indeling SGA: ------------- SELECT * FROM V$SGA; free memory shared pool: ------------------------ SELECT * FROM v$sgastat WHERE name = 'free memory'; hit ratio shared pool: ---------------------- SELECT gethits,gets,gethitratio FROM v$librarycache WHERE namespace = 'SQL AREA'; SELECT SUM(PINS) "EXECUTIONS", SUM(RELOADS) "CACHE MISSES WHILE EXECUTING" FROM V$LIBRARYCACHE; SELECT sum(sharable_mem) FROM v$db_object_cache; statistics: ----------- SELECT class, value, name FROM v$sysstat; Executions: ----------- SELECT substr(sql_text,1,90) "SQL", count(*) , sum(executions) "TotExecs" FROM v$sqlarea WHERE executions > 5 GROUP BY substr(sql_text,1,90) HAVING count(*) > 10 ORDER BY 2 ; The values 40,5 and 30 are example values so this query is looking for different statements whose first 40 characters are the same which have only been executed a few times each and there are at least 30 different occurrances in the shared pool. This query uses the idea it is common for literal statements to begin "SELECT col1,col2,col3 FROM table WHERE ..." with the leading portion of each statement being the same. V$SQLAREA: SQL_TEXT VARCHAR2(1000) First thousand characters of the SQL text for the current cursor SHARABLE_MEM NUMBER Amount of shared memory used by a cursor. If multiple child cursors exist, then the sum of all shared memory used by all child cursors. PERSISTENT_MEM NUMBER Fixed amount of memory used for the lifetime of an open cursor. If multiple child cursors exist, the fixed sum of memory used for the lifetime of all the child cursors. RUNTIME_MEM NUMBER Fixed amount of memory required during execution of a cursor. If multiple child cursors exist, the fixed sum of all memory required during execution of all the child cursors. SORTS NUMBER Sum of the number of sorts that were done for all the child cursors VERSION_COUNT NUMBER Number of child cursors that are present in the cache under this parent LOADED_VERSIONS NUMBER Number of child cursors that are present in the cache and have their context heap (KGL heap 6) loaded OPEN_VERSIONS NUMBER The number of child cursors that are currently open under this current parent USERS_OPENING NUMBER The number of users that have any of the child cursors open FETCHES NUMBER Number of fetches associated with the SQL statement EXECUTIONS NUMBER Total number of executions, totalled over all the child cursors USERS_EXECUTING NUMBER Total number of users executing the statement over all child cursors LOADS NUMBER The number of times the object was loaded or reloaded FIRST_LOAD_TIME VARCHAR2(19) Timestamp of the parent creation time INVALIDATIONS NUMBER Total number of invalidations over all the child cursors PARSE_CALLS NUMBER The sum of all parse calls to all the child cursors under this parent DISK_READS NUMBER The sum of the number of disk reads over all child cursors BUFFER_GETS NUMBER The sum of buffer gets over all child cursors ROWS_PROCESSED NUMBER The total number of rows processed on behalf of this SQL statement COMMAND_TYPE NUMBER The Oracle command type definition OPTIMIZER_MODE VARCHAR2(10) Mode under which the SQL statement is executed PARSING_USER_ID NUMBER The user ID of the user that has parsed the very first cursor under this parent PARSING_SCHEMA_ID NUMBER The schema ID that was used to parse this child cursor KEPT_VERSIONS NUMBER The number of child cursors that have been marked to be kept using the DBMS_SHARED_POOL package ADDRESS RAW(4) The address of the handle to the parent for this cursor HASH_VALUE NUMBER The hash value of the parent statement in the library cache MODULE VARCHAR2(64) Contains the name of the module that was executing at the time that the SQL statement was first parsed as set by calling DBMS_APPLICATION_INFO.SET_MODULE MODULE_HASH NUMBER The hash value of the module that is named in the MODULE column ACTION VARCHAR2(64) Contains the name of the action that was executing at the time that the SQL statement was first parsed as set by calling DBMS_APPLICATION_INFO.SET_ACTION ACTION_HASH NUMBER The hash value of the action that is named in the ACTION column SERIALIZABLE_ABORTS NUMBER Number of times the transaction fails to serialize, producing ORA-08177 errors, totalled over all the child cursors IS_OBSOLETE VARCHAR2(1) Indicates whether the cursor has become obsolete (Y) or not (N). This can happen if the number of child cursors is too large. CHILD_LATCH NUMBER Child latch number that is protecting the cursor V$SQL: ------ V$SQL lists statistics on shared SQL area without the GROUP BY clause and contains one row for each child of the original SQL text entered. Column Datatype Description SQL_TEXT VARCHAR2(1000) First thousand characters of the SQL text for the current cursor SHARABLE_MEM NUMBER Amount of shared memory used by this child cursor (in bytes) PERSISTENT_MEM NUMBER Fixed amount of memory used for the lifetime of this child cursor (in bytes) RUNTIME_MEM NUMBER Fixed amount of memory required during the execution of this child cursor SORTS NUMBER Number of sorts that were done for this child cursor LOADED_VERSIONS NUMBER Indicates whether the context heap is loaded (1) or not (0) OPEN_VERSIONS NUMBER Indicates whether the child cursor is locked (1) or not (0) USERS_OPENING NUMBER Number of users executing the statement FETCHES NUMBER Number of fetches associated with the SQL statement EXECUTIONS NUMBER Number of executions that took place on this object since it was brought into the library cache USERS_EXECUTING NUMBER Number of users executing the statement LOADS NUMBER Number of times the object was either loaded or reloaded FIRST_LOAD_TIME VARCHAR2(19) Timestamp of the parent creation time INVALIDATIONS NUMBER Number of times this child cursor has been invalidated PARSE_CALLS NUMBER Number of parse calls for this child cursor DISK_READS NUMBER Number of disk reads for this child cursor BUFFER_GETS NUMBER Number of buffer gets for this child cursor ROWS_PROCESSED NUMBER Total number of rows the parsed SQL statement returns COMMAND_TYPE NUMBER Oracle command type definition OPTIMIZER_MODE VARCHAR2(10) Mode under which the SQL statement is executed OPTIMIZER_COST NUMBER Cost of this query given by the optimizer PARSING_USER_ID NUMBER User ID of the user who originally built this child cursor PARSING_SCHEMA_ID NUMBER Schema ID that was used to originally build this child cursor KEPT_VERSIONS NUMBER Indicates whether this child cursor has been marked to be kept pinned in the cache using the DBMS_SHARED_POOL package ADDRESS RAW(4) Address of the handle to the parent for this cursor TYPE_CHK_HEAP RAW(4) Descriptor of the type check heap for this child cursor HASH_VALUE NUMBER Hash value of the parent statement in the library cache PLAN_HASH_VALUE NUMBER Numerical representation of the SQL plan for this cursor. Comparing one PLAN_HASH_VALUE to another easily identifies whether or not two plans are the same (rather than comparing the two plans line by line). CHILD_NUMBER NUMBER Number of this child cursor MODULE VARCHAR2(64) Contains the name of the module that was executing at the time that the SQL statement was first parsed, which is set by calling DBMS_APPLICATION_INFO.SET_MODULE MODULE_HASH NUMBER Hash value of the module listed in the MODULE column ACTION VARCHAR2(64) Contains the name of the action that was executing at the time that the SQL statement was first parsed, which is set by calling DBMS_APPLICATION_INFO.SET_ACTION ACTION_HASH NUMBER Hash value of the action listed in the ACTION column SERIALIZABLE_ABORTS NUMBER Number of times the transaction fails to serialize, producing ORA-08177 errors, per cursor OUTLINE_CATEGORY VARCHAR2(64) If an outline was applied during construction of the cursor, then this column displays the category of that outline. Otherwise the column is left blank. CPU_TIME NUMBER CPU time (in microseconds) used by this cursor for parsing/executing/fetching ELAPSED_TIME NUMBER Elapsed time (in microseconds) used by this cursor for parsing/executing/fetching OUTLINE_SID NUMBER Outline session identifier CHILD_ADDRESS RAW(4) Address of the child cursor SQLTYPE NUMBER Denotes the version of the SQL language used for this statement REMOTE VARCHAR2(1) (Y/N) Identifies whether the cursor is remote mapped or not OBJECT_STATUS VARCHAR2(19) Status of the cursor (VALID/INVALID) LITERAL_HASH_VALUE NUMBER Hash value of the literals which are replaced with system-generated bind variables and are to be matched, when CURSOR_SHARING is used. This is not the hash value for the SQL statement. If CURSOR_SHARING is not used, then the value is 0. LAST_LOAD_TIME VARCHAR2(19) IS_OBSOLETE VARCHAR2(1) Indicates whether the cursor has become obsolete (Y) or not (N). This can happen if the number of child cursors is too large. CHILD_LATCH NUMBER Child latch number that is protecting the cursor Checking for high version counts: -------------------------------- SELECT address, hash_value, version_count , users_opening , users_executing, substr(sql_text,1,40) "SQL" FROM v$sqlarea WHERE version_count > 10 ; "Versions" of a statement occur where the SQL is character for character identical but the underlying objects or binds etc.. are different. Finding statement/s which use lots of shared pool memory: -------------------------------------------------------- SELECT substr(sql_text,1,60) "Stmt", count(*), sum(sharable_mem) "Mem", sum(users_opening) "Open", sum(executions) "Exec" FROM v$sql GROUP BY substr(sql_text,1,60) HAVING sum(sharable_mem) > 20000 ; SELECT substr(sql_text,1,100) "Stmt", count(*), sum(sharable_mem) "Mem", sum(users_opening) "Open", sum(executions) "Exec" FROM v$sql GROUP BY substr(sql_text,1,60) HAVING sum(executions) > 200 ; SELECT substr(sql_text,1,100) "Stmt", count(*), sum(executions) "Exec" FROM v$sql GROUP BY substr(sql_text,1,100) HAVING sum(executions) > 200 ; where MEMSIZE is about 10% of the shared pool size in bytes. This should show if there are similar literal statements, or multiple versions of a statements which account for a large portion of the memory in the shared pool. 1.2 statistics: --------------- - Rule based / Cost based - apply EXPLAIN PLAN in query - ANALYZE COMMAND: ANALYZE TABLE EMPLOYEE COMPUTE STATISTICS; ANALYZE TABLE EMPLOYEE COMPUTE STATISTICS FOR ALL INDEXES; ANALYZE INDEX scott.indx1 COMPUTE STATISTICS; ANALYZE TABLE EMPLOYEE ESTIMATE STATISTICS SAMPLE 10 PERCENT; ALTER TABLE EMPLOYEE DELETE STATISTICS; - DBMS_UTILITY.ANALYZE_SCHEMA() procedure: DBMS_UTILITY.ANALYZE_SCHEMA ( schema VARCHAR2, method VARCHAR2, estimate_rows NUMBER DEFAULT NULL, estimate_percent NUMBER DEFAULT NULL, method_opt VARCHAR2 DEFAULT NULL); DBMS_UTILITY.ANALYZE_DATABASE ( method VARCHAR2, estimate_rows NUMBER DEFAULT NULL, estimate_percent NUMBER DEFAULT NULL, method_opt VARCHAR2 DEFAULT NULL); method=compute, estimate, delete To exexcute: exec DBMS_UTILITY.ANALYZE_SCHEMA('CISADM','COMPUTE'); 1.3 Storage parameters: ----------------------- segement: pctfree, pctused, number AND size of extends in STORAGE clause - very low updates : pctfree low - if updates, oltp : pctfree 10, pctused 40 - if only inserts : pctfree low 1.4 rebuild indexes on regular basis: ----------------------------------------- alter index SCOTT.EMPNO_INDEX rebuild tablespace INDEX storage (initial 5M next 5M pctincrease 0); You should next use the ANALYZE TABLE COMPUTE STATISTICS command 1.5 Is an index used in a query?: --------------------------------- De WHERE clause of a query must use the 'leading column' of (one of the) index(es): Suppose an index 'indx1' exists on EMPLOYEE(city, state, zip) Suppose a user issues the query: SELECT .. FROM EMPLOYEE WHERE state='NY' Then this query will not use that index! Therfore you must pay attention to the cardinal column of any index. 1.6 set transaction parameters: ------------------------------- ONLY ORACLE 7,8,8i: Suppose you must perform an action which will generate a lot of redo and rollback. If you want to influence which rollback segment will be used in your transactions, you can use the statement set transaction use rollback segment SEGMENT_NAME 1.7 Reduce fragmentation of a dictionary managed tablespace: ------------------------------------------------------------ alter tablespace DATA coalesce; 1.8 normalisation of tables: ---------------------------- The more tables are 'normalized', the higher the performance costs for queries joining tables 1.9 commits na zoveel rows: ---------------------------- declare i number := 0; cursor s1 is SELECT * FROM tab1 WHERE col1 = 'value1' FOR UPDATE; begin for c1 in s1 loop update tab1 set col1 = 'value2' WHERE current of s1; i := i + 1; -- Commit after every X records if i > 1000 then commit; i := 0; end if; end loop; commit; end; / -- ------------------------------ CREATE TABLE TEST ( ID NUMBER(10) NULL, DATUM DATE NULL, NAME VARCHAR2(10) NULL ); declare i number := 1000; begin while i>1 loop insert into TEST values (1, sysdate+i,'joop'); i := i - 1; commit; end loop; commit; end; / -- ------------------------------ CREATE TABLE TEST2 ( i number NULL, ID NUMBER(10) NULL, DATUM DATE NULL, DAG VARCHAR2(10) NULL, NAME VARCHAR2(10) NULL ); declare i number := 1; j date; k varchar2(10); begin while i<1000000 loop j:=sysdate+i; k:=TO_CHAR(SYSDATE+i,'DAY'); insert into TEST2 values (i,1, j, k,'joop'); i := i + 1; commit; end loop; commit; end; / -- ------------------------------ CREATE TABLE TEST3 ( ID NUMBER(10) NULL, DATUM DATE NULL, DAG VARCHAR2(10) NULL, VORIG VARCHAR2(10) NULL, NAME VARCHAR2(10) NULL ); declare i number := 1; j date; k varchar2(10); l varchar2(10); begin while i<1000 loop j:=sysdate+i; k:=TO_CHAR(SYSDATE+i,'DAY'); l:=TO_CHAR(SYSDATE+i-1,'DAY'); insert into TEST3 (ID,DATUM,DAG,VORIG,NAME) values (i, j, k, l,'joop'); i := i + 1; commit; end loop; commit; end; / 1.10 explain plan commAND, autotrace: ------------------------------------- 1 explain plan commAND: ----------------------- First execute the utlxplan.sql script. This script will create the PLAN_TABLE table, needed for storage of performance data. Now it's possible to do the following: -- optionally, delete the former performance data DELETE FROM plan_table WHERE statement_id = 'XXX'; COMMIT; -- now you can run the query that is to be analyzed EXPLAIN PLAN SET STATEMENT_ID = 'XXX' FOR SELECT * FROM EMPLOYEE WHERE city > 'Y%'; To view results, you can use the utlxpls.sql script. 2. set autotrace on / off ------------------------- Deze maakt ook gebruik van de PLAN_TABLE en de PLUSTRACE role moet bestaan. Desgewenst kan het plustrce.sql script worden uitgevoerd (onder SYS). Opmerking: Execution plan / access path bij een join query: - nested loop: 1 table is de driving table met full table scan of gebruik van index, en de tweede table wordt benadert m.b.v. een index van de tweede table gebaseerd op de WHERE clause. - merge join: als er geen bruikbare index is, worden alle rows opgehaald, gesorteerd, en gejoined naar een resultset. - Hash join: bepaalde init.ora parameters moeten aanwezig zijn (HASH_JOIN_ENABLE=TRUE, HASH_AREA_SIZE= , of via ALTER SESSION SET HASH_JOIN_ENABLED=TRUE). Meestal zeer effectief bij joins van een kleine table met een grote table. De kleine table is de driving table in memory en het vervolg is een algolritme wat lijkt op de nested loop Kan ook worden afgedwongen met een hint: SELECT /*+ USE_HASH(COMPANY) */ COMPANY.Name, SUM(Dollar_Amount) FROM COMPANY, SALES WHERE COMPANY.Company_ID = SALES.Company_ID GROUP BY COMPANY.Name; 3 SQL trace en TKPROFF ---------------------- SQL trace kan geactiveerd worden via init.ora of via ALTER SESSION SET SQL_TRACE=TRUE DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(sid, serial#, TRUE); DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(12, 398, TRUE); DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(12, 398, FALSE); DBMS_SUPPORT.START_TRACE_IN_SESSION(12,398); Turn SQL tracing on in session 448. The trace information will get written to user_dump_dest. SQL> exec dbms_system.set_sql_trace_in_session(448,2288,TRUE); Turn SQL tracing off in session 448 SQL> exec dbms_system.set_sql_trace_in_session(448,2288,FALSE); Init.ora: Max_dump_file_size in OS blocks SQL_TRACE=TRUE (kan zeer grote files opleveren, is voor alle sessions) USER_DUMP_DEST= lokatie trace files 1.12 Indien de CBO niet het beste access path gebruikt: hints in query: ----------------------------------------------------------------------- Goal hints: ALL_ROWS, FIRST_ROWS, CHOOSE, RULE Access methods hints: FULL, ROWID, CLUSTER, HASH, INDEX SELECT /*+ INDEX(emp_pk) */ FROM emp WHERE empno=12345; SELECT /*+ RULE */ ename, dname FROM emp, dept WHERE emp.deptno=dept.deptno ============================================== 3. Data dictonary queries m.b.t perfoRMANce: ============================================== 3.1 Reads AND writes in files: ------------------------------ V$FILESTAT, V$DATAFILE - Relative File I/O (1) SELECT fs.file#, df.file#, substr(df.name, 1, 50), fs.phyrds, fs.phywrts, df.status FROM v$filestat fs, v$datafile df WHERE fs.file#=df.file# - Relative File I/O (2) set pagesize 60 linesize 80 newpage 0 feedback off ttitle skip centre 'Datafile IO Weights' skip centre column Total_IO format 999999999 column Weigt format 999.99 column file_name format A40 break on drive skip 2 compute sum of Weight on Drive SELECT substr(DF.Name, 1, 6) Drive, DF.Name File_Name, FS.Phyblkrd+FS.Phyblkwrt Total_IO, 100*(FS.Phyblkrd+FS.Phyblkwrt) / MaxIO Weight FROM V$FILESTAT FS, V$DATAFILE DF, (SELECT MAX(Phyblkrd+Phyblkwrt) MaxIO FROM V$FILESTAT) WHERE DF.File#=FS.File# ORDER BY Weight desc / 3.2 undocumented init parameters: --------------------------------- SELECT * FROM SYS.X$KSPPI WHERE SUBSTR(KSPPINM,1,1) = '_'; 3.3 Kans op gebruik index of niet?: ----------------------------------- Kijk in DBA_TAB_COLUMNS.NUM_DISTINCT DBA_TABLES.NUM_ROWS als num_distinct in de buurt komt van num_rows : index favoriet i.p.v. full table Kijk in DBA_INDEXES, USER_INDEXES.CLUSTERING_FACTOR als clustering_factor = aantal blocks: ordered 3.4 snel overzicht hit ratio buffer cache: ------------------------------------------ Hit ratio= (LR - PR) / LR Stel er zijn nauwelijk Physical Reads PR, ofwel PR=0, dan is de Hit Ratio=LR/LR=1 Er worden dan geen blocks van disk gelezen. Praktijk: Hit ratio moet gemiddeld wel zo > 0,8 - 0,9 V$sess_io en v$sysstat en v$session kunnen geraadpleegd worden om de hit ratio te bepalen. V$sess_io: sid, consistent_gets, physical_reads V$session: sid, username SELECT name, value FROM v$sysstat WHERE name IN ('db block gets', 'consistent gets','physical reads'); SELECT (1-(pr.value/(dbg.value+cg.value)))*100 FROM v$sysstat pr, v$sysstat dbg, v$sysstat cg WHERE pr.name = 'physical reads' AND dbg.name = 'db block gets' AND cg.name = 'consistent gets'; -- uitgebeidere query m.b.t. hit ratio CLEAR SET HEAD ON SET VERIFY OFF col HitRatio format 999.99 heading 'Hit Ratio' col CGets format 9999999999999 heading 'Consistent Gets' col DBGets format 9999999999999 heading 'DB Block Gets' col PhyGets format 9999999999999 heading 'Physical Reads' SELECT substr(Username, 1, 10), v$sess_io.sid, consistent_gets, block_gets, physical_reads, 100*(consistent_gets+block_gets-physical_reads)/ (consistent_gets+block_gets) HitRatio FROM v$session, v$sess_io WHERE v$session.sid = v$sess_io.sid AND (consistent_gets+block_gets) > 0 AND Username is NOT NULL / SELECT 'Hit Ratio' Database, cg.value CGets, db.value DBGets, pr.value PhyGets, 100*(cg.value+db.value-pr.value)/(cg.value+db.value) HitRatio FROM v$sysstat db, v$sysstat cg, v$sysstat pr WHERE db.name = 'db block gets' AND cg.name = 'consistent gets' AND pr.name = 'physical reads' / 3.6 Wat zijn de actieve transacties?: ------------------------------------- SELECT substr(username, 1, 10), substr(terminal, 1, 10), substr(osuser, 1, 10), t.start_time, r.name, t.used_ublk "ROLLB BLKS", decode(t.space, 'YES', 'SPACE TX', decode(t.recursive, 'YES', 'RECURSIVE TX', decode(t.noundo, 'YES', 'NO UNDO TX', t.status) )) status FROM sys.v_$transaction t, sys.v_$rollname r, sys.v_$session s WHERE t.xidusn = r.usn AND t.ses_addr = s.saddr 3.7 sid's, resource belasting en locks: --------------------------------------- SELECT sid, lmode, ctime, block FROM v$lock SELECT s.sid, substr(s.username, 1, 10), substr(s.schemaname, 1, 10), substr(s.osuser, 1, 10), substr(s.program, 1, 10), s.command, l.lmode, l.block FROM v$session s, v$lock l WHERE s.sid=l.sid; SELECT l.addr, s.saddr, l.sid, s.sid, l.type, l.lmode, s.status, substr(s.schemaname, 1, 10), s.lockwait, s.row_wait_obj# FROM v$lock l, v$session s WHERE l.addr=s.saddr SELECT sid, substr(owner, 1, 10), substr(object, 1, 10) FROM v$access SID Session number that is accessing an object OWNER Owner of the object OBJECT Name of the object TYPE Type identifier for the object SELECT substr(s.username, 1, 10), s.sid, t.log_io, t.phy_io FROM v$session s, v$transaction t WHERE t.ses_addr=s.saddr 3.8 latch use in SGA (locks op process): ---------------------------------------- SELECT c.name,a.gets,a.misses,a.sleeps, a.immediate_gets,a.immediate_misses,b.pid FROM v$latch a, v$latchholder b, v$latchname c WHERE a.addr = b.laddr(+) AND a.latch# = c.latch# AND (c.name like 'redo%' or c.name like 'row%') ORDER BY a.latch#; column latch_name format a40 SELECT name latch_name, gets, misses, round(decode(gets-misses,0,1,gets-misses)/ decode(gets,0,1,gets),3) hit_ratio FROM v$latch WHERE name = 'redo allocation'; column latch_name format a40 SELECT name latch_name, immediate_gets, immediate_misses, round(decode(immediate_gets-immediate_misses,0,1, immediate_gets-immediate_misses)/ decode(immediate_gets,0,1,immediate_gets),3) hit_ratio FROM v$latch WHERE name = 'redo copy'; column name format a40 column value format a10 SELECT name,value FROM v$parameter WHERE name in ('log_small_entry_max_size','log_simultaneous_copies', 'cpu_count'); -- latches en locks in beeld set pagesize 23 set pause on set pause 'Hit any key...' col sid format 999999 col serial# format 999999 col username format a12 trunc col process format a8 trunc col terminal format a12 trunc col type format a12 trunc col lmode format a4 trunc col lrequest format a4 trunc col object format a73 trunc SELECT s.sid, s.serial#, decode(s.process, null, decode(substr(p.username,1,1), '?', upper(s.osuser), p.username), decode( p.username, 'ORACUSR ', upper(s.osuser), s.process) ) process, nvl(s.username, 'SYS ('||substr(p.username,1,4)||')') username, decode(s.terminal, null, rtrim(p.terminal, chr(0)), upper(s.terminal)) terminal, decode(l.type, -- Long locks 'TM', 'DML/DATA ENQ', 'TX', 'TRANSAC ENQ', 'UL', 'PLS USR LOCK', -- Short locks 'BL', 'BUF HASH TBL', 'CF', 'CONTROL FILE', 'CI', 'CROSS INST F', 'DF', 'DATA FILE ', 'CU', 'CURSOR BIND ', 'DL', 'DIRECT LOAD ', 'DM', 'MOUNT/STRTUP', 'DR', 'RECO LOCK ', 'DX', 'DISTRIB TRAN', 'FS', 'FILE SET ', 'IN', 'INSTANCE NUM', 'FI', 'SGA OPN FILE', 'IR', 'INSTCE RECVR', 'IS', 'GET STATE ', 'IV', 'LIBCACHE INV', 'KK', 'LOG SW KICK ', 'LS', 'LOG SWITCH ', 'MM', 'MOUNT DEF ', 'MR', 'MEDIA RECVRY', 'PF', 'PWFILE ENQ ', 'PR', 'PROCESS STRT', 'RT', 'REDO THREAD ', 'SC', 'SCN ENQ ', 'RW', 'ROW WAIT ', 'SM', 'SMON LOCK ', 'SN', 'SEQNO INSTCE', 'SQ', 'SEQNO ENQ ', 'ST', 'SPACE TRANSC', 'SV', 'SEQNO VALUE ', 'TA', 'GENERIC ENQ ', 'TD', 'DLL ENQ ', 'TE', 'EXTEND SEG ', 'TS', 'TEMP SEGMENT', 'TT', 'TEMP TABLE ', 'UN', 'USER NAME ', 'WL', 'WRITE REDO ', 'TYPE='||l.type) type, decode(l.lmode, 0, 'NONE', 1, 'NULL', 2, 'RS', 3, 'RX', 4, 'S', 5, 'RSX', 6, 'X', to_char(l.lmode) ) lmode, decode(l.request, 0, 'NONE', 1, 'NULL', 2, 'RS', 3, 'RX', 4, 'S', 5, 'RSX', 6, 'X', to_char(l.request) ) lrequest, decode(l.type, 'MR', decode(u.name, null, 'DICTIONARY OBJECT', u.name||'.'||o.name), 'TD', u.name||'.'||o.name, 'TM', u.name||'.'||o.name, 'RW', 'FILE#='||substr(l.id1,1,3)|| ' BLOCK#='||substr(l.id1,4,5)||' ROW='||l.id2, 'TX', 'RS+SLOT#'||l.id1||' WRP#'||l.id2, 'WL', 'REDO LOG FILE#='||l.id1, 'RT', 'THREAD='||l.id1, 'TS', decode(l.id2, 0, 'ENQUEUE', 'NEW BLOCK ALLOCATION'), 'ID1='||l.id1||' ID2='||l.id2) object FROM sys.v_$lock l, sys.v_$session s, sys.obj$ o, sys.user$ u, sys.v_$process p WHERE s.paddr = p.addr(+) AND l.sid = s.sid AND l.id1 = o.obj#(+) AND o.owner# = u.user#(+) AND l.type <> 'MR' UNION ALL /*** LATCH HOLDERS ***/ SELECT s.sid, s.serial#, s.process, s.username, s.terminal, 'LATCH', 'X', 'NONE', h.name||' ADDR='||rawtohex(laddr) FROM sys.v_$process p, sys.v_$session s, sys.v_$latchholder h WHERE h.pid = p.pid AND p.addr = s.paddr UNION ALL /*** LATCH WAITERS ***/ SELECT s.sid, s.serial#, s.process, s.username, s.terminal, 'LATCH', 'NONE', 'X', name||' LATCH='||p.latchwait FROM sys.v_$session s, sys.v_$process p, sys.v_$latch l WHERE latchwait is not null AND p.addr = s.paddr AND p.latchwait = l.addr / SELECT v.SID, v.BLOCK_GETS, v.BLOCK_CHANGES, w.USERNAME, w.OSUSER, w.TERMINAL FROM v$sess_io v, V$session w WHERE v.SID=w.SID ORDER BY v.SID; SQL> desc v$sess_io Name Null? Type ----------------------------- -------- -------------------- SID NUMBER BLOCK_GETS NUMBER CONSISTENT_GETS NUMBER PHYSICAL_READS NUMBER BLOCK_CHANGES NUMBER CONSISTENT_CHANGES NUMBER SQL> desc v$session; Name Null? Type ----------------------------- -------- -------------------- SADDR RAW(8) SID NUMBER SERIAL# NUMBER AUDSID NUMBER PADDR RAW(8) USER# NUMBER USERNAME VARCHAR2(30) COMMAND NUMBER OWNERID NUMBER TADDR VARCHAR2(16) LOCKWAIT VARCHAR2(16) STATUS VARCHAR2(8) SERVER VARCHAR2(9) SCHEMA# NUMBER SCHEMANAME VARCHAR2(30) OSUSER VARCHAR2(30) PROCESS VARCHAR2(12) MACHINE VARCHAR2(64) TERMINAL VARCHAR2(30) PROGRAM VARCHAR2(48) TYPE VARCHAR2(10) SQL_ADDRESS RAW(8) SQL_HASH_VALUE NUMBER SQL_ID VARCHAR2(13) SQL_CHILD_NUMBER NUMBER PREV_SQL_ADDR RAW(8) PREV_HASH_VALUE NUMBER PREV_SQL_ID VARCHAR2(13) PREV_CHILD_NUMBER NUMBER PLSQL_ENTRY_OBJECT_ID NUMBER PLSQL_ENTRY_SUBPROGRAM_ID NUMBER PLSQL_OBJECT_ID NUMBER PLSQL_SUBPROGRAM_ID NUMBER MODULE VARCHAR2(48) MODULE_HASH NUMBER ACTION VARCHAR2(32) ACTION_HASH NUMBER CLIENT_INFO VARCHAR2(64) FIXED_TABLE_SEQUENCE NUMBER ROW_WAIT_OBJ# NUMBER ROW_WAIT_FILE# NUMBER ROW_WAIT_BLOCK# NUMBER ROW_WAIT_ROW# NUMBER LOGON_TIME DATE LAST_CALL_ET NUMBER PDML_ENABLED VARCHAR2(3) FAILOVER_TYPE VARCHAR2(13) FAILOVER_METHOD VARCHAR2(10) FAILED_OVER VARCHAR2(3) RESOURCE_CONSUMER_GROUP VARCHAR2(32) PDML_STATUS VARCHAR2(8) PDDL_STATUS VARCHAR2(8) PQ_STATUS VARCHAR2(8) CURRENT_QUEUE_DURATION NUMBER CLIENT_IDENTIFIER VARCHAR2(64) BLOCKING_SESSION_STATUS VARCHAR2(11) BLOCKING_INSTANCE NUMBER BLOCKING_SESSION NUMBER SEQ# NUMBER EVENT# NUMBER EVENT VARCHAR2(64) P1TEXT VARCHAR2(64) P1 NUMBER P1RAW RAW(8) P2TEXT VARCHAR2(64) P2 NUMBER P2RAW RAW(8) P3TEXT VARCHAR2(64) P3 NUMBER P3RAW RAW(8) WAIT_CLASS_ID NUMBER WAIT_CLASS# NUMBER WAIT_CLASS VARCHAR2(64) WAIT_TIME NUMBER SECONDS_IN_WAIT NUMBER STATE VARCHAR2(19) SERVICE_NAME VARCHAR2(64) SQL_TRACE VARCHAR2(8) SQL_TRACE_WAITS VARCHAR2(5) SQL_TRACE_BINDS VARCHAR2(5) SQL> ======================================================== 4. IMP and EXP, IMPDP and EXPDP, and SQL*Loader Examples ======================================================== 4.1 EXPDP and IMPDP examples: ============================= New for Oracle 10g, are the impdp and expdp utilities. EXPDP practice/practice PARFILE=par1.par EXPDP hr/hr DUMPFILE=export_dir:hr_schema.dmp LOGFILE=export_dir:hr_schema.explog EXPDP system/******** PARFILE=c:\rmancmd\dpe_1.expctl Oracle 10g provides two new views, DBA_DATAPUMP_JOBS and DBA_DATAPUMP_SESSIONS that allow the DBA to monitor the progress of all DataPump operations. But ofcourse there are other views as well: SQL> select view_name from dba_views where view_name like '%PUMP%'; VIEW_NAME DATAPUMP_PATHS DATAPUMP_PATHMAP DATAPUMP_TABLE_DATA DATAPUMP_OBJECT_CONNECT DATAPUMP_DDL_TRANSFORM_PARAMS DATAPUMP_REMAP_OBJECTS V_$DATAPUMP_JOB V_$DATAPUMP_SESSION GV_$DATAPUMP_JOB GV_$DATAPUMP_SESSION USER_DATAPUMP_JOBS DBA_DATAPUMP_JOBS DBA_DATAPUMP_SESSIONS AQ$KUPC$DATAPUMP_QUETAB_S AQ$_KUPC$DATAPUMP_QUETAB_F AQ$KUPC$DATAPUMP_QUETAB AQ$KUPC$DATAPUMP_QUETAB_R SELECT owner_name ,job_name ,operation ,job_mode ,state ,degree ,attached_sessions FROM dba_datapump_jobs ; SELECT DPS.owner_name ,DPS.job_name ,S.osuser FROM dba_datapump_sessions DPS ,v$session S WHERE S.saddr = DPS.saddr ; Oracle Data Pump is a newer, faster and more flexible alternative to the "exp" and "imp" utilities used in previous Oracle versions. In addition to basic import and export functionality data pump provides a PL/SQL API and support for external tables. For the examples to work we must first unlock the SCOTT account and create a directory object it can access: CONN sys/password@db10g AS SYSDBA ALTER USER scott IDENTIFIED BY tiger ACCOUNT UNLOCK; GRANT CREATE ANY DIRECTORY TO scott; CREATE OR REPLACE DIRECTORY test_dir AS '/u01/app/oracle/oradata/'; GRANT READ, WRITE ON DIRECTORY test_dir TO scott; -- Table Exports/Imports: ------------------------- The TABLES parameter is used to specify the tables that are to be exported. The following is an example of the table export and import syntax: $ expdp scott/tiger@db10g tables=EMP,DEPT directory=TEST_DIR dumpfile=EMP_DEPT.dmp logfile=expdpEMP_DEPT.log $ impdp scott/tiger@db10g tables=EMP,DEPT directory=TEST_DIR dumpfile=EMP_DEPT.dmp logfile=impdpEMP_DEPT.log For example output files see expdpEMP_DEPT.log and impdpEMP_DEPT.log. The "TABLE_EXISTS_ACTION=APPEND" parameter allows data to be imported into existing tables. -- Schema Exports/Imports: -------------------------- The OWNER parameter of exp has been replaced by the SCHEMAS parameter which is used to specify the schemas to be exported. The following is an example of the schema export and import syntax: $ expdp scott/tiger@db10g schemas=SCOTT directory=TEST_DIR dumpfile=SCOTT.dmp logfile=expdpSCOTT.log $ impdp scott/tiger@db10g schemas=SCOTT directory=TEST_DIR dumpfile=SCOTT.dmp logfile=impdpSCOTT.log For example output files see expdpSCOTT.log and impdpSCOTT.log. -- Database Exports/Imports: ---------------------------- The FULL parameter indicates that a complete database export is required. The following is an example of the full database export and import syntax: $ expdp system/password@db10g full=Y directory=TEST_DIR dumpfile=DB10G.dmp logfile=expdpDB10G.log $ impdp system/password@db10g full=Y directory=TEST_DIR dumpfile=DB10G.dmp logfile=impdpDB10G.log For an example output file see expdpDB10G.log. Example 1. EXPDP parfile ------------------------ JOB_NAME=NightlyDRExport DIRECTORY=export_dir DUMPFILE=export_dir:fulldb_%U.dmp LOGFILE=export_dir:NightlyDRExport.explog FULL=Y PARALLEL=2 FILESIZE=650M CONTENT=ALL STATUS=30 ESTIMATE_ONLY=Y Example 2. EXPDP parfile, only for getting an estimate of export size --------------------------------------------------------------- JOB_NAME=EstimateOnly DIRECTORY=export_dir LOGFILE=export_dir:EstimateOnly.explog FULL=Y CONTENT=DATA_ONLY ESTIMATE=STATISTICS ESTIMATE_ONLY=Y STATUS=60 Example 3. EXPDP parfile, only 1 schema, writing to multiple files with %U variable, limited to 650M ---------------------------------------------------------------------------------------------- JOB_NAME=SH_TABLESONLY DIRECTORY=export_dir DUMPFILE=export_dir:SHONLY_%U.dmp LOGFILE=export_dir:SH_TablesOnly.explog SCHEMAS=SH PARALLEL=2 FILESIZE=650M STATUS=60 Example 4. EXPDP parfile, multiple tables, writing to multiple files with %U variable, limited ---------------------------------------------------------------------------------------- JOB_NAME=HR_PAYROLL_REFRESH DIRECTORY=export_dir DUMPFILE=export_dir:HR_PAYROLL_REFRESH_%U.dmp LOGFILE=export_dir:HR_PAYROLL_REFRESH.explog STATUS=20 FILESIZE=132K CONTENT=ALL TABLES=HR.EMPLOYEES,HR.DEPARTMENTS,HR.PAYROLL_CHECKS,HR.PAYROLL_HOURLY,HR.PAYROLL_SALARY,HR.PAYROLL_TRANSACTIONS Example 5. EXPDP parfile, Exports all objects in the HR schema, including metadata, asof just before midnight on April 10, 2005 ------------------------------------------------------------------------------------------------------------------------- JOB_NAME=HREXPORT DIRECTORY=export_dir DUMPFILE=export_dir:HREXPORT_%U.dmp LOGFILE=export_dir:2005-04-10_HRExport.explog SCHEMAS=HR CONTENTS=ALL FLASHBACK_TIME=TO_TIMESTAMP"('04-10-2005 23:59', 'MM-DD-YYYY HH24:MI')" Example 6. IMPDP parfile, Imports data +only+ into selected tables in the HR schema, Multiple dump files will be used ---------------------------------------------------------------------------------------------------------------------- JOB_NAME=HR_PAYROLL_IMPORT DIRECTORY=export_dir DUMPFILE=export_dir:HR_PAYROLL_REFRESH_%U.dmp LOGFILE=export_dir:HR_PAYROLL_IMPORT.implog STATUS=20 TABLES=HR.PAYROLL_CHECKS,HR.PAYROLL_HOURLY,HR.PAYROLL_SALARY,HR.PAYROLL_TRANSACTIONS CONTENT=DATA_ONLY TABLE_EXISTS_ACTION=TRUNCATE Example 7. IMPDP parfile,3 tables in the SH schema are the only tables to be refreshed,These tables will be truncated before loading -------------------------------------------------------------------------------------------------------------------------------- DIRECTORY=export_dir JOB_NAME=RefreshSHTables DUMPFILE=export_dir:fulldb_%U.dmp LOGFILE=export_dir:RefreshSHTables.implog STATUS=30 CONTENT=DATA_ONLY SCHEMAS=SH INCLUDE=TABLE:"IN('COUNTRIES','CUSTOMERS','PRODUCTS','SALES')" TABLE_EXISTS_ACTION=TRUNCATE Example IMPDP parfile,Generates SQLFILE output showing the DDL statements,Note that this code is +not+ executed! ---------------------------------------------------------------------------------------------------------------- DIRECTORY=export_dir JOB_NAME=GenerateImportDDL DUMPFILE=export_dir:hr_payroll_refresh_%U.dmp LOGFILE=export_dir:GenerateImportDDL.implog SQLFILE=export_dir:GenerateImportDDL.sql INCLUDE=TABLE Example: schedule a procedure which uses DBMS_DATAPUMP ------------------------------------------------------ BEGIN DBMS_SCHEDULER.CREATE_JOB ( job_name => 'HR_EXPORT' ,job_type => 'PLSQL_BLOCK' ,job_action => 'BEGIN HR.SP_EXPORT;END;' ,start_date => '04/18/2005 23:00:00.000000' ,repeat_interval => 'FREQ=DAILY' ,enabled => TRUE ,comments => 'Performs HR Schema Export nightly at 11 PM' ); END; / ====================================== How to use the NETWORK_LINK paramater: ====================================== Note 1: ======= Lora, the DBA at Acme Bank, is at the center of attention in a high-profile meeting of the bank's top management team. The objective is to identify ways of enabling end users to slice and dice the data in the company's main data warehouse. At the meeting, one idea presented is to create several small data marts—each based on a particular functional area—that can each be used by specialized teams. To effectively implement the data mart approach, the data specialists must get data into the data marts quickly and efficiently. The challenge the team faces is figuring out how to quickly refresh the warehouse data to the data marts, which run on heterogeneous platforms. And that's why Lora is at the meeting. What options does she propose for moving the data? An experienced and knowledgeable DBA, Lora provides the meeting attendees with three possibilities, as follows: Using transportable tablespaces Using Data Pump (Export and Import) Pulling tablespaces This article shows Lora's explanation of these options, including their implementation details and their pros and cons. Transportable Tablespaces: Lora starts by describing the transportable tablespaces option. The quickest way to transport an entire tablespace to a target system is to simply transfer the tablespace's underlying files, using FTP (file transfer protocol) or rcp (remote copy). However, just copying the Oracle data files is not sufficient; the target database must recognize and import the files and the corresponding tablespace before the tablespace data can become available to end users. Using transportable tablespaces involves copying the tablespace files and making the data available in the target database. A few checks are necessary before this option can be considered. First, for a tablespace TS1 to be transported to a target system, it must be self-contained. That is, all the indexes, partitions, and other dependent segments of the tables in the tablespace must be inside the tablespace. Lora explains that if a set of tablespaces contains all the dependent segments, the set is considered to be self-contained. For instance, if tablespaces TS1 and TS2 are to be transferred as a set and a table in TS1 has an index in TS2, the tablespace set is self-contained. However, if another index of a table in TS1 is in tablespace TS3, the tablespace set (TS1, TS2) is not self-contained. To transport the tablespaces, Lora proposes using the Data Pump Export utility in Oracle Database 10g. Data Pump is Oracle's next-generation data transfer tool, which replaces the earlier Oracle Export (EXP) and Import (IMP) tools. Unlike those older tools, which use regular SQL to extract and insert data, Data Pump uses proprietary APIs that bypass the SQL buffer, making the process extremely fast. In addition, Data Pump can extract specific objects, such as a particular stored procedure or a set of tables from a particular tablespace. Data Pump Export and Import are controlled by jobs, which the DBA can pause, restart, and stop at will. Lora has run a test before the meeting to see if Data Pump can handle Acme's requirements. Lora's test transports the TS1 and TS2 tablespaces as follows: 1. Check that the set of TS1 and TS2 tablespaces is self- contained. Issue the following command: BEGIN SYS.DBMS_TTS.TRANSPORT_SET_CHECK ('TS1','TS2'); END; 2. Identify any nontransportable sets. If no rows are selected, the tablespaces are self-contained: SELECT * FROM SYS.TRANSPORT_SET_VIOLATIONS; no rows selected 3. Ensure the tablespaces are read-only: SELECT STATUS FROM DBA_TABLESPACES WHERE TABLESPACE_NAME IN ('TS1','TS2'); STATUS --------- READ ONLY READ ONLY 4. Transfer the data files of each tablespace to the remote system, into the directory /u01/oradata, using a transfer mechanism such as FTP or rcp. 5. In the target database, create a database link to the source database (named srcdb in the line below). CREATE DATABASE LINK srcdb USING 'srcdb'; 6. In the target database, import the tablespaces into the database, using Data Pump Import. impdp lora/lora123 TRANSPORT_DATAFILES="'/u01/oradata/ts1_1.dbf','/u01/oradata/ts2_1.dbf'" NETWORK_LINK='srcdb' TRANSPORT_TABLESPACES=\(TS1,TS2\) NOLOGFILE=Y This step makes the TS1 and TS2 tablespaces and their data available in the target database. Note that Lora doesn't export the metadata from the source database. She merely specifies the value srcdb, the database link to the source database, for the parameter NETWORK_LINK in the impdp command above. Data Pump Import fetches the necessary metadata from the source across the database link and re-creates it in the target. 7. Finally, make the TS1 and TS2 tablespaces in the source database read-write. ALTER TABLESPACE TS1 READ WRITE; ALTER TABLESPACE TS2 READ WRITE; Note 2: ======= One of the most significant characteristics of an import operation is its mode, because the mode largely determines what is imported. The specified mode applies to the source of the operation, either a dump file set or another database if the NETWORK_LINK parameter is specified. The NETWORK_LINK parameter initiates a network import. This means that the impdp client initiates the import request, typically to the local database. That server contacts the remote source database referenced by the database link in the NETWORK_LINK parameter, retrieves the data, and writes it directly back to the target database. There are no dump files involved. In the following example, the source_database_link would be replaced with the name of a valid database link that must already exist. impdp hr/hr TABLES=employees DIRECTORY=dpump_dir1 NETWORK_LINK=source_database_link EXCLUDE=CONSTRAINT This example results in an import of the employees table (excluding constraints) from the source database. The log file is written to dpump_dir1, specified on the DIRECTORY parameter. 4.2 Export / Import examples: ============================= In all Oracle versions 7,8,8i,9i,10g you can use the exp and imp utilities. exp system/manager file=expdat.dmp compress=Y owner=(HARRY, PIET) exp system/manager file=hr.dmp owner=HR indexes=Y exp system/manager file=expdat.dmp TABLES=(john.SALES) imp system/manager file=hr.dmp full=Y buffer=64000 commit=Y imp system/manager file=expdat.dmp FROMuser=ted touser=john indexes=N commit=Y buffer=64000 imp rm_live/rm file=dump.dmp tables=(employee) imp system/manager file=expdat.dmp FROMuser=ted touser=john buffer=4194304 c:\> cd [oracle_db_home]\bin c:\> set nls_lang=american_america.WE8ISO8859P15 # export NLS_LANG=AMERICAN_AMERICA.UTF8 # export NLS_LANG=AMERICAN_AMERICA.AL32UTF8 c:\> imp system/manager fromuser=mis_owner touser=mis_owner file=[yourexport.dmp] FROM Oracle8i one can use the QUERY= export parameter to SELECTively unload a subset of the data FROM a table. Look at this example: exp scott/tiger tables=emp query=\"WHERE deptno=10\" -- Export metadata only: The Export utility is used to export the metadata describing the objects contained in the transported tablespace. For our example scenario, the Export command could be: EXP TRANSPORT_TABLESPACE=y TABLESPACES=ts_temp_sales FILE=jan_sales.dmp This operation will generate an export file, jan_sales.dmp. The export file will be small, because it contains only metadata. In this case, the export file will contain information describing the table temp_jan_sales, such as the column names, column datatype, and all other information that the target Oracle database will need in order to access the objects in ts_temp_sales. $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ Extended example: ----------------- CASE 1: ======= We create a user Albert on a 10g DB. This user will create a couple of tables with referential constraints (PK-FK relations). Then we will export this user, drop the user, and do an import. See what we have after the import. -- User: create user albert identified by albert default tablespace ts_cdc temporary tablespace temp QUOTA 10M ON sysaux QUOTA 20M ON users QUOTA 50M ON TS_CDC ; -- GRANTS: GRANT create session TO albert; GRANT create table TO albert; GRANT create sequence TO albert; GRANT create procedure TO albert; GRANT connect TO albert; GRANT resource TO albert; -- connect albert/albert -- create tables create table LOC -- table of locations ( LOCID int, CITY varchar2(16), constraint pk_loc primary key (locid) ); create table DEPT -- table of departments ( DEPID int, DEPTNAME varchar2(16), LOCID int, constraint pk_dept primary key (depid), constraint fk_dept_loc foreign key (locid) references loc(locid) ); create table EMP -- table of employees ( EMPID int, EMPNAME varchar2(16), DEPID int, constraint pk_emp primary key (empid), constraint fk_emp_dept foreign key (depid) references dept(depid) ); -- show constraints: SQL> select CONSTRAINT_NAME, CONSTRAINT_TYPE,TABLE_NAME,R_CONSTRAINT_NAME from user_constraints; CONSTRAINT_NAME C TABLE_NAME R_CONSTRAINT_NAME ------------------------------ - ------------------------------ ------------------------------ FK_EMP_DEPT R EMP PK_DEPT FK_DEPT_LOC R DEPT PK_LOC PK_LOC P LOC PK_DEPT P DEPT PK_EMP P EMP -- insert some data: INSERT INTO LOC VALUES (1,'Amsterdam'); INSERT INTO LOC VALUES (2,'Haarlem'); INSERT INTO LOC VALUES (3,null); INSERT INTO LOC VALUES (4,'Utrecht'); INSERT INTO DEPT VALUES (1,'Sales',1); INSERT INTO DEPT VALUES (2,'PZ',1); INSERT INTO DEPT VALUES (3,'Management',2); INSERT INTO DEPT VALUES (4,'RD',3); INSERT INTO DEPT VALUES (5,'IT',4); INSERT INTO EMP VALUES (1,'Joop',1); INSERT INTO EMP VALUES (2,'Gerrit',2); INSERT INTO EMP VALUES (3,'Harry',2); INSERT INTO EMP VALUES (4,'Christa',3); INSERT INTO EMP VALUES (5,null,4); INSERT INTO EMP VALUES (6,'Nina',5); INSERT INTO EMP VALUES (7,'Nadia',5); -- make an export C:\oracle\expimp>exp '/@test10g2 as sysdba' file=albert.dat owner=albert Export: Release 10.2.0.1.0 - Production on Sat Mar 1 08:03:59 2008 Copyright (c) 1982, 2005, Oracle. All rights reserved. Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production With the Partitioning, OLAP and Data Mining options Export done in WE8MSWIN1252 character set and AL16UTF16 NCHAR character set server uses AL32UTF8 character set (possible charset conversion) About to export specified users ... . exporting pre-schema procedural objects and actions . exporting foreign function library names for user ALBERT . exporting PUBLIC type synonyms . exporting private type synonyms . exporting object type definitions for user ALBERT About to export ALBERT's objects ... . exporting database links . exporting sequence numbers . exporting cluster definitions . about to export ALBERT's tables via Conventional Path ... . . exporting table DEPT 5 rows exported . . exporting table EMP 7 rows exported . . exporting table LOC 4 rows exported . exporting synonyms . exporting views . exporting stored procedures . exporting operators . exporting referential integrity constraints . exporting triggers . exporting indextypes . exporting bitmap, functional and extensible indexes . exporting posttables actions . exporting materialized views . exporting snapshot logs . exporting job queues . exporting refresh groups and children . exporting dimensions . exporting post-schema procedural objects and actions . exporting statistics Export terminated successfully without warnings. C:\oracle\expimp> -- drop user albert SQL>drop user albert cascade - create user albert See above -- do the import C:\oracle\expimp>imp '/@test10g2 as sysdba' file=albert.dat fromuser=albert touser=albert Import: Release 10.2.0.1.0 - Production on Sat Mar 1 08:09:26 2008 Copyright (c) 1982, 2005, Oracle. All rights reserved. Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production With the Partitioning, OLAP and Data Mining options Export file created by EXPORT:V10.02.01 via conventional path import done in WE8MSWIN1252 character set and AL16UTF16 NCHAR character set import server uses AL32UTF8 character set (possible charset conversion) . importing ALBERT's objects into ALBERT . . importing table "DEPT" 5 rows imported . . importing table "EMP" 7 rows imported . . importing table "LOC" 4 rows imported About to enable constraints... Import terminated successfully without warnings. C:\oracle\expimp> - connect albert/albert SQL> select * from emp; EMPID EMPNAME DEPID ---------- ---------------- ---------- 1 Joop 1 2 Gerrit 2 3 Harry 2 4 Christa 3 5 4 6 Nina 5 7 Nadia 5 7 rows selected. SQL> select * from loc; LOCID CITY ---------- ---------------- 1 Amsterdam 2 Haarlem 3 4 Utrecht SQL> select * from dept; DEPID DEPTNAME LOCID ---------- ---------------- ---------- 1 Sales 1 2 PZ 1 3 Management 2 4 RD 3 5 IT 4 -- show constraints: SQL> select CONSTRAINT_NAME, CONSTRAINT_TYPE,TABLE_NAME,R_CONSTRAINT_NAME from user_constraints; CONSTRAINT_NAME C TABLE_NAME R_CONSTRAINT_NAME ------------------------------ - ------------------------------ ------------------------------ FK_DEPT_LOC R DEPT PK_LOC FK_EMP_DEPT R EMP PK_DEPT PK_DEPT P DEPT PK_EMP P EMP PK_LOC P LOC Everything is back again. CASE 2: ======= We are not going to drop the user, but empty the tables: SQL> alter table dept disable constraint FK_DEPT_LOC; SQL> alter table emp disable constraint FK_EMP_DEPT; SQL> alter table dept disable constraint PK_DEPT; SQL> alter table emp disable constraint pk_emp; SQL> alter table loc disable constraint pk_loc; SQL> truncate table emp; SQL> truncate table loc; SQL> truncate table dept; -- do the import C:\oracle\expimp>imp '/@test10g2 as sysdba' file=albert.dat ignore=y fromuser=albert touser=albert Import: Release 10.2.0.1.0 - Production on Sat Mar 1 08:25:27 2008 Copyright (c) 1982, 2005, Oracle. All rights reserved. Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production With the Partitioning, OLAP and Data Mining options Export file created by EXPORT:V10.02.01 via conventional path import done in WE8MSWIN1252 character set and AL16UTF16 NCHAR character set import server uses AL32UTF8 character set (possible charset conversion) . importing ALBERT's objects into ALBERT . . importing table "DEPT" 5 rows imported . . importing table "EMP" 7 rows imported . . importing table "LOC" 4 rows imported About to enable constraints... IMP-00017: following statement failed with ORACLE error 2270: "ALTER TABLE "EMP" ENABLE CONSTRAINT "FK_EMP_DEPT"" IMP-00003: ORACLE error 2270 encountered ORA-02270: no matching unique or primary key for this column-list IMP-00017: following statement failed with ORACLE error 2270: "ALTER TABLE "DEPT" ENABLE CONSTRAINT "FK_DEPT_LOC"" IMP-00003: ORACLE error 2270 encountered ORA-02270: no matching unique or primary key for this column-list Import terminated successfully with warnings. So the data gets imported, but we have a problem with the FOREIGN KEYS: SQL> select CONSTRAINT_NAME, CONSTRAINT_TYPE,TABLE_NAME,R_CONSTRAINT_NAME, STATUS from user_constrai nts; CONSTRAINT_NAME C TABLE_NAME R_CONSTRAINT_NAME STATUS ------------------------------ - ------------------------------ ------------------------------ ----- FK_DEPT_LOC R DEPT PK_LOC DISABLED FK_EMP_DEPT R EMP PK_DEPT DISABLED PK_LOC P LOC DISABLED PK_EMP P EMP DISABLED PK_DEPT P DEPT DISABLED alter table dept enable constraint pk_dept; alter table emp enable constraint pk_emp; alter table loc enable constraint pk_loc; alter table dept enable constraint FK_DEPT_LOC; alter table emp enable constraint FK_EMP_DEPT; alter table dept enable constraint PK_DEPT; SQL> select CONSTRAINT_NAME, CONSTRAINT_TYPE,TABLE_NAME,R_CONSTRAINT_NAME, STATUS from user_constraints; CONSTRAINT_NAME C TABLE_NAME R_CONSTRAINT_NAME STATUS ------------------------------ - ------------------------------ ------------------------------ ----- FK_DEPT_LOC R DEPT PK_LOC ENABLED FK_EMP_DEPT R EMP PK_DEPT ENABLED PK_DEPT P DEPT ENABLED PK_EMP P EMP ENABLED PK_LOC P LOC ENABLED SQL> Everything is back again. $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ What is exported?: ------------------ Tables, indexes, data, database links gets exported. Example: -------- exp system/manager file=oemuser.dmp owner=oemuser Verbonden met: Oracle9i Enterprise Edition Release 9.0.1.4.0 - Production With the Partitioning option JServer Release 9.0.1.4.0 - Production. Export is uitgevoerd in WE8MSWIN1252 tekenset en AL16UTF16 NCHAR-tekenset. Export van opgegeven gebruikers gaat beginnen ... . pre-schema procedurele objecten en acties wordt geÙxporteerd. . bibliotheeknamen van verwijzende functie voor gebruiker OEMUSER worden geÙxpo teerd . objecttypedefinities voor gebruiker OEMUSER worden geÙxporteerd Export van objecten van OEMUSER gaat beginnen ... . databasekoppelingen worden geÙxporteerd. . volgnummers worden geÙxporteerd. . clusterdefinities worden geÙxporteerd. . export van tabellen van OEMUSER gaat beginnen ... via conventioneel pad ... . . tabel CUSTOMERS wordt geÙxporteerd.Er zijn 2 rijen geÙxporteerd. . synoniemen worden geÙxporteerd. . views worden geÙxporteerd. . opgeslagen procedures worden geÙxporteerd. . operatoren worden geÙxporteerd. . referentiÙle integriteitsbeperkingen worden geÙxporteerd. . triggers worden geÙxporteerd. . indextypen worden geÙxporteerd. . bitmap, functionele en uit te breiden indexen worden geÙxporteerd. . acties post-tabellen worden geÙxporteerd . snapshots worden geÙxporteerd. . logs voor snapshots worden geÙxporteerd. . takenwachtrijen worden geÙxporteerd . herschrijfgroepen en kinderen worden geÙxporteerd . dimensies worden geÙxporteerd. . post-schema procedurele objecten en acties wordt geÙxporteerd. . statistieken worden geÙxporteerd. Export is succesvol beÙindigd zonder waarschuwingen. D:\temp> Can one import tables to a different tablespace? ------------------------------------------------- Import the dump file using the INDEXFILE= option Edit the indexfile. Remove remarks and specify the correct tablespaces. Run this indexfile against your database, this will create the required tables in the appropriate tablespaces Import the table(s) with the IGNORE=Y option. Change the default tablespace for the user: Revoke the "UNLIMITED TABLESPACE" privilege FROM the user Revoke the user's quota FROM the tablespace FROM WHERE the object was exported. This forces the import utility to create tables in the user's default tablespace. Make the tablespace to which you want to import the default tablespace for the user Import the table Can one export to multiple files?/ Can one beat the Unix 2 Gig limit? --------------------------------------------------------------------- FROM Oracle8i, the export utility supports multiple output files. exp SCOTT/TIGER FILE=D:\F1.dmp,E:\F2.dmp FILESIZE=10m LOG=scott.log Use the following technique if you use an Oracle version prior to 8i: Create a compressed export on the fly. # create a named pipe mknod exp.pipe p # read the pipe - output to zip file in the background gzip < exp.pipe > scott.exp.gz & # feed the pipe exp userid=scott/tiger file=exp.pipe ... Some famous Errors: ------------------- Error 1: -------- EXP-00008: ORACLE error 6550 encountered ORA-06550: line 1, column 31: PLS-00302: component 'DBMS_EXPORT_EXTENSION' must be declared 1. The errors indicate that $ORACLE_HOME/rdbms/admin/CATALOG.SQL and $ORACLE_HOME/rdbms/admin/CATPROC.SQL Should be run again, as has been previously suggested. Were these scripts run connected as SYS? Try SELECT OBJECT_NAME, OBJECT_TYPE FROM DBA_OBJECTS WHERE STATUS = 'INVALID' AND OWNER = 'SYS'; Do you have invalid objects? Is DBMS_EXPORT_EXTENSION invalid? If so, try compiling it manually: ALTER PACKAGE DBMS_EXPORT_EXTENSION COMPILE BODY; If you receive errors during manual compilation, please show errors for further information. 2. Or possibly different imp/exp versions are run to another version of the database. The problem can be resolved by copying the higher version CATEXP.SQL and executed in the lesser version RDBMS. 3. Other fix: If there are problems in exp/imp from single byte to multibyte databases: - Analyze which tables/rows could be affected by national characters before running the export - Increase the size of affected rows. - Export the table data once again. Error 2: -------- EXP-00091: Exporting questionable statistics. Hi. This warning is generated because the statistics are questionable due to the client character set difference from the server character set. There is an article which discusses the causes of questionable statistics available via the MetaLink Advanced Search option by Doc ID: Doc ID: 159787.1 9i: Import STATISTICS=SAFE If you do not want this conversion to occur, you need to ensure the client NLS environment performing the export is set to match the server. Fix ~~~~ a) If the statistics of a table are not required to include in export take the export with parameter STATISTICS=NONE Example: $exp scott/tiger file=emp1.dmp tables=emp STATISTICS=NONE b) In case, the statistics are need to be included can use STATISTICS=ESTIMATE or COMPUTE (default is Estimate). Error 3: -------- EXP-00056: ORACLE error 1403 encountered ORA-01403: no data found EXP-00056: ORACLE error 1403 encountered ORA-01403: no data found EXP-00000: Export terminated unsuccessfully You can't export any DB with an exp utility of a newer version. The exp version must be equal or older than the DB version Doc ID : Note:281780.1 Content Type: TEXT/PLAIN Subject: Oracle 9.2.0.4.0: Schema Export Fails with ORA-1403 (No Data Found) on Exporting Cluster Definitions Creation Date: 29-AUG-2004 Type: PROBLEM Last Revision Date: 29-AUG-2004 Status: PUBLISHED The information in this article applies to: - Oracle Server - Enterprise Edition - Version: 9.2.0.4 to 9.2.0.4 - Oracle Server - Personal Edition - Version: 9.2.0.4 to 9.2.0.4 - Oracle Server - Standard Edition - Version: 9.2.0.4 to 9.2.0.4 This problem can occur on any platform. ERRORS ------ EXP-56 ORACLE error encountered ORA-1403 no data found EXP-0: Export terminated unsuccessfully SYMPTOMS -------- A schema level export with the 9.2.0.4 export utility from a 9.2.0.4 or higher release database in which XDB has been installed, fails when exporting the cluster definitions with: ... . exporting cluster definitions EXP-00056: ORACLE error 1403 encountered ORA-01403: no data found EXP-00000: Export terminated unsuccessfully You can confirm that XDB has been installed in the database: SQL> SELECT substr(comp_id,1,15) comp_id, status, substr(version,1,10) version, substr(comp_name,1,30) comp_name FROM dba_registry ORDER BY 1; COMP_ID STATUS VERSION COMP_NAME --------------- ----------- ---------- ------------------------------ ... XDB INVALID 9.2.0.4.0 Oracle XML Database XML VALID 9.2.0.6.0 Oracle XDK for Java XOQ LOADED 9.2.0.4.0 Oracle OLAP API You create a trace file of the ORA-1403 error: SQL> SHOW PARAMETER user_dump SQL> ALTER SYSTEM SET EVENTS '1403 trace name errorstack level 3'; System altered. -- Re-run the export SQL> ALTER SYSTEM SET EVENTS '1403 trace name errorstack off'; System altered. The trace file that was written to your USER_DUMP_DEST directory, shows: ksedmp: internal or fatal error ORA-01403: no data found Current SQL statement for this session: SELECT xdb_uid FROM SYS.EXU9XDBUID You can confirm that you have no invalid XDB objects in the database: SQL> SET lines 200 SQL> SELECT status, object_id, object_type, owner||'.'||object_name "OWNER.OBJECT" FROM dba_objects WHERE owner='XDB' AND status != 'VALID' ORDER BY 4,2; no rows selected Note: If you do have invalid XDB objects, and the same ORA-1403 error occurs when performing a full database export, see the solution mentioned in: [NOTE:255724.1] "Oracle 9i: Full Export Fails with ORA-1403 (No Data Found) on Exporting Cluster Defintions" CHANGES ------- You recently restored the database from a backup or you recreated the controlfile, or you performed Operating System actions on your database tempfiles. CAUSE ----- The Temporary tablespace does not have any tempfiles. Note that the errors are different when exporting with a 9.2.0.3 or earlier export utility: . exporting cluster definitions EXP-00056: ORACLE error 1157 encountered ORA-01157: cannot identify/lock data file 201 - see DBWR trace file ORA-01110: data file 201: 'M:\ORACLE\ORADATA\M9201WA\TEMP01.DBF' ORA-06512: at "SYS.DBMS_LOB", line 424 ORA-06512: at "SYS.DBMS_METADATA", line 1140 ORA-06512: at line 1 EXP-00000: Export terminated unsuccessfully The errors are also different when exporting with a 9.2.0.5 or later export utility: . exporting cluster definitions EXP-00056: ORACLE error 1157 encountered ORA-01157: cannot identify/lock data file 201 - see DBWR trace file ORA-01110: data file 201: 'M:\ORACLE\ORADATA\M9205WA\TEMP01.DBF' EXP-00000: Export terminated unsuccessfully FIX --- 1. If the controlfile does not have any reference to the tempfile(s), add the tempfile(s): SQL> SET lines 200 SQL> SELECT status, enabled, name FROM v$tempfile; no rows selected SQL> ALTER TABLESPACE temp ADD TEMPFILE 'M:\ORACLE\ORADATA\M9204WA\TEMP01.DBF' REUSE; or: If the controlfile has a reference to the tempfile(s), but the files are missing on disk, re-create the temporary tablespace, e.g.: SQL> SET lines 200 SQL> CREATE TEMPORARY TABLESPACE temp2 TEMPFILE 'M:\ORACLE\ORADATA\M9204WA\TEMP201.DBF' SIZE 100m AUTOEXTEND ON NEXT 100M MAXSIZE 2000M; SQL> ALTER DATABASE DEFAULT TEMPORARY TABLESPACE temp2; SQL> DROP TABLESPACE temp; SQL> CREATE TEMPORARY TABLESPACE temp TEMPFILE 'M:\ORACLE\ORADATA\M9204WA\TEMP01.DBF' SIZE 100m AUTOEXTEND ON NEXT 100M MAXSIZE 2000M; SQL> ALTER DATABASE DEFAULT TEMPORARY TABLESPACE temp; SQL> SHUTDOWN IMMEDIATE SQL> STARTUP SQL> DROP TABLESPACE temp2 INCLUDING CONTENTS AND DATAFILES; 2. Now re-run the export. Other errors: ------------- Doc ID : Note:175624.1 Content Type: TEXT/X-HTML Subject: Oracle Server - Export and Import FAQ Creation Date: 08-FEB-2002 Type: FAQ Last Revision Date: 16-FEB-2005 Status: PUBLISHED PURPOSE ======= This Frequently Asked Questions (FAQ) provides common Export and Import issues in the following sections: - GENERIC - LARGE FILES - INTERMEDIA - TOP EXPORT DEFECTS - COMPATIBILITY - TABLESPACE - ADVANCED QUEUING - TOP IMPORT DEFECTS - PARAMETERS - ORA-942 - REPLICATION - PERFORMANCE - NLS - FREQUENT ERRORS GENERIC ======= Question: What is actually happening when I export and import data? See Note 61949.1 "Overview of Export and Import in Oracle7" Question: What is important when doing a full database export or import? See Note 10767.1 "How to perform full system Export/Import" Question: Can data corruption occur using export & import (version 8.1.7.3 to 9.2.0)? See Note 199416.1 "ALERT: EXP Can Produce Dump File with Corrupted Data" Question: How to Connect AS SYSDBA when Using Export or Import? See Note 277237.1 "How to Connect AS SYSDBA when Using Export or Import" COMPATIBILITY ============= Question: Which version should I use when moving data between different database releases? See Note 132904.1 "Compatibility Matrix for Export & Import Between Different Oracle Versions" See Note 291024.1 "Compatibility and New Features when Transporting Tablespaces with Export and Import" See Note 76542.1 "NT: Exporting from Oracle8, Importing Into Oracle7" Question: How to resolve the IMP-69 error when importing into a database? See Note 163334.1 "Import Gets IMP-00069 when Importing 8.1.7 Export" See Note 1019280.102 "IMP-69 on Import" PARAMETERS ========== Question: What is the difference between a Direct Path and a Conventional Path Export? See Note 155477.1 "Parameter DIRECT: Conventional Path Export versus Direct Path Export" Question: What is the meaning of the Export parameter CONSISTENT=Y and when should I use it? See Note 113450.1 "When to Use CONSISTENT=Y During an Export" Question: How to use the Oracle8i/9i Export parameter QUERY=... and what does it do? See Note 91864.1 "Query= Syntax in Export in 8i" See Note 277010.1 "How to Specify a Query in Oracle10g Export DataPump and Import DataPump" Question: How to create multiple export dumpfiles instead of one large file? See Note 290810.1 "Parameter FILESIZE - Make Export Write to Multiple Export Files" PERFORMANCE =========== Question: Import takes so long to complete. How can I improve the performance of Import? See Note 93763.1 "Tuning Considerations when Import is slow" Question: Why has export performance decreased after creating tables with LOB columns? See Note 281461.1 "Export and Import of Table with LOB Columns (like CLOB and BLOB) has Slow Performance" LARGE FILES =========== Question: Which commands to use for solving Export dump file problems on UNIX platforms? See Note 30528.1 "QREF: Export/Import/SQL*Load Large Files in Unix - Quick Reference" Question: How to solve the EXP-15 and EXP-2 errors when Export dump file is larger than 2Gb? See Note 62427.1 "2Gb or Not 2Gb - File limits in Oracle" See Note 1057099.6 "Unable to export when export file grows larger than 2GB" See Note 290810.1 "Parameter FILESIZE - Make Export Write to Multiple Export Files" Question: How to export to a tape device by using a named pipe? See Note 30428.1 "Exporting to Tape on Unix System" TABLESPACE ========== Question: How to transport tablespace between different versions? See Note 291024.1 "Compatibility and New Features when Transporting Tablespaces with Export and Import" Question: How to move tables to a different tablespace and/or different user? See Note 1012307.6 "Moving Tables Between Tablespaces Using EXPORT/IMPORT" See Note 1068183.6 "How to change the default tablespace when importing using the INDEXFILE option" Question: How can I export all tables of a specific tablespace? See Note 1039292.6 "How to Export Tables for a specific Tablespace" ORA-942 ======= Question: How to resolve an ORA-942 during import of the ORDSYS schema? See Note 109576.1 "Full Import shows Errors when adding Referential Constraint on Cartrige Tables" Question: How to resolve an ORA-942 during import of a snapshot (log) into a different schema? See Note 1017292.102 "IMP-00017 IMP-00003 ORA-00942 USING FROMUSER/TOUSER ON SNAPSHOT [LOG] IMPORT" Question: How to resolve an ORA-942 during import of a trigger on a renamed table? See Note 1020026.102 "ORA-01702, ORA-00942, ORA-25001, When Importing Triggers" Question: How to resolve an ORA-942 during import of one specific table? See Note 1013822.102 "ORA-00942: ON TABLE LEVEL IMPORT" NLS === Question: Which effect has the client's NLS_LANG setting on an export and import? See Note 227332.1 "NLS considerations in Import/Export - Frequently Asked Questions" See Note 15656.1 "Export/Import and NLS Considerations" Question: How to prevent the loss of diacritical marks during an export/import? See Note 96842.1 "Loss Of Diacritics When Performing EXPORT/IMPORT Due To Incorrect Charactersets" INTERMEDIA OBJECTS ================== Question: How to solve an EXP-78 when exporting metadata for an interMedia Text index? See Note 130080.1 "Problems with EXPORT after upgrading from 8.1.5 to 8.1.6" Question: I dropped the ORDSYS schema, but now I get ORA-6550 and PLS-201 when exporting? See Note 120540.1 "EXP-8 PLS-201 After Drop User ORDSYS" ADVANCED QUEUING OBJECTS ======================== Question: Why does export show ORA-1403 and ORA-6512 on an AQ object, after an upgrade? See Note 159952.1 "EXP-8 and ORA-1403 When Performing A Full Export" Question: How to resolve export errors on DBMS_AQADM_SYS and DBMS_AQ_SYS_EXP_INTERNAL? See Note 114739.1 "ORA-4068 while performing full database export" REPLICATION OBJECTS =================== Question: How to resolve import errors on DBMS_IJOB.SUBMIT for Replication jobs? See Note 137382.1 "IMP-3, PLS-306 Unable to Import Oracle8i JobQueues into Oracle8" Question: How to reorganize Replication base tables with Export and Import? See Note 1037317.6 "Move Replication System Tables using Export/Import for Oracle 8.X" FREQUENTLY REPORTED EXPORT/IMPORT ERRORS ======================================== EXP-00002: Error in writing to export file Note 1057099.6 "Unable to export when export file grows larger than 2GB" EXP-00002: error in writing to export file The export file could not be written to disk anymore, probably because the disk is full or the device has an error. Most of the time this is followed by a device (filesystem) error message indicating the problem. Possible causes are file systems that do not support a certain limit (eg. dump file size > 2Gb) or a disk/filesystem that ran out of space. EXP-00003: No storage definition found for segment(%s,%s) (EXP-3 EXP-0) Note 274076.1 "EXP-00003 When Exporting From Oracle9i 9.2.0.5.0 with a Pre-9.2.0.5.0 Export Utility" Note 124392.1 "EXP-3 while exporting Rollback Segment definitions during FULL Database Export" EXP-00067: "Direct path cannot export %s which contains object or lob data." Note 1048461.6 "EXP-00067 PERFORMING DIRECT PATH EXPORT" EXP-00079: Data in table %s is protected (EXP-79) Note 277606.1 "How to Prevent EXP-00079 or EXP-00080 Warning (Data in Table xxx is Protected) During Export" EXP-00091: Exporting questionable statistics Note 159787.1 "9i: Import STATISTICS=SAFE" IMP-00016: Required character set conversion (type %lu to %lu) not supported Note 168066.1 "IMP-16 When Importing Dumpfile into a Database Using Multibyte Characterset" IMP-00020: Long column too large for column buffer size Note 148740.1 "ALERT: Export of table with dropped functional index may cause IMP-20 on import" ORA-00904: Invalid column name (EXP-8 ORA-904 EXP-0) Note 106155.1 "EXP-00008 ORA-1003 ORA-904 During Export" Note 172220.1 "Export of Database fails with EXP-00904 and ORA-01003" Note 158048.1 "Oracle8i Export Fails on Synonym Export with EXP-8 and ORA-904" Note 130916.1 "ORA-904 using EXP73 against Oracle8/8i Database" Note 1017276.102 "Oracle8i Export Fails on Synonym Export with EXP-8 and ORA-904" ORA-01406: Fetched column value was truncated (EXP-8 ORA-1406 EXP-0) Note 163516.1 "EXP-0 and ORA-1406 during Export of Object Types" ORA-01422: Exact fetch returns more than requested number of rows Note 221178.1 "PLS-201 and ORA-06512 at 'XDB.DBMS_XDBUTIL_INT' while Exporting Database" Note 256548.1 "Export of Database with XDB Throws ORA-1422 Error" ORA-01555: Snapshot too old Note 113450.1 "When to Use CONSISTENT=Y During an Export" ORA-04030: Out of process memory when trying to allocate %s bytes (%s,%s) (IMP-3 ORA-4030 ORA-3113) Note 165016.1 "Corrupt Packages When Export/Import Wrapper PL/SQL Code" ORA-06512: at "SYS.DBMS_STATS", line ... (IMP-17 IMP-3 ORA-20001 ORA-6512) Note 123355.1 "IMP-17 and IMP-3 errors referring dbms_stats package during import" ORA-29344: Owner validation failed - failed to match owner 'SYS' Note 294992.1 "Import DataPump: Transport Tablespace Fails with ORA-39123 and 29344 (Failed to match owner SYS)" ORA-29516: Aurora assertion failure: Assertion failure at %s (EXP-8 ORA-29516 EXP-0) Note 114356.1 "Export Fails With ORA-29516 Aurora Assertion Failure EXP-8" PLS-00103: Encountered the symbol "," when expecting one of the following ... (IMP-17 IMP-3 ORA-6550 PLS-103) Note 123355.1 "IMP-17 and IMP-3 errors referring dbms_stats package during import" Note 278937.1 "Import DataPump: ORA-39083 and PLS-103 when Importing Statistics Created with Non "." NLS Decimal Character" EXPORT TOP ISSUES CAUSED BY DEFECTS =================================== Release : 8.1.7.2 and below Problem : Export may fail with ORA-1406 when exporting object type definitions Solution : apply patch-set 8.1.7.3 Workaround: no, see Note 163516.1 "EXP-0 and ORA-1406 during Export of Object Types" Bug 1098503 Release : Oracle8i (8.1.x) and Oracle9i (9.x) Problem : EXP-79 when Exporting Protected Tables Solution : this is not a defect Workaround: N/A, see Note 277606.1 "How to Prevent EXP-00079 or EXP-00080 Warning (Data in Table xxx is Protected) During Export" Bug 2410612 Release : 8.1.7.3 and higher and 9.0.1.2 and higher Problem : Conventional export may produce an export file with corrupt data Solution : 8.1.7.5 and 9.2.0.x or check for Patch 2410612 (for 8.1.7.x), 2449113 (for 9.0.1.x) Workaround: yes, see Note 199416.1 "ALERT: Client Program May Give Incorrect Query Results (EXP Can Produce Dump File with Corrupted Data)" Release : Oracle8i (8.1.x) Problem : Full database export fails with EXP-3: no storage definition found for segment Solution : Oracle9i (9.x) Workaround: yes, see Note 124392.1 "EXP-3 while exporting Rollback Segment definitions during FULL Database Export" Bug 2900891 Release : 9.0.1.4 and below and 9.2.0.3 and below Problem : Export with 8.1.7.3 and 8.1.7.4 from Oracle9i fails with invalid identifier SPOLICY (EXP-8 ORA-904 EXP-0) Solution : 9.2.0.4 or 9.2.0.5 Workaround: yes, see Bug 2900891 how to recreate view sys.exu81rls Bug 2685696 Release : 9.2.0.3 and below Problem : Export fails when exporting triggers in call to XDB.DBMS_XDBUTIL_INT (EXP-56 ORA-1422 ORA-6512) Solution : 9.2.0.4 or check for Patch 2410612 (for 9.2.0.2 and 9.2.0.3) Workaround: yes, see Note 221178.1 "ORA-01422 ORA-06512: at "XDB.DBMS_XDBUTIL_INT" while exporting full database" Bug 2919120 Release : 9.2.0.4 and below Problem : Export fails when exporting triggers in call to XDB.DBMS_XDBUTIL_INT (EXP-56 ORA-1422 ORA-6512) Solution : 9.2.0.5 or check for Patch 2919120 (for 9.2.0.4) Workaround: yes, see Note 256548.1 "Export of Database with XDB Throws ORA-1422 Error" IMPORT TOP ISSUES CAUSED BY DEFECTS =================================== Bug 1335408 Release : 8.1.7.2 and below Problem : Bad export file using a locale with a ',' decimal seperator (IMP-17 IMP-3 ORA-6550 PLS-103) Solution : apply patch-set 8.1.7.3 or 8.1.7.4 Workaround: yes, see Note 123355.1 "IMP-17 and IMP-3 errors referring DBMS_STATS package during import" Bug 1879479 Release : 8.1.7.2 and below and 9.0.1.2 and below Problem : Export of a wrapped package can result in a corrupt package being imported (IMP-3 ORA-4030 ORA-3113 ORA-7445 ORA-600[16201]). Solution : in Oracle8i with 8.1.7.3 and higher; in Oracle9iR1 with 9.0.1.3 and higher Workaround: no, see Note 165016.1 "Corrupt Packages When Export/Import Wrapper PL/SQL Code" Bug 2067904 Release : Oracle8i (8.1.7.x) and 9.0.1.2 and below Problem : Trigger-name causes call to DBMS_DDL.SET_TRIGGER_FIRING_PROPERTY to fail during Import (IMP-17 IMP-3 ORA-931 ORA-23308 ORA-6512). Solution : in Oracle9iR1 with patchset 9.0.1.3 Workaround: yes, see Note 239821.1 "ORA-931 or ORA-23308 in SET_TRIGGER_FIRING_PROPERTY on Import of Trigger in 8.1.7.x and 9.0.1.x" Bug 2854856 Release : Oracle8i (8.1.7.x) and 9.0.1.2 and below Problem : Schema-name causes call to DBMS_DDL.SET_TRIGGER_FIRING_PROPERTY to fail during Import (IMP-17 IMP-3 ORA-911 ORA-6512). Solution : in Oracle9iR2 with patchset 9.2.0.4 Workaround: yes, see Note 239890.1 "ORA-911 in SET_TRIGGER_FIRING_PROPERTY on Import of Trigger in 8.1.7.x and Oracle9i" 4.3 SQL*Loader examples: -======================= SQL*Loader is used for loading data from text files into Oracle tables. The text file can have fixed column positions or columns separated by a special character, for example an ",". to call sqlloader sqlldr system/manager control=smssoft.ctl sqlldr parfile=bonus.par Example 1: ---------- BONUS.PAR: userid=scott control=bonus.ctl bad=bonus.bad log=bonus.log discard=bonus.dis rows=2 errors=2 skip=0 BONUS.CTL: LOAD DATA INFILE bonus.dat APPEND INTO TABLE BONUS (name position(01:08) char, city position(09:19) char, salary position(20:22) integer external) Now you can use the command: $ sqlldr parfile=bonus.par Example 2: ---------- LOAD1.CTL: LOAD DATA INFILE 'PLAYER.TXT' INTO TABLE BASEBALL_PLAYER FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"' (player_id,last_name,first_name,middle_initial,start_date) SQLLDR system/manager CONTROL=LOAD1.CTL LOG=LOAD1.LOG BAD=LOAD1.BAD DISCARD=LOAD1.DSC Example 3: another controlfile: ------------------------------ SMSSOFT.CTL: LOAD DATA INFILE 'SMSSOFT.TXT' TRUNCATE INTO TABLE SMSSOFTWARE FIELDS TERMINATED BY ',' OPTIONALLY ENCLOSED BY '"' (DWMACHINEID, SERIALNUMBER, NAME, SHORTNAME, SOFTWARE, CMDB_ID, LOGONNAME) Example 4: another controlfile: ------------------------------- LOAD DATA INFILE * BADFILE 'd:\stage\loader\load.bad' DISCARDFILE 'd:\stage\loader\load.dsc' APPEND INTO TABLE TEST FIELDS TERMINATED BY "" TRAILING NULLCOLS ( c1, c2 char, c3 date(8) "DD-MM-YY" ) BEGINDATA 1X25-12-00 2Y31-12-00 Note: The placeholder is only for illustration purposes, in the acutal implementation, one would use a real tab character which is not visible. - Convential path load: When the DIRECT=Y parameter is not used, the convential path is used. This means that essentially INSERT statements are used, triggers and referential integrety are in normal use, and that the buffer cache is used. - Direct path load: Buffer cache is not used. Existing used blocks are not used. New blocks are written as needed. Referential integrety and triggers are disabled during the load. Example 5: ---------- The following shows the control file (sh_sales.ctl) loading the sales table: LOAD DATA INFILE sh_sales.dat APPEND INTO TABLE sales FIELDS TERMINATED BY "|" (PROD_ID, CUST_ID, TIME_ID, CHANNEL_ID, PROMO_ID, QUANTITY_SOLD, AMOUNT_SOLD) It can be loaded with the following command: $ sqlldr sh/sh control=sh_sales.ctl direct=true 4.4 Creation of new table on basis of existing table: ===================================================== CREATE TABLE EMPLOYEE_2 AS SELECT * FROM EMPLOYEE CREATE TABLE temp_jan_sales NOLOGGING TABLESPACE ts_temp_sales AS SELECT * FROM sales WHERE time_id BETWEEN '31-DEC-1999' AND '01-FEB-2000'; insert into t SELECT * FROM t2; insert into DSA_IMPORT SELECT * FROM MDB_DW_COMPONENTEN@SALES 4.5 Copy commAND om data uit een remote database te halen: ========================================================== set copycommit 1 set arraysize 1000 copy FROM HR/PASSWORD@loc - create EMPLOYEE - using SELECT * FROM employee - WHERE state='NM' 4.6 Simple differences between table versions: ============================================== SELECT * FROM new_version MINUS SELECT * FROM old_version; SELECT * FROM old_version MINUS SELECT * FROM new_version; ======================================================= 5. Add, Move AND Size Datafiles, tablespaces, logfiles: ======================================================= 5.1 ADD OR DROP REDO LOGFILE GROUP: =================================== ADD: ---- alter database add logfile group 4 ('/db01/oracle/CC1/log_41.dbf', '/db02/oracle/CC1/log_42.dbf') size 5M; ALTER DATABASE ADD LOGFILE ('/oracle/dbs/log1c.rdo', '/oracle/dbs/log2c.rdo') SIZE 500K; ALTER DATABASE ADD LOGFILE ('/oracle/dbs/log1c.rdo', '/oracle/dbs/log2c.rdo') SIZE 500K; Add logfile plus group: ALTER DATABASE ADD LOGFILE GROUP 4 ('/dbms/tdbaeduc/educslot/recovery/redo_logs/redo04.log') SIZE 50M; ALTER DATABASE ADD LOGFILE GROUP 5 ('/dbms/tdbaeduc/educslot/recovery/redo_logs/redo05.log') SIZE 50M; ALTER DATABASE ADD LOGFILE ('G:\ORADATA\AIRM\REDO05.LOG') SIZE 20M; DROP: ----- -An instance requires at least two groups of online redo log files, regardless of the number of members in the groups. (A group is one or more members.) -You can drop an online redo log group only if it is inactive. If you need to drop the current group, first force a log switch to occur. ALTER DATABASE DROP LOGFILE GROUP 3; ALTER DATABASE DROP LOGFILE 'G:\ORADATA\AIRM\REDO02.LOG'; 5.2 ADD REDO LOGFILE MEMBER: ============================ alter database add logfile member '/db03/oracle/CC1/log_3c.dbf' to group 4; Note: More on ONLINE LOGFILES: ------------------------------ -- Log Files Without Redundancy LOGFILE GROUP 1 '/u01/oradata/redo01.log'SIZE 10M, GROUP 2 '/u02/oradata/redo02.log'SIZE 10M, GROUP 3 '/u03/oradata/redo03.log'SIZE 10M, GROUP 4 '/u04/oradata/redo04.log'SIZE 10M -- Log Files With Redundancy LOGFILE GROUP 1 ('/u01/oradata/redo1a.log','/u05/oradata/redo1b.log') SIZE 10M, GROUP 2 ('/u02/oradata/redo2a.log','/u06/oradata/redo2b.log') SIZE 10M, GROUP 3 ('/u03/oradata/redo3a.log','/u07/oradata/redo3b.log') SIZE 10M, GROUP 4 ('/u04/oradata/redo4a.log','/u08/oradata/redo4b.log') SIZE 10M -- Related Queries View information on log files SELECT * FROM gv$log; View information on log file history SELECT thread#, first_change#, TO_CHAR(first_time,'MM-DD-YY HH12:MIPM'), next_change# FROM gv$log_history; -- Forcing log file switches ALTER SYSTEM SWITCH LOGFILE; -- Clear A Log File If It Has Become Corrupt ALTER DATABASE CLEAR LOGFILE GROUP ; This statement overcomes two situations where dropping redo logs is not possible: If there are only two log groups The corrupt redo log file belongs to the current group. ALTER DATABASE CLEAR LOGFILE GROUP 4; -- Clear A Log File If It Has Become Corrupt And Avoid Archiving ALTER DATABASE CLEAR UNARCHIVED LOGFILE GROUP ; -- Use this version of clearing a log file if the corrupt log file has not been archived. ALTER DATABASE CLEAR UNARCHIVED LOGFILE GROUP 3; Managing Log File Groups Adding a redo log file group ALTER DATABASE ADD LOGFILE ('', '') SIZE ; ALTER DATABASE ADD LOGFILE ('/oracle/dbs/log1c.rdo', '/oracle/dbs/log2c.rdo') SIZE 500K; Adding a redo log file group and specifying the group number ALTER DATABASE ADD LOGFILE GROUP ('') SIZE ; ALTER DATABASE ADD LOGFILE GROUP 4 ('c:\temp\newlog1.log') SIZE 100M; Relocating redo log files ALTER DATABASE RENAME FILE '' TO ''; conn / as sysdba SELECT member FROM v_$logfile; SHUTDOWN; host $ cp /u03/logs/log1a.log /u04/logs/log1a.log $ cp /u03/logs/log1b.log /u05/logs/log1b.log $ exit startup mount ALTER DATABASE RENAME FILE '/u03/logs/log1a.log' TO '/u04/oradata/log1a.log'; ALTER DATABASE RENAME FILE '/u04/logs/log1b.log' TO '/u05/oradata/log1b.log'; ALTER DATABASE OPEN host $ rm /u03/logs/log1a.log $ rm /u03/logs/log1b.log $ exit SELECT member FROM v_$logfile; Drop a redo log file group ALTER DATABASE DROP LOGFILE GROUP ; ALTER DATABASE DROP LOGFILE GROUP 4; Managing Log File Members Adding log file group members ALTER DATABASE ADD LOGFILE MEMBER '' TO GROUP ; ALTER DATABASE ADD LOGFILE MEMBER '/oracle/dbs/log2b.rdo' TO GROUP 2; Dropping log file group members ALTER DATABASE DROP LOGFILE MEMBER ''; ALTER DATABASE DROP LOGFILE MEMBER '/oracle/dbs/log3c.rdo'; Dumping Log Files Dumping a log file to trace ALTER SYSTEM DUMP LOGFILE '' DBA MIN DBA MAX ; or ALTER SYSTEM DUMP LOGFILE '' TIME MIN TIME MIN conn uwclass/uwclass alter session set nls_date_format='MM/DD/YYYY HH24:MI:SS'; SELECT SYSDATE FROM dual; CREATE TABLE test AS SELECT owner, object_name, object_type FROM all_objects WHERE SUBSTR(object_name,1,1) BETWEEN 'A' AND 'W'; INSERT INTO test (owner, object_name, object_type) VALUES ('UWCLASS', 'log_dump', 'TEST'); COMMIT; conn / as sysdba SELECT ((SYSDATE-1/1440)-TO_DATE('01/01/2007','MM/DD/YYYY'))*86400 ssec FROM dual; ALTER SYSTEM DUMP LOGFILE 'c:\oracle\product\oradata\orabase\redo01.log' TIME MIN 579354757; Disable Log Archiving Stop log file archiving The following is undocumented and unsupported and should be used only with great care and following through tests. One might consider this for loading a data warehouse. Be sure to restart logging as soon as the load is complete or the system will be at extremely high risk. The rest of the database remains unchanged. The buffer cache works in exactly the same way, old buffers get overwritten, old dirty buffers get written to disk. It's just the process of physically flushing the redo buffer that gets disabled. I used it in a very large test environment where I wanted to perform a massive amount of changes (a process to convert blobs to clobs actually) and it was going to take days to complete. By disabling logging, I completed the task in hours and if anything untoward were to have happened, I was quite happy to restore the test database back from backup. ~ the above paraphrased from a private email from Richard Foote. conn / as sysdba SHUTDOWN; STARTUP MOUNT EXCLUSIVE; ALTER DATABASE NOARCHIVELOG; ALTER DATABASE OPEN; ALTER SYSTEM SET "_disable_logging"=TRUE; 5.3 RESIZE DATABASE FILE: ========================= alter database datafile '/db05/oracle/CC1/data01.dbf' rezise 400M; (increase or decrease size) alter tablespace DATA datafile '/db05/oracle/CC1/data01.dbf' rezise 400M; (increase or decrease size) 5.4 ADD FILE TO TABLESPACE: =========================== alter tablespace DATA add datafile '/db05/oracle/CC1/data02.dbf' size 50M autoextend ON maxsize unlimited; 5.5 ALTER STORAGE FOR FILE: =========================== alter database datafile '/db05/oracle/CC1/data01.dbf' autoextend ON maxsize unlimited; alter database datafile '/oradata/temp/temp.dbf' autoextend off; The AUTOEXTEND option cannot be turned OFF at for the entire tablespace with a single command. Each datafile within the tablespace must explicitly turn off the AUTOEXTEND option via the ALTER DATABASE command. +447960585647 5.6 MOVE OF DATA FILE: ====================== connect internal shutdown mv /db01/oracle/CC1/data01.dbf /db02/oracle/CC1 connect / as SYSDBA startup mount CC1 alter database rename file '/db01/oracle/CC1/data01.dbf' to '/db02/oracle/CC1/data01.dbf'; alter database open; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/sysaux01.dbf' to '/dbms/tdbaplay/playdwhs/database/default/sysaux01.dbf'; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/system01.dbf' to '/dbms/tdbaplay/playdwhs/database/default/system01.dbf'; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/temp01.dbf' to '/dbms/tdbaplay/playdwhs/database/default/temp01.dbf'; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/undotbs01.dbf' to '/dbms/tdbaplay/playdwhs/database/default/undotbs01.dbf'; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/users01.dbf' to '/dbms/tdbaplay/playdwhs/database/default/users01.dbf'; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/redo01.log' to '/dbms/tdbaplay/playdwhs/recovery/redo_logs/redo01.log'; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/redo02.log' to '/dbms/tdbaplay/playdwhs/recovery/redo_logs/redo02.log'; alter database rename file '/dbms/tdbaplay/playdwhs/database/playdwhs/redo03.log' to '/dbms/tdbaplay/playdwhs/recovery/redo_logs/redo03.log'; 5.7 MOVE OF REDO LOG FILE: ========================== connect internal shutdown mv /db05/oracle/CC1/redo01.dbf /db02/oracle/CC1 connect / as SYSDBA startup mount CC1 alter database rename file '/db05/oracle/CC1/redo01.dbf' to '/db02/oracle/CC1/redo01.dbf'; alter database open; in case of problems: ALTER DATABASE CLEAR LOGFILE GROUP n example: -------- shutdown immediate op Unix: mv /u01/oradata/spltst1/redo01.log /u02/oradata/spltst1/ mv /u03/oradata/spltst1/redo03.log /u02/oradata/spltst1/ startup mount pfile=/apps/oracle/admin/SPLTST1/pfile/init.ora alter database rename file '/u01/oradata/spltst1/redo01.log' to '/u02/oradata/spltst1/redo01.log'; alter database rename file '/u03/oradata/spltst1/redo03.log' to '/u02/oradata/spltst1/redo03.log'; alter database open; 5.8 Put a datafile or tablespace ONLINE or OFFLINE: =================================================== alter tablespace data offline; alter tablespace data online; alter database datafile 8 offline; alter database datafile 8 online; 5.9 ALTER DEFAULT STORAGE: ========================== alter tablespace AP_INDEX_SMALL default storage (initial 5M next 5M pctincrease 0); 5.10 CREATE TABLESPACE STORAGE PARAMETERS: ========================================== locally managed 9i style: -- autoallocate: ---------------- CREATE TABLESPACE DEMO DATAFILE '/u02/oracle/data/lmtbsb01.dbf' size 100M extent management local autoallocate; -- uniform size, 1M is default: ------------------------------- CREATE TABLESPACE LOBS DATAFILE 'f:\oracle\oradata\pegacc\lobs01.dbf' SIZE 3000M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 64K; CREATE TABLESPACE LOBS2 DATAFILE 'f:\oracle\oradata\pegacc\lobs02.dbf' SIZE 3000M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 1M; CREATE TABLESPACE CISTS_01 DATAFILE '/u04/oradata/pilactst/cists_01.dbf' SIZE 1000M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 128K; CREATE TABLESPACE CISTS_01 DATAFILE '/u01/oradata/spldev1/cists_01.dbf' SIZE 400M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 128K; CREATE TABLESPACE PUB DATAFILE 'C:\ORACLE\ORADATA\TEST10G\PUB.DBF' SIZE 50M EXTENT MANAGEMENT LOCAL AUTOALLOCATE SEGMENT SPACE MANAGEMENT AUTO; CREATE TABLESPACE STAGING DATAFILE 'C:\ORACLE\ORADATA\TEST10G\STAGING.DBF' SIZE 50M EXTENT MANAGEMENT LOCAL AUTOALLOCATE SEGMENT SPACE MANAGEMENT AUTO; CREATE TABLESPACE RMAN DATAFILE 'C:\ORACLE\ORADATA\RMAN\RMAN.DBF' SIZE 100M EXTENT MANAGEMENT LOCAL AUTOALLOCATE SEGMENT SPACE MANAGEMENT AUTO; CREATE TABLESPACE CISTS_01 DATAFILE '/u07/oradata/spldevp/cists_01.dbf' SIZE 1200M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 128K; CREATE TABLESPACE USERS DATAFILE '/u06/oradata/splpack/users01.dbf' SIZE 50M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 128K; CREATE TABLESPACE INDX DATAFILE '/u06/oradata/splpack/indx01.dbf' SIZE 100M EXTENT MANAGEMENT LOCAL UNIFORM SIZE CREATE TEMPORARY TABLESPACE TEMP TEMPFILE '/u07/oradata/spldevp/temp01.dbf' SIZE 200M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 10M; ALTER DATABASE DEFAULT TEMPORARY TABLESPACE TEMP; ALTER TABLESPACE CISTS_01 ADD DATAFILE '/u03/oradata/splplay/cists_02.dbf' SIZE 1000M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 128K; ALTER TABLESPACE UNDOTBS ADD DATAFILE '/dbms/tdbaprod/prodross/database/default/undotbs03.dbf' SIZE 2000M; alter tablespace DATA add datafile '/db05/oracle/CC1/data02.dbf' size 50M autoextend ON maxsize unlimited; -- segment management manual or automatic: -- --------------------------------------- We can have a locally managed tablespace, but the segment space management, via the free lists and the pct_free and pct_used parameters, be still used manually. To specify manual space management, use the SEGMENT SPACE MANAGEMENT MANUAL clause CREATE TABLESPACE INDX2 DATAFILE '/u06/oradata/bcict2/indx09.dbf' SIZE 5000M EXTENT MANAGEMENT LOCAL AUTOALLOCATE SEGMENT SPACE MANAGEMENT MANUAL; or if you want segement space management to be automatic: CREATE TABLESPACE INDX2 DATAFILE '/u06/oradata/bcict2/indx09.dbf' SIZE 5000M EXTENT MANAGEMENT LOCAL AUTOALLOCATE SEGMENT SPACE MANAGEMENT AUTO; -- temporary tablespace: ------------------------ CREATE TEMPORARY TABLESPACE TEMP TEMPFILE '/u04/oradata/pilactst/temp01.dbf' SIZE 200M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 10M; create user cisadm identified by cisadm default tablespace cists_01 temporary tablespace temp; create user cisuser identified by cisuser default tablespace cists_01 temporary tablespace temp; create user cisread identified by cisread default tablespace cists_01 temporary tablespace temp; grant connect to cisadm; grant connect to cisuser; grant connect to cisread; grant resource to cisadm; grant resource to cisuser; grant resource to cisread; CREATE TEMPORARY TABLESPACE TEMP TEMPFILE '/u04/oradata/bcict2/tempt01.dbf' SIZE 5000M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 100M; alter tablespace TEMP add tempfile '/u04/oradata/bcict2/temp02.dbf' SIZE 5000M; alter tablespace UNDO add file '/u04/oradata/bcict2/undo07.dbf' size 500M; ALTER DATABASE datafile '/u04/oradata/bcict2/undo07.dbf' RESIZE 3000M; CREATE TEMPORARY TABLESPACE TEMP2 TEMPFILE '/u04/oradata/bcict2/temp01.dbf' SIZE 5000M EXTENT MANAGEMENT LOCAL UNIFORM SIZE 100M; ALTER TABLESPACE TEMP ADD TEMPFILE '/u04/oradata/bcict2/tempt4.dbf' SIZE 5000M; 1 /u03/oradata/bcict2/temp.dbf 2 /u03/oradata/bcict2/temp01.dbf 3 /u03/oradata/bcict2/temp02.dbf ALTER DATABASE TEMPFILE '/u02/oracle/data/lmtemp02.dbf' DROP INCLUDING DATAFILES; The extent management clause is optional for temporary tablespaces because all temporary tablespaces are created with locally managed extents of a uniform size. The Oracle default for SIZE is 1M. But if you want to specify another value for SIZE, you can do so as shown in the above statement. The AUTOALLOCATE clause is not allowed for temporary tablespaces. If you get errors: ------------------ If the controlfile does not have any reference to the tempfile(s), add the tempfile(s): SQL> SET lines 200 SQL> SELECT status, enabled, name FROM v$tempfile; no rows selected SQL> ALTER TABLESPACE temp ADD TEMPFILE 'M:\ORACLE\ORADATA\M9204WA\TEMP01.DBF' REUSE; or: If the controlfile has a reference to the tempfile(s), but the files are missing on disk, re-create the temporary tablespace, e.g.: SQL> SET lines 200 SQL> CREATE TEMPORARY TABLESPACE temp2 TEMPFILE 'M:\ORACLE\ORADATA\M9204WA\TEMP201.DBF' SIZE 100m AUTOEXTEND ON NEXT 100M MAXSIZE 2000M; SQL> ALTER DATABASE DEFAULT TEMPORARY TABLESPACE temp2; SQL> DROP TABLESPACE temp; SQL> CREATE TEMPORARY TABLESPACE temp TEMPFILE 'M:\ORACLE\ORADATA\M9204WA\TEMP01.DBF' SIZE 100m AUTOEXTEND ON NEXT 100M MAXSIZE 2000M; SQL> ALTER DATABASE DEFAULT TEMPORARY TABLESPACE temp; SQL> SHUTDOWN IMMEDIATE SQL> STARTUP SQL> DROP TABLESPACE temp2 INCLUDING CONTENTS AND DATAFILES; -- undo tablespace: -- ---------------- CREATE UNDO TABLESPACE undotbs_02 DATAFILE '/u01/oracle/rbdb1/undo0201.dbf' SIZE 2M REUSE AUTOEXTEND ON; ALTER SYSTEM SET UNDO_TABLESPACE = undotbs_02; -- ROLLBACK TABLESPACE: -- -------------------- create tablespace RBS datafile '/disk01/oracle/oradata/DB1/rbs01.dbf' size 25M default storage ( initial 500K next 500K pctincrease 0 minextents 2 ); ####################################################################################### CREATE TABLESPACE "DRSYS" LOGGING DATAFILE '/u02/oradata/pegacc/drsys01.dbf' SIZE 20M REUSE AUTOEXTEND ON NEXT 1024K MAXSIZE UNLIMITED EXTENT MANAGEMENT LOCAL SEGMENT SPACE MANAGEMENT AUTO ; CREATE TABLESPACE "INDX" LOGGING DATAFILE '/u02/oradata/pegacc/indx01.dbf' SIZE 100M REUSE AUTOEXTEND ON NEXT 1024K MAXSIZE UNLIMITED EXTENT MANAGEMENT LOCAL SEGMENT SPACE MANAGEMENT AUTO ; CREATE TABLESPACE "TOOLS" LOGGING DATAFILE '/u02/oradata/pegacc/tools01.dbf' SIZE 100M REUSE AUTOEXTEND ON NEXT 1024K MAXSIZE UNLIMITED EXTENT MANAGEMENT LOCAL SEGMENT SPACE MANAGEMENT AUTO ; CREATE TABLESPACE "USERS" LOGGING DATAFILE '/u02/oradata/pegacc/users01.dbf' SIZE 1000M REUSE AUTOEXTEND ON NEXT 1024K MAXSIZE UNLIMITED EXTENT MANAGEMENT LOCAL SEGMENT SPACE MANAGEMENT AUTO ; CREATE TABLESPACE "XDB" LOGGING DATAFILE '/u02/oradata/pegacc/xdb01.dbf' SIZE 20M REUSE AUTOEXTEND ON NEXT 1024K MAXSIZE UNLIMITED EXTENT MANAGEMENT LOCAL SEGMENT SPACE MANAGEMENT AUTO ; CREATE TABLESPACE "LOBS" LOGGING DATAFILE '/u02/oradata/pegacc/lobs01.dbf' SIZE 2000M REUSE AUTOEXTEND ON NEXT 1024K MAXSIZE UNLIMITED EXTENT MANAGEMENT LOCAL UNIFORM SIZE 1M ; ####################################################################################### General form of a 8i type statement: CREATE TABLESPACE DATA DATAFILE 'G:\ORADATA\RCDB\DATA01.DBF' size 100M EXTENT MANAGEMENT DICTIONARY default storage ( initial 512K next 512K minextents 1 pctincrease 0 ) minimum extent 512K logging online peRMANENTt; More info: ---------- By declaring a tablespace as DICTIONARY managed, you are specifying that extent management for segments in this tablespace will be managed using the dictionary tables sys.fet$ and sys.uet$. Oracle updates these tables in the data dictionary whenever an extent is allocated, or freed for reuse. This is the default in Oracle8i when no extent management clause is used in the CREATE TABLESPACE statement. The sys.fet$ table is clustered in the C_TS# cluster. Because it is created without a SIZE clause, one block will be reserved in the cluster for each tablespace. Although, if a tablespace has more free extents than can be contained in a single cluster block, then cluster block chaining will occur which can significantly impact performance on the data dictionary and space management transactions in particular. Unfortunately, chaining in this cluster cannot be repaired without recreating the entire database. Preferably, the number of free extents in a tablespace should never be greater than can be recorded in the primary cluster block for that tablespace, which is about 500 free extents for a database with an 8K database block size. Used extents, on the other hand, are recorded in the data dictionary table sys.uet$, which is clustered in the C_FILE#_BLOCK# cluster. Unlike the C_TS# cluster, C_FILE#_BLOCK# is sized on the assumption that segments will have an average of just 4 or 5 extents each. Unless your data dictionary was specifically customized prior to database creation to allow for more used extents per segment, then creating segments with thousands of extents (like mentioned in the previous section) will cause excessive cluster block chaining in this cluster. The major dilemma with an excessive number of used and/or free extents is that they can misrepresent the operations of the dictionary cache LRU mechanism. Extents should therefore not be allowed to grow into the thousands, not because of the impact of full table scans, but rather the performance of the data dictionary and dictionary cache. A Locally Managed Tablespace is a tablespace that manages its own extents by maintaining a bitmap in each datafile to keep track of the free or used status of blocks in that datafile. Each bit in the bitmap corresponds to a block or a group of blocks. When the extents are allocated or freed for reuse, Oracle simply changes the bitmap values to show the new status of the blocks. These changes do not generate rollback information because they do not update tables in the data dictionary (except for tablespace quota information). This is the default in Oracle9i. If COMPATIBLE is set to 9.0.0, then the default extent management for any new tablespace is locally managed in Oracle9i. If COMPATIBLE is less than 9.0.0, then the default extent management for any new tablespace is dictionary managed in Oracle9i. While free space is represented in a bitmap within the tablespace, used extents are only recorded in the extent map in the segment header block of each segment, and if necessary, in additional extent map blocks within the segment. Keep in mind though, that this information is not cached in the dictionary cache. It must be obtained from the database block every time that it is required, and if those blocks are not in the buffer cache, that involves I/O and potentially lots of it. Take for example a query against DBA_EXTENTS. This query would be required to read every segment header and every additional extent map block in the entire database. It is for this reason that it is recommended that the number of extents per segment in locally managed tablespaces be limited to the number of rows that can be contained in the extent map with the segment header block. This would be approximately - (db_block_size / 16) - 7. For a database with a db block size of 8K, the above formula would be 505 extents. Drop a tempfile from the TEMP tablespace: alter tablespace TEMP drop tempfile '/dbms/tdbaaccp/accproca/database/default/temp02.dbf'; 5.11 DEALLOCATE EN OPSPOREN VAN UNUSED SPACE IN EEN TABLE: ========================================================== alter table emp deallocate unused; alter table emp deallocate unused keep 100K; alter table emp allocate extent ( size 100K datafile '/db05/oradata/CC1/user05.dbf'); Deze datafile moet in dezelfde tablespace bestaan. -- gebruik van de dbms_space.unused_space package declare var1 number; var2 number; var3 number; var4 number; var5 number; var6 number; var7 number; begin dbms_space.unused_space('AUTOPROV1', 'MACADDRESS_INDEX', 'INDEX', var1, var2, var3, var4, var5, var6, var7); dbms_output.put_line('OBJECT_NAME = NOG ZON SLECHTE INDEX'); dbms_output.put_line('TOTAL_BLOCKS ='||var1); dbms_output.put_line('TOTAL_BYTES ='||var2); dbms_output.put_line('UNUSED_BLOCKS ='||var3); dbms_output.put_line('UNUSED_BYTES ='||var4); dbms_output.put_line('LAST_USED_EXTENT_FILE_ID ='||var5); dbms_output.put_line('LAST_USED_EXTENT_BLOCK_ID ='||var6); dbms_output.put_line('LAST_USED_BLOCK ='||var7); end; / 5.12 CREATE TABLE: ================== -- STORAGE PARAMETERS EXAMPLE: -- --------------------------- create table emp ( id number, name varchar(2) ) tablespace users pctfree 10 storage (initial 1024K next 1024K pctincrease 10 minextents 2); ALTER a COLUMN: =============== ALTER TABLE GEWEIGERDETRANSACTIE MODIFY (VERBRUIKTIJD DATE); -- Creation of new table on basis of existing table: -- ------------------------------------------------- CREATE TABLE EMPLOYEE_2 AS SELECT * FROM EMPLOYEE insert into t SELECT * FROM t2; insert into DSA_IMPORT SELECT * FROM MDB_DW_COMPONENTEN@SALES -- Creation of a table with an autoincrement: -- ------------------------------------------ CREATE SEQUENCE seq_customer INCREMENT BY 1 START WITH 1 MAXVALUE 99999 NOCYCLE; CREATE SEQUENCE seq_employee INCREMENT BY 1 START WITH 1218 MAXVALUE 99999 NOCYCLE; CREATE SEQUENCE seq_a INCREMENT BY 1 START WITH 1 MAXVALUE 99999 NOCYCLE; CREATE TABLE CUSTOMER ( CUSTOMER_ID NUMBER (10) NOT NULL, NAAM VARCHAR2 (30) NOT NULL, CONSTRAINT PK_CUSTOMER PRIMARY KEY ( CUSTOMER_ID ) USING INDEX TABLESPACE INDX PCTFREE 10 STORAGE ( INITIAL 16K NEXT 16K PCTINCREASE 0 )) TABLESPACE USERS PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 STORAGE ( INITIAL 80K NEXT 80K PCTINCREASE 0 MINEXTENTS 1 MAXEXTENTS 2147483645 ) NOCACHE; CREATE OR REPLACE TRIGGER tr_CUSTOMER_ins BEFORE INSERT ON CUSTOMER FOR EACH ROW BEGIN SELECT seq_customer.NEXTVAL INTO :NEW.CUSTOMER_ID FROM dual; END; CREATE SEQUENCE seq_brains_verbruik INCREMENT BY 1 START WITH 1750795 MAXVALUE 100000000 NOCYCLE; CREATE OR REPLACE TRIGGER tr_PARENTEENHEID_ins BEFORE INSERT ON PARENTEENHEID FOR EACH ROW BEGIN SELECT seq_brains_verbruik.NEXTVAL INTO :NEW.VERBRUIKID FROM dual; END; 5.13 REBUILD OF INDEX: ====================== ALTER INDEX emp_pk REBUILD -- online 8.16 or higher NOLOGGING TABLESPACE INDEX_BIG PCTFREE 10 STORAGE ( INITIAL 5M NEXT 5M pctincrease 0 ); ALTER INDEX emp_ename INITRANS 5 MAXTRANS 10 STORAGE (PCTINCREASE 50); In situations where you have B*-tree index leaf blocks that can be freed up for reuse, you can merge those leaf blocks using the following statement: ALTER INDEX vmoore COALESCE; DROP INDEX emp_ename: -- Basic example of creating an index: CREATE INDEX emp_ename ON emp(ename) TABLESPACE users STORAGE (INITIAL 20K NEXT 20k PCTINCREASE 75) PCTFREE 0; If you have a LMT, you can just do: create index cust_indx on customers(id) nologging; This statement is without storage parameters. -- Dropping an index: DROP INDEX emp_ename: 5.14 MOVE TABLE TO OTHER TABLESPACE: ==================================== ALTER TABLE CHARLIE.CUSTOMERS MOVE TABLESPACE USERS2 5.15 SYNONYM (pointer to an object): ==================================== example: create public synonym EMPLOYEE for HARRY.EMPLOYEE; 5.16 DATABASE LINK: =================== CREATE PUBLIC DATABASE LINK SALESLINK CONNECT TO FRONTEND IDENTIFIED BY cygnusx1 USING 'SALES'; SELECT * FROM employee@MY_LINK; For example, using a database link to database sales.division3.acme.com, a user or application can reference remote data as follows: SELECT * FROM scott.emp@sales.division3.acme.com; # emp table in scott's schema SELECT loc FROM scott.dept@sales.division3.acme.com; If GLOBAL_NAMES is set to FALSE, then you can use any name for the link to sales.division3.acme.com. For example, you can call the link foo. Then, you can access the remote database as follows: SELECT name FROM scott.emp@foo; # link name different FROM global name Synonyms for Schema Objects: Oracle lets you create synonyms so that you can hide the database link name FROM the user. A synonym allows access to a table on a remote database using the same syntax that you would use to access a table on a local database. For example, assume you issue the following query against a table in a remote database: SELECT * FROM emp@hq.acme.com; You can create the synonym emp for emp@hq.acme.com so that you can issue the following query instead to access the same data: SELECT * FROM emp; View DATABASE LINKS: select substr(owner,1,10), substr(db_link,1,50), substr(username,1,25), substr(host,1,40), created from dba_db_links 5.17 TO CLEAR TABLESPACE TEMP: ============================== alter tablespace TEMP default storage (pctincrease 0); alter session set events 'immediate trace name DROP_SEGMENTS level TS#+1'; 5.18 RENAME OF OBJECT: ====================== RENAME sales_staff TO dept_30; RENAME emp2 TO emp; 5.19 CREATE PROFILE: ==================== CREATE PROFILE DEVELOPER LIMIT IDLE_TIME 10; CREATE PROFILE PRIOLIMIT LIMIT SESSIONS_PER_USER 10; ALTER USER GERRIT PROFILE DEVELOPER; ALTER PROFILE EXTERNLIMIT LIMIT PASSWORD_REUSE_TIME 90 PASSWORD_REUSE_MAX UNLIMITED; ALTER PROFILE EXTERNLIMIT LIMIT SESSIONS_PER_USER 20 IDLE_TIME 20; 5.20 RECOMPILE OF FUNCTION, PACKAGE, PROCEDURE: =============================================== ALTER FUNCTION schema.function COMPILE; example: ALTER FUNCTION oe.get_bal COMPILE; ALTER PACKAGE schema.package COMPILE specification/body/package example ALTER PACKAGE emp_mgmt COMPILE PACKAGE; ALTER PROCEDURE schema.procedure COMPILE; example ALTER PROCEDURE hr.remove_emp COMPILE; TO FIND OBJECTS: SELECT 'ALTER '||decode( object_type, 'PACKAGE SPECIFICATION' ,'PACKAGE' ,'PACKAGE BODY' ,'PACKAGE' ,object_type) ||' '||owner ||'.'|| object_name ||' COMPILE ' ||decode( object_type, 'PACKAGE SPECIFICATION' ,'SPECIFACTION' ,'PACKAGE BODY' ,'BODY' , NULL) ||';' FROM dba_objects WHERE status = 'INVALID'; 5.21 CREATE PACKAGE: ==================== A package is a set of related functions and / or routines. Packages are used to group together PL/SQL code blocks which make up a common application or are attached to a single business function. Packages consist of a specification and a body. The package specification lists the public interfaces to the blocks within the package body. The package body contains the public and private PL/SQL blocks which make up the application, private blocks are not defined in the package specification and cannot be called by any routine other than one defined within the package body. The benefits of packages are that they improve the organisation of procedure and function blocks, allow you to update the blocks that make up the package body without affecting the specification (which is the object that users have rights to) and allow you to grant execute rights once instead of for each and every block. To create a package specification we use a variation on the CREATE command, all we need put in the specification is each PL/SQL block header that will be public within the package. An example follows :- CREATE OR REPLACE PACKAGE MYPACK1 AS PROCEDURE MYPROC1 (REQISBN IN NUMBER, MYVAR1 IN OUT CHAR,TCOST OUT NUMBER); FUNCTION MYFUNC1; END MYPACK1; To create a package body we now specify each PL/SQL block that makes up the package, note that we are not creating these blocks separately (no CREATE OR REPLACE is required for the procedure and function definitions). An example follows :- CREATE OR REPLACE PACKAGE BODY MYPACK1 AS PROCEDURE MYPROC1 (REQISBN IN NUMBER, MYVAR1 IN OUT CHAR, TCOST OUT NUMBER) TEMP_COST NUMBER(10,2)) IS BEGIN SELECT COST FROM JD11.BOOK INTO TEMP_COST WHERE ISBN = REQISBN; IF TEMP_COST > 0 THEN UPDATE JD11.BOOK SET COST = (TEMP_COST*1.175) WHERE ISBN = REQISBN; ELSE UPDATE JD11.BOOK SET COST = 21.32 WHERE ISBN = REQISBN; END IF; TCOST := TEMP_COST; COMMIT; EXCEPTION WHEN NO_DATA_FOUND THEN INSERT INTO JD11.ERRORS (CODE, MESSAGE) VALUES(99, 'ISBN NOT FOUND'); END MYPROC1; FUNCTION MYFUNC1 RETURN NUMBER IS RCOST NUMBER(10,2); BEGIN SELECT COST FROM JD11.BOOK INTO RCOST WHERE ISBN = 21; RETURN (RCOST); END MYFUNC1; END MYPACK1; You can execute a public package block like this :- EXECUTE :PCOST := JD11.MYPACK1.MYFUNC1 - WHERE JD11 is the schema name that owns the package. You can use DROP PACKAGE and DROP PACKAGE BODY to remove the package objects FROM the database. CREATE OR REPLACE PACKAGE schema.package CREATE PACKAGE emp_mgmt AS FUNCTION hire (last_name VARCHAR2, job_id VARCHAR2, manager_id NUMBER, salary NUMBER, commission_pct NUMBER, department_id NUMBER) RETURN NUMBER; FUNCTION create_dept(department_id NUMBER, location NUMBER) RETURN NUMBER; PROCEDURE remove_emp(employee_id NUMBER); PROCEDURE remove_dept(department_id NUMBER); PROCEDURE increase_sal(employee_id NUMBER, salary_incr NUMBER); PROCEDURE increase_comm(employee_id NUMBER, comm_incr NUMBER); no_comm EXCEPTION; no_sal EXCEPTION; END emp_mgmt; / Before you can call this package's procedures and functions, you must define these procedures and functions in the package body. 5.22 View a view: ================= set long 2000 SELECT text FROM sys.dba_views WHERE view_name = 'CONTROL_PLAZA_V'; 5.23 ALTER SYSTEM: ================== ALTER SYSTEM CHECKPOINT; ALTER SYSTEM ENABLE/DISABLE RESTRICTED SESSION; ALTER SYSTEM FLUSH SHARED_POOL; ALTER SYSTEM SWITCH LOGFILE; ALTER SYSTEM SUSPEND/RESUME; ALTER SYSTEM SET RESOURCE_LIMIT = TRUE; ALTER SYSTEM SET LICENSE_MAX_USERS = 300; ALTER SYSTEM SET GLOBAL_NAMES=FALSE; ALTER SYSTEM SET COMPATIBLE = '9.2.0' SCOPE=SPFILE; 5.24 HOW TO ENABLE OR DISABLE TRIGGERS: ======================================= Disable enable trigger: ALTER TRIGGER Reorder DISABLE; ALTER TRIGGER Reorder ENABLE; Or in 1 time for all triggers on a table: ALTER TABLE Inventory DISABLE ALL TRIGGERS; 5.25 DIASABLING AND ENABLING AN INDEX: ====================================== alter index HEAT_CUSTOMER_POSTAL_CODE unusable; alter index HEAT_CUSTOMER_POSTAL_CODE rebuild; 5.26 CREATE A VIEW: =================== CREATE VIEW v1 AS SELECT LPAD(' ',40-length(size_tab.size_col)/2,' ') size_col FROM size_tab; CREATE VIEW X AS SELECT * FROM gebruiker@aptest 5.27 MAKE A USER: ================= CREATE USER jward IDENTIFIED BY aZ7bC2 DEFAULT TABLESPACE data_ts QUOTA 100M ON test_ts QUOTA 500K ON data_ts TEMPORARY TABLESPACE temp_ts PROFILE clerk; GRANT connect TO jward; create user jaap identified by jaap default tablespace users temporary tablespace temp; grant connect to jaap; grant resource to jaap; Dynamic queries: ---------------- -- CREATE USER AND GRANT PERMISSION STATEMENTS -- dynamic querieS SELECT 'CREATE USER '||USERNAME||' identified by '||USERNAME||' default tableSpace '|| DEFAULT_TABLESPACE||' temporary tableSpace '||TEMPORARY_TABLESPACE||';' FROM DBA_USERS WHERE USERNAME NOT IN ('SYS','SYSTEM','OUTLN','CTXSYS','ORDSYS','MDSYS'); SELECT 'GRANT CREATE SeSSion to '||USERNAME||';' FROM DBA_USERS WHERE USERNAME NOT IN ('SYS','SYSTEM','OUTLN','CTXSYS','ORDSYS','MDSYS'); SELECT 'GRANT connect to '||USERNAME||';' FROM DBA_USERS WHERE USERNAME NOT IN ('SYS','SYSTEM','OUTLN','CTXSYS','ORDSYS','MDSYS'); SELECT 'GRANT reSource to '||USERNAME||';' FROM DBA_USERS WHERE USERNAME NOT IN ('SYS','SYSTEM','OUTLN','CTXSYS','ORDSYS','MDSYS'); SELECT 'GRANT unlimited tableSpace to '||USERNAME||';' FROM DBA_USERS WHERE USERNAME NOT IN ('SYS','SYSTEM','OUTLN','CTXSYS','ORDSYS','MDSYS'); Becoming another user: ====================== - Do the query: select 'ALTER USER '||username||' IDENTIFIED BY VALUES '||''''||password||''''||';' from dba_users; - change the password - do what you need to do as the other account - change the password back to the original value -- grant to SELECT 'ALTER TABLE RM_LIVE.'||table_name||' disable constraint '||constraint_name||';' from dba_constraints where owner='RM_LIVE' and CONSTRAINT_TYPE='R'; SELECT 'ALTER TABLE RM_LIVE.'||table_name||' disable constraint '||constraint_name||';' from dba_constraints where owner='RM_LIVE' and CONSTRAINT_TYPE='P'; 5.28 CREATE A SEQUENCE: ======================= Sequences are database objects from which multiple users can generate unique integers. You can use sequences to automatically generate primary key values. CREATE SEQUENCE INCREMENT BY START WITH MAXVALUE CYCLE ; CREATE SEQUENCE department_seq INCREMENT BY 1 START WITH 1 MAXVALUE 99999 NOCYCLE; 5.29 STANDARD USERS IN 9i: ========================== CTXSYS is the primary schema for interMedia. MDSYS, ORDSYS, and ORDPLUGINS are schemas required when installing any of the cartridges. MTSSYS is required for the Oracle Service for MTS and is specific to NT. OUTLN is an integral part of the database required for the plan stability feature in Oracle8i. While the interMedia and cartridge schemas can be recreated by running their associated scripts as needed, I am not 100% on the steps associated with the MTSSYS user. Unfortunately, the OUTLN user is created at database creation time when sql.bsq is run. The OUTLN user owns the package OUTLN_PKG which is used to manage stored outlines and their outline categories. There are other tables (base tables), indexes, grants, and synonyms related to this package. By default, are automatically created during database creation : SCOTT by script $ORACLE_HOME/rdbms/admin/utlsampl.sql OUTLN by script $ORACLE_HOME/rdbms/admin/sql.bsq Optionally: DBSNMP if Enterprise Manager Intelligent Agent is installed TRACESVR if Enterprise Manager is installed AURORA$ORB$UNAUTHENTICATED \ AURORA$JIS$UTILITY$ -- if Oracle Servlet Engine (OSE) is installed OSE$HTTP$ADMIN / MDSYS if Oracle Spatial option is installed ORDSYS if interMedia Audio option is installed ORDPLUGINS if interMedia Audio option is installed CTXSYS if Oracle Text option is installed REPADMIN if Replication Option is installed LBACSYS if Oracle Label Security option is installed ODM if Oracle Data Mining option is installed ODM_MTR idem OLAPSYS if OLAP option is installed WMSYS if Oracle Workspace Manager script owmctab.plb is executed. ANONYMOUS if catqm.sql catalog script for SQL XML management XDB is executed 5.30 FORCED LOGGING: ==================== alter database no force logging; If a database is in force logging mode, all changes, except those in temporary tablespaces, will be logged, independently from any nologging specification. It is also possible to put arbitrary tablespaces into force logging mode: alter tablespace force logging. A force logging might take a while to complete because alter database add supplemental log data; ALTER DATABASE DROP SUPPLEMENTAL LOG DATA; ALTER TABLESPACE TDBA_CDC NO FORCE LOGGING; ==================================================== ORACLE INSTALLATIONS ON SOLARIS, LINUX, AIX, VMS: ==================================================== 6: Install on Solaris 7: Install on Linux 8: Install on OpenVMS 9: Install on AIX ================================== 6.1. Install Oracle 92 on Solaris: ================================== 6.1 Tutorial 1: =============== Short Guide to install Oracle 9.2.0 on SUN Solaris 8 -------------------------------------------------------------------------------- The Oracle 9i Distribution can be found on Oracle Technet (http://technet.oracle.com) The following, short Installation Guide shows how to install Oracle 9.2.0 for SUN Solaris 8. You may download our scripts to create a database, we suggest this way and NOT using DBASSIST. Besides this scripts, you can download our SQLNET configuration files TNSNAMES.ORA. LISTENER.ORA and SQLNET.ORA. Check Hardware Requirements Operating System Software Requirements Java Runtime Environment (JRE) Check Software Limits Setup the Solaris Kernel Create Unix Group «dba» Create Unix User «oracle» Setup ORACLE environment ($HOME/.profile) as follows Install from CD-ROM ... ... or Unpacking downloaded installation files Check oraInst.loc File Install with Installer in interactive mode Create the Database Start Listener Automatically Start / Stop the Database Install Oracle Options (optional) Download Scripts for Sun Solaris For our installation, we used the following ORACLE_HOME and ORACLE_SID, please adjust these parameters for your own environment. ORACLE_HOME = /opt/oracle/product/9.2.0 ORACLE_SID = TYP2 -------------------------------------------------------------------------------- Check Hardware Requirements Minimal Memory: 256 MB Minimal Swap Space: Twice the amount of the RAM To determine the amount of RAM memory installed on your system, enter the following command. $ /usr/sbin/prtconf To determine the amount of SWAP installed on your system, enter the following command and multiply the BLOCKS column by 512. $ swap -l Use the latest kernel patch from Sun Microsystems (http://sunsolve.sun.com) Operating System Software Requirements Use the latest kernel patch from Sun Microsystems. - Download the Patch from: http://sunsolve.sun.com - Read the README File included in the Patch - Usually the only thing you have to do is: $ cd $ ./install_custer $ cat /var/sadm/install_data/_log $ showrev -p - Reboot the system To determine your current operating system information: $ uname -a To determine which operating system patches are installed: $ showrev -p To determine which operating system packages are installed: $ pkginfo -i [package_name] To determine if your X-windows system is working properly on your local system, but you can redirect the X-windows output to another system. $ xclock To determine if you are using the correct system executables: $ /usr/bin/which make $ /usr/bin/which ar $ /usr/bin/which ld $ /usr/bin/which nm Each of the four commands above should point to the /usr/ccs/bin directory. If not, add /usr/ccs/bin to the beginning of the PATH environment variable in the current shell. Java Runtime Environment (JRE) The JRE shipped with Oracle9i is used by Oracle Java applications such as the Oracle Universal Installer is the only one supported. You should not modify this JRE, unless it is done through a patch provided by Oracle Support Services. The inventory can contain multiple versions of the JRE, each of which can be used by one or more products or releases. The Installer creates the oraInventory directory the first time it is run to keep an inventory of products that it installs on your system as well as other installation information. The location of oraInventory is defined in /var/opt/oracle/oraInst.loc. Products in an ORACLE_HOME access the JRE through a symbolic link in $ORACLE_HOME/JRE to the actual location of a JRE within the inventory. You should not modify the symbolic link. Check Software Limits Oracle9i includes native support for files greater than 2 GB. Check your shell to determine whether it will impose a limit. To check current soft shell limits, enter the following command: $ ulimit -Sa To check maximum hard limits, enter the following command: $ ulimit -Ha The file (blocks) value should be multiplied by 512 to obtain the maximum file size imposed by the shell. A value of unlimited is the operating system default and is the maximum value of 1 TB. Setup the Solaris Kernel Set to the sum of the PROCESSES parameter for each Oracle database, adding the largest one twice, then add an additional 10 for each database. For example, consider a system that has three Oracle instances with the PROCESSES parameter in their initSID.ora files set to the following values: ORACLE_SID=TYP1, PROCESSES=100 ORACLE_SID=TYP2, PROCESSES=100 ORACLE_SID=TYP3, PROCESSES=200 The value of SEMMNS is calculated as follows: SEMMNS = [(A=100) + (B=100)] + [(C=200) * 2] + [(# of instances=3) * 10] = 630 Setting parameters too high for the operating system can prevent the machine from booting up. Refer to Sun Microsystems Sun SPARC Solaris system administration documentation for parameter limits. * * Kernel Parameters on our SUN Enterprise with 640MB for Oracle 9 * set shmsys:shminfo_shmmax=4294967295 set shmsys:shminfo_shmmin=1 set shmsys:shminfo_shmmni=100 set shmsys:shminfo_shmseg=10 set semsys:seminfo_semmni=100 set semsys:seminfo_semmsl=100 set semsys:seminfo_semmns=2500 set semsys:seminfo_semopm=100 set semsys:seminfo_semvmx=32767 -- remarks: The parameter for shared memory (shminfo_shmmax) can be set to the maximum value; it will not impact Solaris in any way. The values for semaphores (seminfo_semmni and seminfo_semmns) depend on the number of clients you want to collect data from. As a rule of the thumb, the values should be set to at least (2*nr of clients + 15). You will have to reboot the system after making changes to the /etc/system file. Solaris doesn't automatically allocate shared memory, unless you specify the value in /etc/system and reboot. Were I you, i'd put in lines in /etc/system that look something like this: only the first value is *really* important. It specifies the maximum amount of shared memory to allocate. I'd make this parameter be about 70-75% of your physical ram (assuming you have nothing else on this machine running besides Oracle ... if not, adjust down accordingly). Then this value will dictate your maximum SGA size as you build your database. set shmsys:shminfo_shmmax=4294967295 set shmsys:shminfo_shmmin=1 set shmsys:shminfo_shmmni=100 set shmsys:shminfo_shmseg=10 set semsys:seminfo_semmsl=256 set semsys:seminfo_semmns=1024 set semsys:seminfo_semmni=400 -- end remarks Create Unix Group «dba» $ groupadd -g 400 dba $ groupdel dba Create Unix User «oracle» $ useradd -u 400 -c "Oracle Owner" -d /export/home/oracle \ -g "dba" -m -s /bin/ksh oracle Setup ORACLE environment ($HOME/.profile) as follows # Setup ORACLE environment ORACLE_HOME=/opt/oracle/product/9.2.0; export ORACLE_HOME ORACLE_SID=TYP2; export ORACLE_SID ORACLE_TERM=xterm; export ORACLE_TERM TNS_ADMIN=/export/home/oracle/config/9.2.0; export TNS_ADMIN NLS_LANG=AMERICAN_AMERICA.WE8ISO8859P1; export NLS_LANG ORA_NLS33=$ORACLE_HOME/ocommon/nls/admin/data; export ORA_NLS33 LD_LIBRARY_PATH=$ORACLE_HOME/lib:/lib:/usr/lib:/usr/openwin/lib LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/dt/lib:/usr/ucblib:/usr/local/lib export LD_LIBRARY_PATH # Set up the search paths: PATH=/bin:/usr/bin:/usr/sbin:/opt/bin:/usr/ccs/bin:/opt/local/GNU/bin PATH=$PATH:/opt/local/bin:/opt/NSCPnav/bin:$ORACLE_HOME/bin PATH=$PATH:/usr/local/samba/bin:/usr/ucb:. export PATH # CLASSPATH must include the following JRE location(s): CLASSPATH=$ORACLE_HOME/JRE:$ORACLE_HOME/jlib:$ORACLE_HOME/rdbms/jlib CLASSPATH=$CLASSPATH:$ORACLE_HOME/network/jlib Install from CD-ROM ... Usually the CD-ROM will be mounted automatically by the Solaris Volume Manager, if not, do it as follows as user root. $ su root $ mkdir /cdrom $ mount -r -F hsfs /dev/.... /cdrom exit or CTRL-D ... or Unpacking downloaded installation files If you downloaded database installation files from Oracle site (901solaris_disk1.cpio.gz, 901solaris_disk2.cpio.gz and 901solaris_disk3.cpio.gz) gunzip them somewhere and you'll get three .cpio files. The best way to download the huge files is to use the tool GetRight ( http://www.getright.com/ ) $ cd $ mkdir Disk1 Disk2 Disk3 $ cd Disk1 $ gunzip 901solaris_disk1.cpio.gz $ cat 901solaris_disk1.cpio | cpio -icd This will extract all the files for Disk1, repeat steps for Disk2 and D3isk3. Now you should have three directories (Disk1, Disk2 and Disk3) containing installation files. Check oraInst.loc File If you used Oracle before on your system, then you must edit the Oracle Inventory File, usually located in: /var/opt/oracle/oraInst.loc inventory_loc=/opt/oracle/product/oraInventory Install with Installer in interactive mode Install Oracle 9i with Oracle Installer $ cd /Disk1 $ DISPLAY=:0.0 $ export DISPLAY $ ./runInstaller example display: $ export DISPLAY=192.168.1.10:0.0 Answer the questions in the Installer, we use the following install directories Inventory Location: /opt/oracle/product/oraInventory Oracle Universal Installer in: /opt/oracle/product/oui Java Runtime Environment in: /opt/oracle/product/jre/1.1.8 Edit the Database Startup Script /var/opt/oracle/oratab TYP2:/opt/oracle/product/9.2.0:Y Create the Database Edit and save the CREATE DATABASE File initTYP2.sql in $ORACLE_HOME/dbs, or create a symbolic-Link from $ORACLE_HOME/dbs to your Location. $ cd $ORACLE_HOME/dbs $ ln -s /export/home/oracle/config/9.2.0/initTYP2.ora initTYP2.ora $ ls -l initTYP2.ora -> /export/home/oracle/config/9.2.0/initTYP2.ora First start the Instance, just to test your initTYP2.ora file for correct syntax and system resources. $ cd /export/home/oracle/config/9.2.0/ $ sqlplus /nolog SQL> connect / as sysdba SQL> startup nomount SQL> shutdown immediate Now you can create the database SQL> @initTYP2.sql SQL> @shutdown immediate SQL> startup Check the Logfile: initTYP2.log Start Listener $ lsnrctl start LSNRTYP2 Automatically Start / Stop the Database To start the Database automatically on Boot-Time, create or use our Startup Scripts dbora and lsnrora (included in ora_config_sol_920.tar.gz), which must be installed in /etc/init.d. Create symbolic Links from the Startup Directories. lrwxrwxrwx 1 root root S99dbora -> ../init.d/dbora* lrwxrwxrwx 1 root root S99lsnrora -> ../init.d/lsnrora* Install Oracle Options (optional) You may want to install the following Options: Oracle JVM Orcale XML Oracle Spatial Oracle Ultra Search Oracle OLAP Oracle Data Mining Example Schemas Run the following script install_options.sh to enable this options in the database. Before running this scripts adjust the initSID.ora paramaters as follows for the build process. After this, you can reset the paramters to smaller values. parallel_automatic_tuning = false shared_pool_size = 200000000 java_pool_size = 100000000 $ ./install_options.sh Download Scripts for Sun Solaris These Scripts can be used as Templates. Please note, that some Parameters like ORACLE_HOME, ORACLE_SID and PATH must be adjusted on your own Environment. Besides this, you should check the initSID.ora Parameters for your Database (Size, Archivelog, ...) 6.2 Environment oracle user: ---------------------------- typical profile for Oracle account on most unix systems: .profile -------- MAIL=/usr/mail/${LOGNAME:?} umask=022 EDITOR=vi; export EDITOR ORACLE_BASE=/opt/app/oracle; export ORACLE_BASE ORACLE_HOME=$ORACLE_BASE/product/9.2; export ORACLE_HOME ORACLE_SID=OWS; export ORACLE_SID ORACLE_TERM=xterm; export ORACLE_TERM TNS_ADMIN=$ORACLE_HOME/network/admin; export TNS_ADMIN NLS_LANG=AMERICAN_AMERICA.AL16UTF8; export NLS_LANG ORA_NLS33=$ORACLE_HOME/ocommon/nls/admin/data; export ORA_NLS33 LD_LIBRARY_PATH=$ORACLE_HOME/lib:/lib:/usr/lib:/usr/openwin/lib LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/dt/lib:/usr/ucblib:/usr/local/lib export LD_LIBRARY_PATH PATH=.:/usr/bin:/usr/sbin:/sbin:/usr/ucb:/etc:$ORACLE_HOME/lib:/usr/oasys/bin:$ORACLE_HOME/bin:/usr/local/bin: export PATH PS1='$PWD >' DISPLAY=172.17.2.128:0.0 export DISPLAY /etc >more passwd ----------------- root:x:0:1:Super-User:/:/sbin/sh daemon:x:1:1::/: bin:x:2:2::/usr/bin: sys:x:3:3::/: adm:x:4:4:Admin:/var/adm: lp:x:71:8:Line Printer Admin:/usr/spool/lp: uucp:x:5:5:uucp Admin:/usr/lib/uucp: nuucp:x:9:9:uucp Admin:/var/spool/uucppublic:/usr/lib/uucp/uucico smmsp:x:25:25:SendMail Message Submission Program:/: listen:x:37:4:Network Admin:/usr/net/nls: nobody:x:60001:60001:Nobody:/: noaccess:x:60002:60002:No Access User:/: nobody4:x:65534:65534:SunOS 4.x Nobody:/: avdsel:x:1002:100:Albert van der Sel:/export/home/avdsel:/bin/ksh oraclown:x:1001:102:Oracle owner:/export/home/oraclown:/bin/ksh brighta:x:1005:102:Bright Alley:/export/home/brighta:/bin/ksh customer:x:2000:102:Customer account:/export/home/customer:/usr/bin/tcsh /etc >more group ---------------- root::0:root other::1: bin::2:root,bin,daemon sys::3:root,bin,sys,adm adm::4:root,adm,daemon uucp::5:root,uucp mail::6:root tty::7:root,adm lp::8:root,lp,adm nuucp::9:root,nuucp staff::10: daemon::12:root,daemon sysadmin::14: smmsp::25:smmsp nobody::60001: noaccess::60002: nogroup::65534: dba::100:oraclown,brighta oper::101: oinstall::102: ===================================== 7. install Oracle 9i on Linux: ===================================== ==================== 7.1.Article 1: ==================== The Oracle 9i Distribution can be found on Oracle Technet (http://technet.oracle.com) The following short Guide shows how to install and configure Oracle 9.2.0 on RedHat Linux 7.2 / 8.0 You may download our Scripts to create a database, we suggest this way and NOT using DBASSIST. Besides these scripts, you can download our NET configuration files: LISTNER.ORA, TNSNAMES.ORA and SQLNET.ORA. System Requirements Create Unix Group «dba» Create Unix User «oracle» Setup Environment ($HOME/.bash_profile) as follows Mount the Oracle 9i CD-ROM (only if you have the CD) ... ... or Unpacking downloaded installation files Install with Installer in interactive mode Create the Database Create your own DB-Create Script (optional) Start Listener Automatically Start / Stop the Database Setup Kernel Parameters ( if necessary ) Install Oracle Options (optional) Download Scripts for RedHat Linux 7.2 For our installation, we used the following ORACLE_HOME AND ORACLE_SID, please adjust these parameters for your own environment. ORACLE_HOME = /opt/oracle/product/9.2.0 ORACLE_SID = VEN1 -------------------------------------------------------------------------------- System Requirements Oracle 9i needs Kernel Version 2.4 and glibc 2.2, which is included in RedHat Linux 7.2. Component Check with ... ... Output Liunx Kernel Version 2.4 rpm -q kernel kernel-2.4.7-10 System Libraries rpm -q glibc glibc-2.2.4-19.3 Proc*C/C++ rpm -q gcc gcc-2.96-98 Create Unix Group «dba» $ groupadd -g 400 dba Create Unix User «oracle» $ useradd -u 400 -c "Oracle Owner" -d /home/oracle \ -g "dba" -m -s /bin/bash oracle Setup Environment ($HOME/.bash_profile) as follows # Setup ORACLE environment ORACLE_HOME=/opt/oracle/product/9.2.0; export ORACLE_HOME ORACLE_SID=VEN1; export ORACLE_SID ORACLE_TERM=xterm; export ORACLE_TERM ORACLE_OWNER=oracle; export ORACLE_OWNER TNS_ADMIN=/home/oracle/config/9.2.0; export TNS_ADMIN NLS_LANG=AMERICAN_AMERICA.WE8ISO8859P1; export NLS_LANG ORA_NLS33=$ORACLE_HOME/ocommon/nls/admin/data; export ORA_NLS33 CLASSPATH=$ORACLE_HOME/jdbc/lib/classes111.zip LD_LIBRARY_PATH=$ORACLE_HOME/lib; export LD_LIBRARY_PATH ### see JSDK: export CLASSPATH # Set up JAVA and JSDK environment: export JAVA_HOME=/usr/local/jdk export JSDK_HOME=/usr/local/jsdk CLASSPATH=$CLASSPATH:$JAVA_HOME/lib:$JSDK_HOME/lib/jsdk.jar export CLASSPATH # Set up the search paths: PATH=$POSTFIX/bin:$POSTFIX/sbin:$POSTFIX/sendmail PATH=$PATH:/usr/local/jre/bin:/usr/local/jdk/bin:/bin:/sbin:/usr/bin:/usr/sbin PATH=$PATH:/usr/local/bin:$ORACLE_HOME/bin:/usr/local/jsdk/bin PATH=$PATH:/usr/local/sbin:/usr/bin/X11:/usr/X11R6/bin:/root/bin PATH=$PATH:/usr/local/samba/bin export PATH Mount the Oracle 9i CD-ROM (only if you have the CD) ... Mount the CD-ROM as user root. $ su root $ mkdir /cdrom $ mount -t iso9660 /dev/cdrom /cdrom $ exit ... or Unpacking downloaded installation files If you downloaded database installation files from Oracle site (Linux9i_Disk1.cpio.gz, Linux9i_Disk2.cpio.gz and Linux9i_Disk3.cpio.gz) gunzip them somewhere and you'll get three .cpio files. The best way to download the huge files is to use the tool GetRight ( http://www.getright.com/ ) $ cd $ cpio -idmv < Linux9i_Disk1.cpio $ cpio -idmv < Linux9i_Disk2.cpio $ cpio -idmv < Linux9i_Disk3.cpio Now you should have three directories (Disk1, Disk2 and Disk3) containing installation files. Install with Installer in interactive mode Install Oracle 9i with Oracle Installer $ cd Disk1 $ DISPLAY=:0.0 $ export DISPLAY $ ./runInstaller Answer the questions in the Installer, we use the following install directories Inventory Location: /opt/oracle/product/oraInventory Oracle Universal Installer in: /opt/oracle/product/oui Java Runtime Environment in: /opt/oracle/product/jre/1.1.8 Edit the Database Startup Script /etc/oratab VEN1:/opt/oracle/product/9.2.0:Y Create the Database Edit and save the CREATE DATABASE File initVEN1.sql in $ORACLE_HOME/dbs, or create a symbolic-Link from $ORACLE_HOME/dbs to your Location. $ cd $ORACLE_HOME/dbs $ ln -s /home/oracle/config/9.2.0/initVEN1.ora initVEN1.ora $ ls -l initVEN1.ora -> /home/oracle/config/9.2.0/initVEN1.ora First start the Instance, just to test your initVEN1.ora file for correct syntax and system resources. $ cd /home/oracle/config/9.2.0/ $ sqlplus /nolog SQL> connect / as sysdba SQL> startup nomount SQL> shutdown immediate Now you can create the database SQL> @initVEN1.sql SQL> @shutdown immediate SQL> startup Check the Logfile: initVEN1.log Create your own DB-Create Script (optional) You can generate your own DB-Create Script using the Tool: $ORACLE_HOME/bin/dbca Start Listener $ lsnrctl start LSNRVEN1 Automatically Start / Stop the Database To start the Database automatically on Boot-Time, create or use our Startup Scripts dbora and lsnrora (included in ora_config_linux_901.tar.gz), which must be installed in /etc/rc.d/init.d. Create symbolic Links from the Startup Directories in /etc/rc.d (e.g. /etc/rc.d/rc2.d). lrwxrwxrwx 1 root root S99dbora -> ../init.d/dbora* lrwxrwxrwx 1 root root S99lsnrora -> ../init.d/lsnrora* Setup Kernel Parameters ( if necessary ) Oracle9i uses UNIX resources such as shared memory, swap space, and semaphores extensively for interprocess communication. If your kernel parameter settings are insufficient for Oracle9i, you will experience problems during installation and instance startup. The greater the amount of data you can store in memory, the faster your database will operate. In addition, by maintaining data in memory, the UNIX kernel reduces disk I/O activity. Use the ipcs command to obtain a list of the system’s current shared memory and semaphore segments, and their identification number and owner. You can modify the kernel parameters by using the /proc file system. To modify kernel parameters using the /proc file system: 1. Log in as root user. 2. Change to the /proc/sys/kernel directory. 3. Review the current semaphore parameter values in the sem file using the cat or more utility # cat sem The output will list, in order, the values for the SEMMSL, SEMMNS, SEMOPM, and SEMMNI parameters. The following example shows how the output will appear. 250 32000 32 128 In the preceding example, 250 is the value of the SEMMSL parameter, 32000 is the value of the SEMMNS parameter, 32 is the value of the SEMOPM parameter, and 128 is the value of the SEMMNI parameter. 4. Modify the parameter values using the following command: # echo SEMMSL_value SEMMNS_value SEMOPM_value SEMMNI_value > sem In the preceding command, all parameters must be entered in order. 5. Review the current shared memory parameters using the cat or more utility. # cat shared_memory_parameter In the preceding example, the shared_memory_parameter is either the SHMMAX or SHMMNI parameter. The parameter name must be entered in lowercase letters. 6. Modify the shared memory parameter using the echo utility. For example, to modify the SHMMAX parameter, enter the following: # echo 2147483648 > shmmax 7. Write a script to initialize these values during system startup and include the script in your system init files. Refer to the following table to determine if your system shared memory and semaphore kernel parameters are set high enough for Oracle9i. The parameters in the following table are the minimum values required to run Oracle9i with a single database instance. You can put the initialization in the file /etc/rc.d/rc.local # Setup Kernel Parameters for Oracle 9i echo 250 32000 100 128 > /proc/sys/kernel/sem echo 2147483648 > /proc/sys/kernel/shmmax echo 4096 > /proc/sys/kernel/shmmni Install Oracle Options (optional) You may want to install the following Options: Oracle JVM Orcale XML Oracle Spatial Oracle Ultra Search Oracle OLAP Oracle Data Mining Example Schemas Run the following script install_options.sh to enable this options in the database. Before running this scripts adjust the initSID.ora paramaters as follows for the build process. After this, you can reset the paramters to smaller values. parallel_automatic_tuning = false shared_pool_size = 200000000 java_pool_size = 100000000 $ ./install_options.sh Download Scripts for RedHat Linux 7.2 These Scripts can be used as Templates. Please note, that some Parameters like ORACLE_HOME, ORACLE_SID and PATH must be adjusted on your own Environment. Besides this, you should check the initSID.ora Parameters for your Database (Size, Archivelog, ...) ==================== 7.2.Article 2: ==================== Installing Oracle9i (9.2.0.5.0) on Red Hat Linux (Fedora Core 2) by Jeff Hunter, Sr. Database Administrator -------------------------------------------------------------------------------- Contents Overview Swap Space Considerations Configuring Shared Memory Configuring Semaphores Configuring File Handles Create Oracle Account and Directories Configuring the Oracle Environment Configuring Oracle User Shell Limits Downloading / Unpacking the Oracle9i Installation Files Update Red Hat Linux System - (Oracle Metalink Note: 252217.1) Install the Oracle 9.2.0.4.0 RDBMS Software Install the Oracle 9.2.0.5.0 Patchset Post Installation Steps Creating the Oracle Database -------------------------------------------------------------------------------- Overview The following article is a summary of the steps required to successfully install the Oracle9i (9.2.0.4.0) RDBMS software on Red Hat Linux Fedora Core 2. Also included in this article is a detailed overview for applying the Oracle9i (9.2.0.5.0) patchset. Keep in mind the following assumptions throughout this article: When installing Red Hat Linux Fedora Core 2, I install ALL components. (Everything). This makes it easier than trying to troubleshoot missing software components. As of March 26, 2004, Oracle includes the Oracle9i RDBMS software with the 9.2.0.4.0 patchset already included. This will save considerable time since the patchset does not have to be downloaded and installed. We will, however, be applying the 9.2.0.5.0 patchset. Although it is not required, it is recommend to apply the 9.2.0.5.0 patchset. The post installation section includes steps for configuring the Oracle Networking files, configuring the database to start and stop when the machine is cycled, and other miscellaneous tasks. Finally, at the end of this article, we will be creating an Oracle 9.2.0.5.0 database named ORA920 using supplied scripts. -------------------------------------------------------------------------------- Swap Space Considerations Ensure enough swap space is available. Installing Oracle9i requires a minimum of 512MB of memory. (An inadequate amount of swap during the installation will cause the Oracle Universal Installer to either "hang" or "die") To check the amount of memory / swap you have allocated, type either: # free - OR - # cat /proc/swaps - OR - # cat /proc/meminfo | grep MemTotal If you have less than 512MB of memory (between your RAM and SWAP), you can add temporary swap space by creating a temporary swap file. This way you do not have to use a raw device or even more drastic, rebuild your system. As root, make a file that will act as additional swap space, let's say about 300MB: # dd if=/dev/zero of=tempswap bs=1k count=300000 Now we should change the file permissions: # chmod 600 tempswap Finally we format the "partition" as swap and add it to the swap space: # mke2fs tempswap # mkswap tempswap # swapon tempswap -------------------------------------------------------------------------------- Configuring Shared Memory The Oracle RDBMS uses shared memory in UNIX to allow processes to access common data structures and data. These data structures and data are placed in a shared memory segment to allow processes the fastest form of Interprocess Communications (IPC) available. The speed is primarily a result of processes not needing to copy data between each other to share common data and structures - relieving the kernel from having to get involved. Oracle uses shared memory in UNIX to hold its Shared Global Area (SGA). This is an area of memory within the Oracle instance that is shared by all Oracle backup and foreground processes. It is important to size the SGA to efficiently hold the database buffer cache, shared pool, redo log buffer as well as other shared Oracle memory structures. Inadequate sizing of the SGA can have a dramatic decrease in performance of the database. To determine all shared memory limits you can use the ipcs command. The following example shows the values of my shared memory limits on a fresh RedHat Linux install using the defaults: # ipcs -lm ------ Shared Memory Limits -------- max number of segments = 4096 max seg size (kbytes) = 32768 max total shared memory (kbytes) = 8388608 min seg size (bytes) = 1 Let's continue this section with an overview of the parameters that are responsible for configuring the shared memory settings in Linux. SHMMAX The SHMMAX parameter is used to define the maximum size (in bytes) for a shared memory segment and should be set large enough for the largest SGA size. If the SHMMAX is set incorrectly (too low), it is possible that the Oracle SGA (which is held in shared segments) may be limited in size. An inadequate SHMMAX setting would result in the following: ORA-27123: unable to attach to shared memory segment You can determine the value of SHMMAX by performing the following: # cat /proc/sys/kernel/shmmax 33554432 As you can see from the output above, the default value for SHMMAX is 32MB. This is often too small to configure the Oracle SGA. I generally set the SHMMAX parameter to 2GB. NOTE: With a 32-bit Linux operating system, the default maximum size of the SGA is 1.7GB. This is the reason I will often set the SHMMAX parameter to 2GB since it requires a larger value for SHMMAX. On a 32-bit Linux operating system, without Physical Address Extension (PAE), the physical memory is divided into a 3GB user space and a 1GB kernel space. It is therefore possible to create a 2.7GB SGA, but you will need make several changes at the Linux operating system level by changing the mapped base. In the case of a 2.7GB SGA, you would want to set the SHMMAX parameter to 3GB. Keep in mind that the maximum value of the SHMMAX parameter is 4GB. To change the value SHMMAX, you can use either of the following three methods: This is method I use most often. This method sets the SHMMAX on startup by inserting the following kernel parameter in the /etc/sysctl.conf startup file: # echo "kernel.shmmax=2147483648" >> /etc/sysctl.conf If you wanted to dynamically alter the value of SHMMAX without rebooting the machine, you can make this change directly to the /proc file system. This command can be made permanent by putting it into the /etc/rc.local startup file: # echo "2147483648" > /proc/sys/kernel/shmmax You can also use the sysctl command to change the value of SHMMAX: # sysctl -w kernel.shmmax=2147483648 SHMMNI We now look at the SHMMNI parameters. This kernel parameter is used to set the maximum number of shared memory segments system wide. The default value for this parameter is 4096. This value is sufficient and typically does not need to be changed. You can determine the value of SHMMNI by performing the following: # cat /proc/sys/kernel/shmmni 4096 SHMALL Finally, we look at the SHMALL shared memory kernel parameter. This parameter controls the total amount of shared memory (in pages) that can be used at one time on the system. In short, the value of this parameter should always be at least: ceil(SHMMAX/PAGE_SIZE) The default size of SHMALL is 2097152 and can be queried using the following command: # cat /proc/sys/kernel/shmall 2097152 From the above output, the total amount of shared memory (in bytes) that can be used at one time on the system is: SM = (SHMALL * PAGE_SIZE) = 2097152 * 4096 = 8,589,934,592 bytes The default setting for SHMALL should be adequate for our Oracle installation. NOTE: The page size in Red Hat Linux on the i386 platform is 4096 bytes. You can, however, use bigpages which supports the configuration of larger memory page sizes. -------------------------------------------------------------------------------- Configuring Semaphores Now that we have configured our shared memory settings, it is time to take care of configuring our semaphores. A semaphore can be thought of as a counter that is used to control access to a shared resource. Semaphores provide low level synchronization between processes (or threads within a process) so that only one process (or thread) has access to the shared segment, thereby ensureing the integrity of that shared resource. When an application requests semaphores, it does so using "sets". To determine all semaphore limits, use the following: # ipcs -ls ------ Semaphore Limits -------- max number of arrays = 128 max semaphores per array = 250 max semaphores system wide = 32000 max ops per semop call = 32 semaphore max value = 32767 You can also use the following command: # cat /proc/sys/kernel/sem 250 32000 32 128 SEMMSL The SEMMSL kernel parameter is used to control the maximum number of semaphores per semaphore set. Oracle recommends setting SEMMSL to the largest PROCESS instance parameter setting in the init.ora file for all databases hosted on the Linux system plus 10. Also, Oracle recommends setting the SEMMSL to a value of no less than 100. SEMMNI The SEMMNI kernel parameter is used to control the maximum number of semaphore sets on the entire Linux system. Oracle recommends setting the SEMMNI to a value of no less than 100. SEMMNS The SEMMNS kernel parameter is used to control the maximum number of semaphores (not semaphore sets) on the entire Linux system. Oracle recommends setting the SEMMNS to the sum of the PROCESSES instance parameter setting for each database on the system, adding the largest PROCESSES twice, and then finally adding 10 for each Oracle database on the system. To summarize: SEMMNS = sum of PROCESSES setting for each database on the system + ( 2 * [largest PROCESSES setting]) + (10 * [number of databases on system] To determine the maximum number of semaphores that can be allocated on a Linux system, use the following calculation. It will be the lesser of: SEMMNS -or- (SEMMSL * SEMMNI) SEMOPM The SEMOPM kernel parameter is used to control the number of semaphore operations that can be performed per semop system call. The semop system call (function) provides the ability to do operations for multiple semaphores with one semop system call. A semaphore set can have the maximum number of SEMMSL semaphores per semaphore set and is therefore recommended to set SEMOPM equal to SEMMSL. Oracle recommends setting the SEMOPM to a value of no less than 100. Setting Semaphore Kernel Parameters Finally, we see how to set all semaphore parameters using several methods. In the following, the only parameter I care about changing (raising) is SEMOPM. All other default settings should be sufficient for our example installation. This is method I use most often. This method sets all semaphore kernel parameters on startup by inserting the following kernel parameter in the /etc/sysctl.conf startup file: # echo "kernel.sem=250 32000 100 128" >> /etc/sysctl.conf If you wanted to dynamically alter the value of all semaphore kernel parameters without rebooting the machine, you can make this change directly to the /proc file system. This command can be made permanent by putting it into the /etc/rc.local startup file: # echo "250 32000 100 128" > /proc/sys/kernel/sem You can also use the sysctl command to change the value of all semaphore settings: # sysctl -w kernel.sem="250 32000 100 128" -------------------------------------------------------------------------------- Configuring File Handles When configuring our Linux database server, it is critical to ensure that the maximum number of file handles is large enough. The setting for file handles designate the number of open files that you can have on the entire Linux system. Use the following command to determine the maximum number of file handles for the entire system: # cat /proc/sys/fs/file-max 103062 Oracle recommends that the file handles for the entire system be set to at least 65536. In most cases, the default for Red Hat Linux is 103062. I have seen others (Red Hat Linux AS 2.1, Fedora Core 1, and Red Hat version 9) that will only default to 32768. If this is the case, you will want to increase this value to at least 65536. This is method I use most often. This method sets the maximum number of file handles (using the kernel parameter file-max) on startup by inserting the following kernel parameter in the /etc/sysctl.conf startup file: # echo "fs.file-max=65536" >> /etc/sysctl.conf If you wanted to dynamically alter the value of all semaphore kernel parameters without rebooting the machine, you can make this change directly to the /proc file system. This command can be made permanent by putting it into the /etc/rc.local startup file: # echo "65536" > /proc/sys/fs/file-max You can also use the sysctl command to change the maximum number of file handles: # sysctl -w fs.file-max=65536 NOTE: It is also possible to query the current usage of file handles using the following command: # cat /proc/sys/fs/file-nr 1140 0 103062 In the above example output, here is an explanation of the three values from the file-nr command: Total number of allocated file handles. Total number of file handles currently being used. Maximum number of file handles that can be allocated. This is essentially the value of file-max - (see above). NOTE: If you need to increase the value in /proc/sys/fs/file-max, then make sure that the ulimit is set properly. Usually for 2.4.20 it is set to unlimited. Verify the ulimit setting my issuing the ulimit command: # ulimit unlimited -------------------------------------------------------------------------------- Create Oracle Account and Directories Now let's create the Oracle UNIX account all all required directories: Login as the root user id. % su - Create directories. # mkdir -p /u01/app/oracle # mkdir -p /u03/app/oradata # mkdir -p /u04/app/oradata # mkdir -p /u05/app/oradata # mkdir -p /u06/app/oradata Create the UNIX Group for the Oracle User Id. # groupadd -g 115 dba Create the UNIX User for the Oracle Software. # useradd -u 173 -c "Oracle Software Owner" -d /u01/app/oracle -g "dba" -m -s /bin/bash oracle # passwd oracle Changing password for user oracle. New UNIX password: ************ BAD PASSWORD: it is based on a dictionary word Retype new UNIX password: ************ passwd: all authentication tokens updated successfully. Change ownership of all Oracle Directories to the Oracle UNIX User. # chown -R oracle:dba /u01 # chown -R oracle:dba /u03 # chown -R oracle:dba /u04 # chown -R oracle:dba /u05 # chown -R oracle:dba /u06 Oracle Environment Variable Settings NOTE: Ensure to set the environment variable: LD_ASSUME_KERNEL=2.4.1 Failing to set the LD_ASSUME_KERNEL parameter will cause the Oracle Universal Installer to hang! Verify all mount points. Please keep in mind that all of the following mount points can simply be directories if you only have one hard drive. For our installation, we will be using four mount points (or directories) as follows: /u01 : The Oracle RDBMS software will be installed to /u01/app/oracle. /u03 : This mount point will contain the physical Oracle files: Control File 1 Online Redo Log File - Group 1 / Member 1 Online Redo Log File - Group 2 / Member 1 Online Redo Log File - Group 3 / Member 1 /u04 : This mount point will contain the physical Oracle files: Control File 2 Online Redo Log File - Group 1 / Member 2 Online Redo Log File - Group 2 / Member 2 Online Redo Log File - Group 3 / Member 2 /u05 : This mount point will contain the physical Oracle files: Control File 3 Online Redo Log File - Group 1 / Member 3 Online Redo Log File - Group 2 / Member 3 Online Redo Log File - Group 3 / Member 3 /u06 : This mount point will contain the all physical Oracle data files. This will be one large RAID 0 stripe for all Oracle data files. All tablespaces including System, UNDO, Temporary, Data, and Index. -------------------------------------------------------------------------------- Configuring the Oracle Environment After configuring the Linux operating environment, it is time to setup the Oracle UNIX User ID for the installation of the Oracle RDBMS Software. Keep in mind that the following steps need to be performed by the oracle user id. Before delving into the details for configuring the Oracle User ID, I packaged an archive of shell scripts and configuration files to assist with the Oracle preparation and installation. You should download the archive "oracle_920_installation_files_linux.tar" as the Oracle User ID and place it in his HOME directory. Login as the oracle user id. % su - oracle Unpackage the contents of the oracle_920_installation_files_linux.tar archive. After extracting the archive, you will have a new directory called oracle_920_installation_files_linux that contains all required files. The following set of commands descibe how to extract the file and where to copy/extract all required files: $ id uid=173(oracle) gid=115(dba) groups=115(dba) $ pwd /u01/app/oracle $ tar xvf oracle_920_installation_files_linux.tar oracle_920_installation_files_linux/ oracle_920_installation_files_linux/admin.tar oracle_920_installation_files_linux/common.tar oracle_920_installation_files_linux/dbora oracle_920_installation_files_linux/dbshut oracle_920_installation_files_linux/.bash_profile oracle_920_installation_files_linux/dbstart oracle_920_installation_files_linux/ldap.ora oracle_920_installation_files_linux/listener.ora oracle_920_installation_files_linux/sqlnet.ora oracle_920_installation_files_linux/tnsnames.ora oracle_920_installation_files_linux/crontabORA920.txt $ cp oracle_920_installation_files_linux/.bash_profile ~/.bash_profile $ tar xvf oracle_920_installation_files_linux/admin.tar $ tar xvf oracle_920_installation_files_linux/common.tar $ . ~/.bash_profile .bash_profile executed $ -------------------------------------------------------------------------------- Configuring Oracle User Shell Limits Many of the Linux shells (including BASH) implement certain controls over certain critical resources like the number of file descriptors that can be opened and the maximum number of processes available to a user's session. In most cases, you will not need to alter any of these shell limits, but you find yourself getting errors when creating or maintaining the Oracle database, you may want to read through this section. You can use the following command to query these shell limits: # ulimit -a core file size (blocks, -c) 0 data seg size (kbytes, -d) unlimited file size (blocks, -f) unlimited max locked memory (kbytes, -l) unlimited max memory size (kbytes, -m) unlimited open files (-n) 1024 pipe size (512 bytes, -p) 8 stack size (kbytes, -s) 10240 cpu time (seconds, -t) unlimited max user processes (-u) 16383 virtual memory (kbytes, -v) unlimited Maximum Number of Open File Descriptors for Shell Session Let's first talk about the maximum number of open file descriptors for a user's shell session. NOTE: Make sure that throughout this section, that you are logged in as the oracle user account since this is the shell account we want to test! Ok, you are first going to tell me, "But I've already altered my Linux environment by setting the system wide kernel parameter /proc/sys/fs/file-max". Yes, this is correct, but there is still a per user limit on the number of open file descriptors. This typically defaults to 1024. To check that, use the following command: % su - oracle % ulimit -n 1024 If you wanted to change the maximum number of open file descriptors for a user's shell session, you could edit the /etc/security/limits.conf as the root account. For your Linux system, you would add the following lines: oracle soft nofile 4096 oracle hard nofile 101062 The first line above sets the soft limit, which is the number of files handles (or open files) that the Oracle user will have after logging in to the shell account. The hard limit defines the maximum number of file handles (or open files) are possible for the user's shell account. If the oracle user account starts to recieve error messages about running out of file handles, then number of file handles should be increased for the oracle using the user should increase the number of file handles using the hard limit setting. You can increase the value of this parameter to 101062 for the current session by using the following: % ulimit -n 101062 Keep in mind that the above command will only effect the current shell session. If you were to log out and log back in, the value would be set back to its default for that shell session. NOTE: Although you can set the soft and hard file limits higher, it is critical to understand to never set the hard limit for nofile for your shell account equal to /proc/sys/fs/file-max. If you were to do this, your shell session could use up all of the file descriptors for the entire Linux system, which means that the entire Linux system would run out of file descriptors. At this point, you would not be able to initiate any new logins since the system would not be able to open any PAM modules, which are required for login. Notice that I set my hard limit to 101062 and not 103062. In short, I am leaving 2000 spare! We're not totally done yet. We still need to ensure that pam_limits is configured in the /etc/pam.d/system-auth file. The steps defined below sould already be performed with a normal Red Hat Linux installation, but should still be validated! The PAM module will read the /etc/security/limits.conf file. You should have an entry in the /etc/pam.d/system-auth file as follows: session required /lib/security/$ISA/pam_limits.so I typically validate that my /etc/pam.d/system-auth file has the following two entries: session required /lib/security/$ISA/pam_limits.so session required /lib/security/$ISA/pam_unix.so Finally, let's test our new settings for the maximum number of open file descriptors for the oracle shell session. Logout and log back in as the oracle user account then run the following commands. Let's first check all current soft shell limits: $ ulimit -Sa core file size (blocks, -c) 0 data seg size (kbytes, -d) unlimited file size (blocks, -f) unlimited max locked memory (kbytes, -l) unlimited max memory size (kbytes, -m) unlimited open files (-n) 4096 pipe size (512 bytes, -p) 8 stack size (kbytes, -s) 10240 cpu time (seconds, -t) unlimited max user processes (-u) 16383 virtual memory (kbytes, -v) unlimited Finally, let's check all current hard shell limits: $ ulimit -Ha core file size (blocks, -c) unlimited data seg size (kbytes, -d) unlimited file size (blocks, -f) unlimited max locked memory (kbytes, -l) unlimited max memory size (kbytes, -m) unlimited open files (-n) 101062 pipe size (512 bytes, -p) 8 stack size (kbytes, -s) unlimited cpu time (seconds, -t) unlimited max user processes (-u) 16383 virtual memory (kbytes, -v) unlimited The soft limit is now set to 4096 while the hard limit is now set to 101062. NOTE: There may be times when you cannot get access to the root user account to change the /etc/security/limits.conf file. You can set this value in the user's login script for the shell as follows: su - oracle cat >> ~oracle/.bash_profile << EOF ulimit -n 101062 EOF NOTE: For this section, I used the BASH shell. The session values will not always be the same for other shells. Maximum Number of Processes for Shell Session This section is very similar to the previous section, "Maximum Number of Open File Descriptors for Shell Session" and deals with the same concept of soft limits and hard limits as well as configuring pam_limits. For most default Red Hat Linux installations, you will not need to be concerned with the maximum number of user processes as this value is generally high enough! NOTE: For this section, I used the BASH shell. The session values will not always be the same for other shells. Let's start by querying the current limit of the maximum number of processes for the oracle user: % su - oracle % ulimit -u 16383 If you wanted to change the soft and hard limits for the maximum number of processes for the oracle user, (and for that matter, all users), you could edit the /etc/security/limits.conf as the root account. For your Linux system, you would add the following lines: oracle soft nproc 2047 oracle hard nproc 16384 NOTE: There may be times when you cannot get access to the root user account to change the /etc/security/limits.conf file. You can set this value in the user's login script for the shell as follows: su - oracle cat >> ~oracle/.bash_profile << EOF ulimit -u 16384 EOF Miscellaneous Notes To check all current soft shell limits, enter the following command: $ ulimit -Sa core file size (blocks, -c) 0 data seg size (kbytes, -d) unlimited file size (blocks, -f) unlimited max locked memory (kbytes, -l) unlimited max memory size (kbytes, -m) unlimited open files (-n) 4096 pipe size (512 bytes, -p) 8 stack size (kbytes, -s) 10240 cpu time (seconds, -t) unlimited max user processes (-u) 16383 virtual memory (kbytes, -v) unlimited To check maximum hard limits, enter the following command: $ ulimit -Ha core file size (blocks, -c) unlimited data seg size (kbytes, -d) unlimited file size (blocks, -f) unlimited max locked memory (kbytes, -l) unlimited max memory size (kbytes, -m) unlimited open files (-n) 101062 pipe size (512 bytes, -p) 8 stack size (kbytes, -s) unlimited cpu time (seconds, -t) unlimited max user processes (-u) 16383 virtual memory (kbytes, -v) unlimited The file (blocks) value should be multiplied by 512 to obtain the maximum file size imposed by the shell. A value of unlimited is the operating system default and typically has a maximum value of 1 TB. NOTE: Oracle9i Release 2 (9.2.0) includes native support for files greater than 2 GB. Check your shell to determine whether it will impose a limit. -------------------------------------------------------------------------------- Downloading / Unpacking the Oracle9i Installation Files Most of the actions throughout the rest of this document should be done as the "oracle" user account unless otherwise noted. If you are not logged in as the "oracle" user account, do so now. Download Oracle9i from Oracle's OTN Site. (If you do not currently have an account with Oracle OTN, you will need to create one. This is a FREE account!) http://www.oracle.com/technology/software/products/oracle9i/htdocs/linuxsoft.html Download the following files to a temporary directory (i.e. /u01/app/oracle/orainstall: ship_9204_linux_disk1.cpio.gz (538,906,295 bytes) (cksum - 245082434) ship_9204_linux_disk2.cpio.gz (632,756,922 bytes) (cksum - 2575824107) ship_9204_linux_disk3.cpio.gz (296,127,243 bytes) (cksum - 96915247) Directions to extract the files. Run "gunzip " on all the files. % gunzip ship_9204_linux_disk1.cpio.gz Extract the cpio archives with the command: "cpio -idmv < " % cpio -idmv < ship_9204_linux_disk1.cpio NOTE: Some browsers will uncompress the files but leave the extension the same (gz) when downloading. If the above steps do not work for you, try skipping step 1 and go directly to step 2 without changing the filename. % cpio -idmv < ship_9204_linux_disk1.cpio.gz You should now have three directories called "Disk1, Disk2 and Disk3" containing the Oracle9i Installation files: /Disk1 /Disk2 /Disk3 -------------------------------------------------------------------------------- Update Red Hat Linux System - (Oracle Metalink Note: 252217.1) The following RPMs, all of which are available on the Red Hat Fedora Core 2 CDs, will need to be updated as per the steps described in Metalink Note: 252217.1 - "Requirements for Installing Oracle 9iR2 on RHEL3". All of these packages will need to be installed as the root user: From Fedora Core 2 / Disk #1 # cd /mnt/cdrom/Fedora/RPMS # rpm -Uvh libpng-1.2.2-22.i386.rpm From Fedora Core 2 / Disk #2 # cd /mnt/cdrom/Fedora/RPMS # rpm -Uvh gnome-libs-1.4.1.2.90-40.i386.rpm From Fedora Core 2 / Disk #3 # cd /mnt/cdrom/Fedora/RPMS # rpm -Uvh compat-libstdc++-7.3-2.96.126.i386.rpm # rpm -Uvh compat-libstdc++-devel-7.3-2.96.126.i386.rpm # rpm -Uvh compat-db-4.1.25-2.1.i386.rpm # rpm -Uvh compat-gcc-7.3-2.96.126.i386.rpm # rpm -Uvh compat-gcc-c++-7.3-2.96.126.i386.rpm # rpm -Uvh openmotif21-2.1.30-9.i386.rpm # rpm -Uvh pdksh-5.2.14-24.i386.rpm From Fedora Core 2 / Disk #4 # cd /mnt/cdrom/Fedora/RPMS # rpm -Uvh sysstat-5.0.1-2.i386.rpm Set gcc296 and g++296 in PATH Put gcc296 and g++296 first in $PATH variable by creating the following symbolic links: # mv /usr/bin/gcc /usr/bin/gcc323 # mv /usr/bin/g++ /usr/bin/g++323 # ln -s /usr/bin/gcc296 /usr/bin/gcc # ln -s /usr/bin/g++296 /usr/bin/g++ Check hostname Make sure the hostname command returns a fully qualified host name by amending the /etc/hosts file if necessary: # hostname Install the 3006854 patch: The Oracle / Linux Patch 3006854 can be downloaded here. # unzip p3006854_9204_LINUX.zip # cd 3006854 # sh rhel3_pre_install.sh -------------------------------------------------------------------------------- Install the Oracle 9.2.0.4.0 RDBMS Software As the "oracle" user account: Set your DISPLAY variable to a valid X Windows display. % DISPLAY=:0.0 % export DISPLAY NOTE: If you forgot to set the DISPLAY environment variable and you get the following error: Xlib: connection to ":0.0" refused by server Xlib: Client is not authorized to connect to Server you will then need to execute the following command to get "runInstaller" working again: % rm -rf /tmp/OraInstall If you don't do this, the Installer will hang without giving any error messages. Also make sure that "runInstaller" has stopped running in the background. If not, kill it. Change directory to the Oracle installation files you downloaded and extracted. Then run: runInstaller. $ su - oracle $ cd orainstall/Disk1 $ ./runInstaller Initializing Java Virtual Machine from /tmp/OraInstall2004-05-02_08-45-13PM/jre/bin/java. Please wait... Screen Name Response Welcome Screen: Click "Next" Inventory Location: Click "OK" UNIX Group Name: Use "dba" Root Script Window: Open another window, login as the root userid, and run "/tmp/orainstRoot.sh". When the script has completed, return to the dialog from the Oracle Installer and hit Continue. File Locations: Leave the "Source Path" at its default setting. For the Destination name, I like to use "OraHome920". You can leave the Destination path at it's default value which should be "/u01/app/oracle/product/9.2.0". Available Products: Select "Oracle9i Database 9.2.0.4.0" and click "Next" Installation Types: Select "Enterprise Edition (2.84GB)" and click "Next" Database Configuration: Select "Software Only" and click "Next" Summary: Click "Install" Running root.sh script. When the "Link" phase is complete, you will be prompted to run the $ORACLE_HOME/root.sh script as the "root" user account. Shutdown any started Oracle processes The Oracle Universal Installer will succeed in starting some Oracle programs, in particular the Oracle HTTP Server (Apache), the Oracle Intelligent Agent, and possibly the Orcle TNS Listener. Make sure all programs are shutdown before attempting to continue in installing the Oracle 9.2.0.5.0 patchset: % $ORACLE_HOME/Apache/Apache/bin/apachectl stop % agentctl stop % lsnrctl stop -------------------------------------------------------------------------------- Install the Oracle 9.2.0.5.0 Patchset Once you have completed installing of the Oracle9i (9.2.0.4.0) RDBMS software, you should now apply the 9.2.0.5.0 patchset. NOTE: The details and instructions for applying the 9.2.0.5.0 patchset in this article is not absolutely necessary. I provide it here simply as a convenience for those how do want to apply the latest patchset. The 9.2.0.5.0 patchset can be downloaded from Oracle Metalink: Patch Number: 3501955 Description: ORACLE 9i DATABASE SERVER RELEASE 2 - PATCH SET 4 VERSION 9.2.0.5.0 Product: Oracle Database Family Release: Oracle 9.2.0.5 Select a Platform or Language: Linux x86 Last Updated: 26-MAR-2004 Size: 313M (328923077 bytes) Use the following steps to install the Oracle10g Universal Installer and then the Oracle 9.2.0.5.0 patchset. To start, let's unpack the Oracle 9.2.0.5.0 to a temporary directory: % cd orapatch % unzip p3501955_9205_LINUX.zip % cpio -idmv < 9205_lnx32_release.cpio Next, we need to install the Oracle10g Universal Installer into the same $ORACLE_HOME we used to install the Oracle9i RDBMS software. NOTE: Using the old Universal Installer that was used to install the Oracle9i RDBMS software, (OUI release 2.2), cannot be used to install the 9.2.0.5.0 patchset and higher! Starting with the Oracle 9.2.0.5.0 patchset, Oracle requires the use of the Oracle10g Universal Installer to apply the 9.2.0.5.0 patchset and to perform all subsequent maintenance operations on the Oracle software $ORACLE_HOME. Let's get this thing started by installing the Oracle10g Universal Installer. This must be done by running the runInstaller that is included with the 9.2.0.5.0 patchset we extracted in the above step: % cd orapatch/Disk1 % ./runInstaller -ignoreSysPrereqs Starting Oracle Universal Installer... Checking installer requirements... Checking operating system version: must be redhat-2.1, UnitedLinux-1.0, redhat-3, SuSE-7 or SuSE-8 Failed <<<< >>> Ignoring required pre-requisite failures. Continuing... Preparing to launch Oracle Universal Installer from /tmp/OraInstall2004-08-30_07-48-15PM. Please wait ... Oracle Universal Installer, Version 10.1.0.2.0 Production Copyright (C) 1999, 2004, Oracle. All rights reserved. Use the following options in the Oracle Universal Installer to install the Oracle10g OUI: Screen Name Response Welcome Screen: Click "Next" File Locations: The "Source Path" should be pointing to the products.xml file by default. For the Destination name, choose the same one you created when installing the Oracle9i software. The name we used in this article was "OraHome920" and the destination path should be "/u01/app/oracle/product/9.2.0". Select a Product to Install: Select "Oracle Universal Installer 10.1.0.2.0" and click "Next" Summary: Click "Install" Exit from the Oracle Universal Installer. Correct the runInstaller symbolic link bug. (Bug 3560961) After the installation of Oracle10g Universal Installer, there is a bug that does NOT update the $ORACLE_HOME/bin/runInstaller symbolic link to point to the new 10g installation location. Since the symbolic link does not get updated, the runInstaller command still points to the old installer (2.2) and will be run instead of the new 10g installer. To correct this, you will need to manually update the $ORACLE_HOME/bin/runInstaller symbolic link: % cd $ORACLE_HOME/bin % ln -s -f $ORACLE_HOME/oui/bin/runInstaller.sh runInstaller We now install the Oracle 9.2.0.5.0 patchset by executing the newly installed 10g Universal Installer: % cd % runInstaller -ignoreSysPrereqs Starting Oracle Universal Installer... Checking installer requirements... Checking operating system version: must be redhat-2.1, UnitedLinux-1.0, redhat-3, SuSE-7 or SuSE-8 Failed <<<< >>> Ignoring required pre-requisite failures. Continuing... Preparing to launch Oracle Universal Installer from /tmp/OraInstall2004-08-30_07-59-30PM. Please wait ... Oracle Universal Installer, Version 10.1.0.2.0 Production Copyright (C) 1999, 2004, Oracle. All rights reserved. Here is an overview of the selections I made while performing the 9.2.0.5.0 patchset install: Screen Name Response Welcome Screen: Click "Next" File Locations: The "Source Path" should be pointing to the products.xml file by default. For the Destination name, choose the same one you created when installing the Oracle9i software. The name we used in this article was "OraHome920" and the destination path should be "/u01/app/oracle/product/9.2.0". Select a Product to Install: Select "Oracle 9iR2 Patchsets 9.2.0.5.0" and click "Next" Summary: Click "Install" Running root.sh script. When the Link phase is complete, you will be prompted to run the $ORACLE_HOME/root.sh script as the "root" user account. Go ahead and run the root.sh script. Exit Universal Installer Exit from the Universal Installer and continue on to the Post Installation section of this article. -------------------------------------------------------------------------------- Post Installation Steps After applying the Oracle 9.2.0.5.0 patchset, we should perform several miscellaneous tasks like configuring the Oracle Networking files and setting up startup and shutdown scripts for then the machine is cycled. Configuring Oracle Networking Files: I already included sample configuration files (contained in the oracle_920_installation_files_linux.tar file) that can be simply copied to their proper location and started. Change to the oracle HOME directory and copy the files as follows: % cd % cd oracle_920_installation_files_linux % cp ldap.ora $ORACLE_HOME/network/admin/ % cp tnsnames.ora $ORACLE_HOME/network/admin/ % cp sqlnet.ora $ORACLE_HOME/network/admin/ % cp listener.ora $ORACLE_HOME/network/admin/ % cd % lsnrctl start Update /etc/oratab: The dbora script (below) relies on an entry in the /etc/oratab. Perform the following actions as the oracle user account: % echo "ORA920:/u01/app/oracle/product/9.2.0:Y" >> /etc/oratab Configuring Startup / Shutdown Scripts: Also included in the oracle_920_installation_files_linux.tar file is a script called dbora. This script can be used by the init process to startup and shutdown the database when the machine is cycled. The following tasks will need to be performed by the root user account: % su - # cp /u01/app/oracle/oracle_920_installation_files_linux/dbora /etc/init.d # chmod 755 /etc/init.d/dbora # ln -s /etc/init.d/dbora /etc/rc3.d/S99dbora # ln -s /etc/init.d/dbora /etc/rc4.d/S99dbora # ln -s /etc/init.d/dbora /etc/rc5.d/S99dbora # ln -s /etc/init.d/dbora /etc/rc0.d/K10dbora # ln -s /etc/init.d/dbora /etc/rc6.d/K10dbora -------------------------------------------------------------------------------- Creating the Oracle Database Finally, let's create an Oracle9i database. This can be done using scripts that I already included with the oracle_920_installation_files_linux.tar download. The scripts are included in the ~oracle/admin/ORA920/create directory. To create the database, perform the following steps: % su - oracle % cd admin/ORA920/create % ./RUN_CRDB.sh After starting the RUN_CRDB.sh, there will be no screen activity until the database creation is complete. You can, however, bring up a new console window to the Linux databse server as the oracle user account, navigate to the same directory you started the database creation from, and tail the crdb.log log file. $ telnet linux3 ... Fedora Core release 2 (Tettnang) Kernel 2.6.5-1.358 on an i686 login: oracle Password: xxxxxx .bash_profile executed [oracle@linux3 oracle]$ cd admin/ORA920/create [oracle@linux3 create]$ tail -f crdb.log ===================================== 8. Install Oracle 9.2.0.2 on OpenVMS: ===================================== VMS: ==== Using OUI to install Oracle9i Release 2 on an OpenVMS System We have a PC running Xcursion and a 16 Processor GS1280 with the 2 built-in disks In the examples we booted on disk DKA0: Oracle account is on disk DKA100. Oracle and the database will be installed on DKA100. Install disk MUST be ODS-5. Installation uses the 9.2 downloaded from the Oracle website. It comes in a Java JAR file. Oracle ships a JRE with its product. However, you will have to install Java on OpenVMS so you can unpack the 9.2 JAR file that comes from the Oracle website Unpack the JAR file as described on the Oracle website. This will create two .BCK files. Follow the instructions in the VMS_9202_README.txt file on how to restore the 2 backup save sets. When the two backup save sets files are restored, you should end up with two directories: [disk1] directory [disk2] directory These directories will be in the root of a disk. In this example they are in the root of DKA100. The OUI requires X-Windows. If the Alpha system you are using does not have a graphic head, use a PC with an X-Windows terminal such as Xcursion. During this install we discovered a problem: Instructions tell you to run @DKA100:[disk1]runinstaller. This will not work because the RUNINSTALLER.COM file is not in the root of DKA100:[disk1]. You must first copy RUNINSTALLER.COM from the dka100:[disk1.000000] directory into dka100:[disk1]: $ Copy dka100:[disk1.000000]runinstaller.com dka100:[disk1] From a terminal window execute: @DKA100:[disk1]runinstaller - Oracle Installer starts Start the installation Click Next to start the installation. - Assign name and directory structure for the Oracle Home ORACLE_HOME Assign a name for your Oracle home. Assign the directory structure for the home, for example Ora_home Dka100:[oracle.oracle9] This is where the OUI will install Oracle. The OUI will create the directories as necessary - Select product to install Select Database. Click Next. - Select type of installation Select Enterprise Edition (or Standard Edition or Custom). Click Next. - Enable RAC Select No. Click Next. - Database summary View list of products that will be installed. Click Install. - Installation begins Installation takes from 45 minutes to an hour. Installation ends Click Exit. Oracle is now installed in DKA100:[oracle.oracle9]. To create the first database, you must first set up Oracle logicals. To do this use a terminal and execute @[.oracle9]orauser . The tool to create and manage databases is DBCA. On the terminal, type DBCA to launch the Database Assistant. Welcome to Database Configuration Assistant DBCA starts. Click Next. Select an operation Select Create a Database. Click Next. Select a template Select New Database. Click Next. Enter database name and SID Enter the name of the database and Oracle System Identifier (SID): In this example, the database name is DB9I. The SID is DB9I1. Click Next. Select database features Select which demo databases are installed. In the example, we selected all possible databases. Click Next. Select default node Select the node in which you want your database to operate by default. In the example, we selected Shared Server Mode. Click Next. Select memory In the example, we selected the default. Click Next. Specify database storage parameters Select the device and directory. Use the UNIX device syntax I.E. For example, DKA100:[oracle.oracle9.database] would be: /DKA100/oracle/oracle9/database/ In the example, we kept the default settings. Click Next. Select database creation options Creating a template saves time when creating a database. Click Finish. Create a template Click OK. Creating and starting Oracle Instance The database builds. If it completes successfully, click Exit. If it does not complete successfully, build it again. Running the database Enter “show system” to see the Oracle database up and running. Set up some files to start and stop the database. Example of a start file This command sets the logicals to manage the database: $ @dka100:[oracle.oracle9]orauser db9i1 The next line starts the Listener (needed for client connects). The final lines start the database. Stop database example Example of how to stop the database. Test database server Use the Enterprise Manager console to test the database server. Oracle Enterprise Manager Enter address of server and SID. Name the server. Click OK. Databases connect information Select database. Enter system account and password. Change connection box to “AS SYSDBA.” Click OK. Open database Database is opened and exposed. Listener Listener automatically picks up the SID from the database. Start Listener before database and the SID will display in the Listener. If you start the database before the Listener, the SID may not appear immediately. To see if the SID is registered in the Listener, enter: $lsnrctl stat Alter a user User is altered: SQL> alter user oe identified by oe account unlock; SQL> exit Preferred method is to use the Enterprise Manager Console. ================================================== 9. Installation of Oracle 9i on AIX and other UNIX ================================================== AIX: ==== 9.1 Installation of Oracle 9i on AIX Doc ID: Note:201019.1 Content Type: TEXT/PLAIN Subject: AIX: Quick Start Guide - 9.2.0 RDBMS Installation Creation Date: 25-JUN-2002 Type: REFERENCE Last Revision Date: 14-APR-2004 Status: PUBLISHED Quick Start Guide Oracle9i Release 2 (9.2.0) RDBMS Installation AIX Operating System Purpose ======= This document is designed to be a quick reference that can be used when installing Oracle9i Release 2 (9.2.0) on an AIX platform. It is NOT designed to replace the Installation Guide or other documentation. A familiarity with the AIX Operating System is assumed. If more detailed information is needed, please see the Appendix at the bottom of this document for additional resources. Each step should be done in the order that it is listed. These steps are the bare minimum that is necessary for a typical install of the Oracle9i RDBMS. Verify OS version is certified with the RDBMS version ====================================================== The following steps are required to verify your version of the AIX operating system is certified with the version of the RDBMS (Oracle9i Release 2 (9.2.0)): 1. Point your web browser to http://metalink.oracle.com. 2. Click the "Certify & Availability" button near the left. 3. Click the "Certifications" button near the top middle. 4. Click the "View Certifications by Platform" link. 5. Select "IBM RS/6000 AIX" and click "Submit". 6. Select Product Group "Oracle Server" and click "Submit". 7. Select Product "Oracle Server - Enterprise Edition" and click "Submit". 8. Read any general notes at the top of the page. 9. Select "9.2 (9i) 64-bit" and click "Submit". The "Status" column displays the certification status. The links in the "Addt'l Info" and "Install Issue" columns may contain additional information relevant to a given version. Note that if patches are listed under one of these links, your installation is not considered certified unless you apply them. The "Addt'l Info" link also contains information about available patchsets. Installation of patchsets is not required to be considered certified, but they are highly recommended. Pre-Installation Steps for the System Administrator ==================================================== The following steps are required to verify your operating system meets minimum requirements for installation, and should be performed by the root user. For assistance with system administration issues, please contact your system administator or operating system vendor. Use these steps to manually check the operating system requirements before attempting to install Oracle RDBMS software, or you may choose to use the convenient "Unix InstallPrep script" which automates these checks for you. For more information about the script, including download information, please review the following article: Note:189256.1 UNIX: Script to Verify Installation Requirements for Oracle 9.x version of RDBMS The InstallPrep script currently does not check requirements for AIX5L systems. The Following Steps Need to be Performed by the Root User: 1. Configure Operating System Resources: Ensure that the system has at least the following resources: ? 400 MB in /tmp * ? 256 MB of physical RAM memory ? Two times the amount of physical RAM memory for Swap/Paging space (On systems with more than 2 GB of physical RAM memory, the requirements for Swap/Paging space can be lowered, but Swap/Paging space should never be less than physical RAM memory.) * You may also redirect /tmp by setting the TEMP environment variable. This is only recommended in rare circumstances where /tmp cannot be expanded to meet free space requirements. 2. Create an Oracle Software Owner and Group: Create an AIX user and group that will own the Oracle software. (user = oracle, group = dba) ? Use the "smit security" command to create a new group and user Please ensure that the user and group you use are defined in the local /etc/passwd (user) and /etc/group (group) files rather than resolved via a network service such as NIS. 3. Create a Software Mount Point and Datafile Mount Points: Create a mount point for the Oracle software installation. (at least 3.5 GB, typically /u01) Create a second, third, and fourth mount point for the database files. (typically /u02, /u03, and /u04) Use of multiple mount points is not required, but is highly recommended for best performance and ease of recoverability. 4. Ensure that Asynchronous Input Output (AIO) is "Available": Use the following command to check the current AIO status: # lsdev -Cc aio Verify that the status shown is "Available". If the status shown is "Defined", then change the "STATE to be configured at system restart" to "Available" after running the following command: # smit chaio 5. Ensure that the math library is installed on your system: Use the following command to determine if the math library is installed: # lslpp -l bos.adt.libm If this fileset is not installed and "COMMITTED", then you must install it from the AIX operating system CD-ROM from IBM. With the correct CD-ROM mounted, run the following command to begin the process to load the required bos.adt.libm fileset: # smit install_latest AIX5L systems also require the following filesets: # lslpp -l bos.perf.perfstat # lslpp -l bos.perf.libperfstat 6. Download and install JDK 1.3.1 from IBM. At the time this article was created, the JDK could be downloaded from the following URL: http://www.ibm.com/developerworks/java/jdk/aix/index.html Please contact IBM Support if you need assistance downloading or installing the JDK. 7. Mount the Oracle CD-ROM: Mount the Oracle9i Release 2 (9.2.0) CD-ROM using the command: # mount -rv cdrfs /dev/cd0 /cdrom 8. Run the rootpre.sh script: NOTE: You must shutdown ALL Oracle database instances (if any) before running the rootpre.sh script. Do not run the rootpre.sh script if you have a newer version of an Oracle database already installed on this system. Use the following command to run the rootpre.sh script: # /cdrom/rootpre.sh Installation Steps for the Oracle User ======================================= The Following Steps Need to be Performed by the Oracle User: 1. Set Environment Variables Environment variables should be set in the login script for the oracle user. If the oracle user's default shell is the C-shell (/usr/bin/csh), then the login script will be named ".login". If the oracle user's default shell is the Bourne-shell (/usr/bin/bsh) or the Korn-shell (/usr/bin/sh or /usr/bin/ksh), then the login script will be named ".profile". In either case, the login script will be located in the oracle user's home directory ($HOME). The examples below assume that your software mount point is /u01. Parameter Value ----------- ----------------------------- ORACLE_HOME /u01/app/oracle/product/9.2.0 PATH /u01/app/oracle/product/9.2.0/bin:/usr/ccs/bin: /usr/bin/X11: (followed by any other directories you wish to include) ORACLE_SID Set this to what you will call your database instance. (typically 4 characters in length) DISPLAY :0.0 (review Note:153960.1 for detailed information) 2. Set the umask: Set the oracle user's umask to "022" in you ".profile" or ".login" file. Example: umask 022 3. Verify the Environment Log off and log on as the oracle user to ensure all environment variables are set correctly. Use the following command to view them: % env | more Before attempting to run the Oracle Universal Installer (OUI), verify that you can successfully run the following command: % /usr/bin/X11/xclock If this does not display a clock on your display screen, please review the following article: Note:153960.1 FAQ: X Server testing and troubleshooting 4. Start the Oracle Universal Installer and install the RDBMS software: Use the following commands to start the installer: % cd /tmp % /cdrom/runInstaller Respond to the installer prompts as shown below: ? When prompted for whether rootpre.sh has been run by root, enter "y". This should have been done in Pre-Installation step 8 above. ? At the "Welcome Screen", click Next. ? If prompted, enter the directory to use for the "Inventory Location". This can be any directory, but is usually not under ORACLE_HOME because the oraInventory is shared with all Oracle products on the system. ? If prompted, enter the "UNIX Group Name" for the oracle user (dba). ? At the "File Locations Screen", verify the Destination listed is your ORACLE_HOME directory. Also enter a NAME to identify this ORACLE_HOME. The NAME can be anything, but is typically "DataServer" and the first three digits of the version. For example: "DataServer920" ? At the "Available Products Screen", choose Oracle9i Database, then click Next. ? At the "Installation Types Screen", choose Enterprise Edition, then click Next. ? If prompted, click Next at the "Component Locations Screen" to accept the default directories. ? At the "Database Configuration Screen", choose the the configuration based on how you plan to use the database, then click Next. ? If prompted, click Next at the "Privileged Operating System Groups Screen" to accept the default values (your current OS primary group). ? If prompted, enter the Global Database Name in the format "ORACLE_SID.hostname" at the "Database Identification Screen". For example: "TEST.AIXhost". The SID entry should be filled in with the value of ORACLE_SID. Click Next. ? If prompted, enter the directory where you would like to put datafiles at the "Database File Location Screen". Click Next. ? If prompted, select "Use the default character set" (WE8ISO8859P1) at the "Database Character Set Screen". Click Next. ? At the "Choose JDK Home Directory", enter the directory where you have previously installed the JDK 1.3.1 from IBM. This should have been done in Pre-Installation step 6 above. ? At the "Summary Screen", review your choices, then click Install. The install will begin. Follow instructions regarding running "root.sh" and any other prompts. When completed, the install will have created a default database, configured a Listener, and started both for you. Note: If you are having problems changing CD-ROMs when prompted to do so, please review the following article: Note:146566.1 How to Unmount / Eject First Cdrom Your Oracle9i Release 2 (9.2.0) RDBMS installation is now complete and ready for use. Appendix A ========== Documentation is available from the following resources: Oracle9i Release 2 (9.2.0) CD-ROM Disk1 ---------------------------------------- Mount the CD-ROM, then use a web browser to open the file "index.htm" located at the top level directory of the CD-ROM. On this CD-ROM you will find the Installation Guide, Administrator's Reference, and other useful documentation. Oracle Documentation Center --------------------------- Point your web browser to the following URL: http://otn.oracle.com/documentation/content.html Select the highest version CD-pack displayed to ensure you get the most up-to-date information. Unattended install: ------------------- Note 1: ------- This note describes how to start the unattended install of patch 9.2.0.5 on AIX 5L, which can be applied to 9.2.0.2, 9.2.0.3, 9.2.0.4 Shut down the existing Oracle server instance with normal or immediate priority. For example, shutdown all instances (cleanly) if running Parallel Server. Stop all listener, agent and other processes running in or against the ORACLE_HOME that will have the patch set installation. Run slibclean (/usr/sbin/slibclean) as root to remove ant currently unused modules in kernel and library memory. To perform a silent installation requiring no user intervention: Copy the response file template provided in the response directory where you unpacked the patch set tar file. Edit the values for all fields labeled as according to the comments and examples in the template. Start the Oracle Universal Installer from the directory described in Step 4 which applies to your situation. You should pass the full path of the response file template you have edited locally as the last argument with your own value of ORACLE_HOME and FROM_LOCATION. The following is an example of the command: % ./runInstaller -silent -responseFile full_path_to_your_response_file Run the $ORACLE_HOME/root.sh script from a root session. If you are applying the patch set in a cluster database environment, the root.sh script should be run in the same way on both the local node and all participating nodes. Note 2: ------- In order to make an unattended install of 9.2.0.1 on Win2K: Running Oracle Universal Installer and Specifying a Response File To run Oracle Universal Installer and specify the response file: Go to the MS-DOS command prompt. Go to the directory where Oracle Universal Installer is installed. Run the appropriate response file. For example, C:\program files\oracle\oui\install> setup.exe -silent -nowelcome -responseFile filename Where... Description filename Identifies the full path of the specific response file -silent Runs Oracle Universal Installer in complete silent mode. The Welcome window is suppressed automatically. This parameter is optional. If you use -silent, -nowelcome is not necessary. -nowelcome Suppresses the Welcome window that appears during installation. This parameter is optional. Note 3: ------- Unattended install of 9.2.0.5 on Win2K: To perform a silent installation requiring no user intervention: Make a copy of the response file template provided in the response directory where you unzipped the patch set file. Edit the values for all fields labeled as according to the comments and examples in the template. Start Oracle Universal Installer release 10.1.0.2 located in the unzipped area of the patch set. For example, Disk1\setup.exe. You should pass the full path of the response file template you have edited locally as the last argument with your own value of ORACLE_HOME and FROM_LOCATION. The syntax is as follows: setup.exe -silent -responseFile ORACLE_BASE\ORACLE_HOME\response_file_path =============================== 9.2 Oracle and UNIX and other OS: =============================== You have the following options for creating your new Oracle database: - Use the Database Configuration Assistant (DBCA). DBCA can be launched by the Oracle Universal Installer, depending upon the type of install that you select, and provides a graphical user interface (GUI) that guides you through the creation of a database. You can chose not to use DBCA, or you can launch it as a standalone tool at any time in the future to create a database. Run DCBA as % dbca - Create the database manually from a script. If you already have existing scripts for creating your database, you can still create your database manually. However, consider editing your existing script to take advantage of new Oracle features. Oracle provides a sample database creation script and a sample initialization parameter file with the database software files it distributes, both of which can be edited to suit your needs. - Upgrade an existing database. In all cases, the Oracle software needs to be installed on your host machine. 9.1.1 Operating system dependencies: ------------------------------------ First, determine for this version of Oracle, what OS settings must be made, and if any patches must be installed. For example, on Linux, glibc 2.1.3 is needed with Oracle version 8.1.7. Linux could be quite critical with respect to libraries in combination with Oracle. Ook moet er mogelijk shmmax (max size of shared memory segment) en dergelijke kernel parameters worden aangepast. # sysctl -w kernel.shmmax=100000000 # echo "kernel.shmmax = 100000000" >> /etc/sysctl.conf Opmerking: Het onderstaANDe is algemeen, maar is ook afgeleid van een Oracle 8.1.7 installatie op Linux Redhat 6.2 Als de 8.1.7 installatie gedaan wordt is ook nog de Java JDK 1.1.8 nodig. Deze kan gedownload worden van www.blackdown.org Download jdk-1.1.8_v3 jdk118_v3-glibc-2.1.3.tar.bz2 in /usr/local tar xvif jdk118_v3-glibc-2.1.3.tar.bz2 ln -s /usr/local/jdk118_v3 /usr/local/java 9.1.2 Environment variables: ---------------------------- Make sure you have the following environment variables set: ON UNIX: ======== Example 1: ---------- ORACLE_BASE=/u01/app/oracle; export ORACLE_BASE (root voor oracle software) ORACLE_HOME=$ORACLE_BASE/product/8.1.5; export ORACLE_HOME (bepaald de directory waarin de instance software zich bevind) ORACLE_SID=brdb; export ORACLE_SID (bepaald de naam van de huidige instance) ORACLE_TERM=xterm, vt100, ansi of wat ANDers; export ORACLE_TERM ORA_NLSxx=$ORACLE_HOME/ocommon/nls/admin/data; export ORA_NLS (bepaald de nls directory t.b.v. datafiles voor meerdere talen) NLS_LANG="Dutch_The NetherlANDs.WE8ISO8859P1"; export NLS_LANG (Dit specificeert de language, territory en characterset t.b.v de client applicaties. LD_LIBRARY_PATH=/u01/app/oracle/product/8.1.7/lib; export LD_LIBRARY_PATH PATH=$ORACLE_HOME/bin:/bin:/user/bin:/usr/sbin:/bin; export PATH plaats deze variabelen in de oracle user profile file: .profile, of .bash_profile etc.. Example 2: ---------- /dbs01 - - - - - Db directory 1 /dbs01 /app - - - - Constante /dbs01 /app /oracle - - $ORACLE_BASE Oracle base directory /dbs01 /app /oracle /admin - $ORACLE_ADMIN Oracle admin directory /dbs01 /app /oracle /product - - Constante /dbs01 /app /oracle /product /817 $ORACLE_HOME Oracle home directory # LISTENER.ORA Network Configuration File: /dbs01/app/oracle/product/817/network/admin/listener.ora # TNSNAMES.ORA Network Configuration File: /dbs01/app/oracle/product/817/network/admin/tnsnames.ora Example 3: ---------- /dbs01/app/orace Oracle software /dbs02/oradata database files /dbs03/oradata database files .. .. /var/opt/oracle network files /opt/oracle/admin/bin Example 4: ---------- Mountpunt Device Omvang (Mbyte) Doel / /dev/md/dsk/d1 100 Unix Root-filesysteem /usr /dev/md/dsk/d3 1200 Unix usr-filesysteem /var /dev/md/dsk/d4 200 Unix var-filesysteem /home /dev/md/dsk/d5 200 Unix opt-filesysteem /opt /dev/md/dsk/d6 4700 Oracle_Home /u01 /dev/md/dsk/d7 8700 Oracle datafiles /u02 /dev/md/dsk/d8 8700 Oracle datafiles /u03 /dev/md/dsk/d9 8700 Oracle datafiles /u04 /dev/md/dsk/d10 8700 Oracle datafiles /u05 /dev/md/dsk/d110 8700 Oracle datafiles /u06 /dev/md/dsk/d120 8700 Oracle datafiles /u07 /dev/md/dsk/d123 8650 Oracle datafiles Example 5: ---------- initBENE.ora /opt/oracle/product/8.0.6/dbs tnsnames.ora /opt/oracle/product/8.0.6/network/admin listener.ora /opt/oracle/product/8.0.6/network/admin alert log /var/opt/oracle/bene/bdump oratab /var/opt/oracle Example 6: ---------- ORACLE_BASE /u01/app/oracle ORACLE_HOME $ORACLE_BASE/product/10.1.0/db_1 ORACLE_PATH /u01/app/oracle/product/10.1.0/db_1/bin:. Note: The period adds the current working directory to the search path. ORACLE_SID SAL1 ORAENV_ASK NO SQLPATH /home:/home/oracle:/u01/oracle TNS_ADMIN $ORACLE_HOME/network/admin TWO_TASK Function Specifies the default connect identifier to use in the connect string. If this environment variable is set, you do not need to specify the connect identifier in the connect string. For example, if the TWO_TASK environment variable is set to sales, you can connect to a database using the CONNECT username/password command rather than the CONNECT username/password@sales command. Syntax Any connect identifier. Example PRODDB_TCP to identify the SID and Oracle home directory for the instance that you want to shut down, enter the following command: Solaris: $ cat /var/opt/oracle/oratab Other operating systems: $ cat /etc/oratab ON NT/2000: =========== SET ORACLE_BASE=G:\ORACLE SET ORACLE_HOME=G:\ORACLE\ORA81 SET ORACLE_SID=AIRM SET ORA_NLSxxx=G:\ORACLE\ORA81\ocommon\nls\admin\data SET NLS_LANG=AMERICAN_AMERICA.WE8ISO8859P1 ON OpenVMS: =========== When Oracle is installed on VMS, a root directory is chosen which is pointed to by the logical name ORA_ROOT. This directory can be placed anywhere on the VMS system. The majority of code, configuration files and command procedures are found below this root directory. When a new database is created a new directory is created in the root directory to store database specific configuration files. This directory is called [.DB_dbname]. This directory will normally hold the system tablespace data file as well as the database specific startup, shutdown and orauser files. The Oracle environment for a VMS user is set up by running the appropriate ORAUSER_dbname.COM file. This sets up the necessary command symbols and logical names to access the various ORACLE utilities. Each database created on a VMS system will have an ORAUSER file in it's home directory and will be named ORAUSER_dbname.COM, e.g. for a database SALES the file specification could be: ORA_ROOT:[DB_SALES]ORAUSER_SALES.COM To have the environment set up automatically on login, run this command file in your login.com file. To access SQLPLUS use the following command with a valid username and password: $ SQLPLUS username/password SQLDBA is also available on VMS and can be invoked similarly: $ SQLDBA username/password 9.1.3 OFA directory structuur: ------------------------------ Hou je aan OFA. Een voorbeeld voor database PROD: /opt/oracle/product/8.1.6 /opt/oracle/product/8.1.6/admin/PROD /opt/oracle/product/8.1.6/admin/pfile /opt/oracle/product/8.1.6/admin/adhoc /opt/oracle/product/8.1.6/admin/bdump /opt/oracle/product/8.1.6/admin/udump /opt/oracle/product/8.1.6/admin/adump /opt/oracle/product/8.1.6/admin/cdump /opt/oracle/product/8.1.6/admin/create /u02/oradata/PROD /u03/oradata/PROD /u04/oradata/PROD etc.. Example mountpoints and disks: ------------------------------ Mountpunt Device Omvang Doel / /dev/md/dsk/d1 100 Unix Root-filesysteem /usr /dev/md/dsk/d3 1200 Unix usr-filesysteem /var /dev/md/dsk/d4 200 Unix var-filesysteem /home /dev/md/dsk/d5 200 Unix opt-filesysteem /opt /dev/md/dsk/d6 4700 Oracle_Home /u01 /dev/md/dsk/d7 8700 Oracle datafiles /u02 /dev/md/dsk/d8 8700 Oracle datafiles /u03 /dev/md/dsk/d9 8700 Oracle datafiles /u04 /dev/md/dsk/d10 8700 Oracle datafiles /u05 /dev/md/dsk/d110 8700 Oracle datafiles /u06 /dev/md/dsk/d120 8700 Oracle datafiles /u07 /dev/md/dsk/d123 8650 Oracle datafiles 9.1.4 Users en groups: ---------------------- Als je met OS verificatie wilt werken, moet in de init.ora gezet zijn: remote_login_passwordfile=none (passwordfile authentication via exlusive) Benodigde groups in UNIX: group dba. Deze moet voorkomen in de /etc/group file vaak is ook nog nodig de group oinstall groupadd dba groupadd oinstall groupadd oper Maak nu user oracle aan: adduser -g oinstall -G dba -d /home/oracle oracle # groupadd dba # useradd oracle # mkdir /usr/oracle # mkdir /usr/oracle/9.0 # chown -R oracle:dba /usr/oracle # touch /etc/oratab # chown oracle:dba /etc/oratab 9.1.5 mount points en disks: ---------------------------- maak de mount points: mkdir /opt/u01 mkdir /opt/u02 mkdir /opt/u03 mkdir /opt/u04 dit moeten voor een produktie omgeving aparte schijven zijn Geef nu ownership van deze mount points aan user oracle en group oinstall chown -R oracle:oinstall /opt/u01 chown -R oracle:oinstall /opt/u02 chown -R oracle:oinstall /opt/u03 chown -R oracle:oinstall /opt/u04 directories: drwxr-xr-x oracle dba files : -rw-r----- oracle dba : -rw-r--r-- oracle dba chmod 644 * chmod u+x filename chmod ug+x filename 9.1.6 test van user oracle: --------------------------- log in als user oracle en geef de commANDo's $groups laat de groups zien (oinstall, dba) $umask laat 022 zien, zoniet zet dan de line umask 022 in het .profile umask is de default mode van een file of directory wanneer deze aangemaakt wordt. rwxrwxrwx=777 rw-rw-rw-=666 rw-r--r--=644 welke correspondeert met umask 022 Verander nu het .profile of .bash_profile van de user oracle. Plaats de environment variabelen van 9.1 in het profile. log uit en in als user oracle, en test de environment: %env %echo $variablename 9.1.7 Oracle Installer bij 8.1.x op Linux: ------------------------------------------ Log in als user oracle. Draai nu oracle installer: Linux: startx cd /usr/local/src/Oracle8iR3 ./runInstaller of Ga naar install/linux op de CD en run runIns.sh Nu volgt een grafische setup. Beantwoord de vragen. Het kan zijn dat de installer vraagt om scripts uit te voeren zoals: orainstRoot.sh en root.sh Om dit uit te voeren: open een nieuw window su root cd $ORACLE_HOME ./orainstRoot.sh Installatie database op Unix: ----------------------------- $ export PATH=$PATH:$ORACLE_HOME/bin $ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$ORACLE_HOME/lib $ dbca & or $ cat "db1:/usr/oracle/9.0:Y >> /etc/oratab" $ cd $ORACLE_HOME/dbs $ cat initdw.ora |sed s/"#db_name = MY_DB_NAME"/"db_name = db1"/|sed s/#control_files/control_files/ > initdb1.ora Start and create database : $ export PATH=$PATH:$ORACLE_HOME/bin $ export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$ORACLE_HOME/lib $ export ORACLE_SID=db1 $ sqlplus /nolog < apply or capture processes) Each capture process and apply process may use multiple parallel execution servers. The apply process by default needs two parallel servers. So this parameter needs to set to at least 2 even for a single non-parallel apply process. Specify a value for this parameter to ensure that there are enough parallel execution servers. In our installation we went for 12 apply server, so we increased the number of parallel_max_server above this figure of 12. _kghdsidx_count=1 This parameter prevents the shared_pool from being divided among CPUs LOG_PARALLELISM=1 This parameter must be set to 1 at each database that captures events. Parameters set using DBMS_CAPTURE_ADM package: Using the DBMS_CAPTURADM.SET_PARAMETER procedure there a 3 a parameters that are of common usage to affect installation PARALLELISM=3 There may be only one logminer session for the whole ruleset and only one enqueuer process that will push the objects. you can safely define as much as 3 execution capture process per CPU _CHECKPOINT_FREQUENCY=1 Increase the frequency of logminer checkpoints especially in a database with significant LOB or DDL activity. A logminer checkpoint is requested by default every 10Mb of redo mined. _SGA_SIZE Amount of memory available from the shared pool for logminer processing. The default amount of shared_pool memory allocated to logminer is 10Mb. Increase this value especially in environments where large LOBs are processed. 9.11. Older Database installations: ----------------------------------- CREATE DATABASE Examples on 8.x The easiest way to create a 8i, 9i database, is using the "Database Configuration Assistant". Using this tool, you are able to create a database and setup the NET configuration and the listener, in a graphical environment. It is also possible to use a script running in sqlpus (8i,9i) or svrmgrl (only in 8i). Charactersets that are used a lot in europe: WE8ISO8859P15 WE8MMSWIN1252 Example 1: ---------- $ SQLPLUS /nolog CONNECT username/password AS sysdba STARTUP NOMOUMT PFILE= -- Create database CREATE DATABASE rbdb1 CONTROLFILE REUSE LOGFILE '/u01/oracle/rbdb1/redo01.log' SIZE 1M REUSE, '/u01/oracle/rbdb1/redo02.log' SIZE 1M REUSE, '/u01/oracle/rbdb1/redo03.log' SIZE 1M REUSE, '/u01/oracle/rbdb1/redo04.log' SIZE 1M REUSE DATAFILE '/u01/oracle/rbdb1/system01.dbf' SIZE 10M REUSE AUTOEXTEND ON NEXT 10M MAXSIZE 200M CHARACTER SET WE8ISO8859P1; run catalog.sql run catproq.sql -- Create another (temporary) system tablespace CREATE ROLLBACK SEGMENT rb_temp STORAGE (INITIAL 100 k NEXT 250 k); -- Alter temporary system tablespace online before proceding ALTER ROLLBACK SEGMENT rb_temp ONLINE; -- Create additional tablespaces ... -- RBS: For rollback segments -- USERs: Create user sets this as the default tablespace -- TEMP: Create user sets this as the temporary tablespace CREATE TABLESPACE rbs DATAFILE '/u01/oracle/rbdb1/rbs01.dbf' SIZE 5M REUSE AUTOEXTEND ON NEXT 5M MAXSIZE 150M; CREATE TABLESPACE users DATAFILE '/u01/oracle/rbdb1/users01.dbf' SIZE 3M REUSE AUTOEXTEND ON NEXT 5M MAXSIZE 150M; CREATE TABLESPACE temp DATAFILE '/u01/oracle/rbdb1/temp01.dbf' SIZE 2M REUSE AUTOEXTEND ON NEXT 5M MAXSIZE 150M; -- Create rollback segments. CREATE ROLLBACK SEGMENT rb1 STORAGE(INITIAL 50K NEXT 250K) tablespace rbs; CREATE ROLLBACK SEGMENT rb2 STORAGE(INITIAL 50K NEXT 250K) tablespace rbs; CREATE ROLLBACK SEGMENT rb3 STORAGE(INITIAL 50K NEXT 250K) tablespace rbs; CREATE ROLLBACK SEGMENT rb4 STORAGE(INITIAL 50K NEXT 250K) tablespace rbs; -- Bring new rollback segments online and drop the temporary system one ALTER ROLLBACK SEGMENT rb1 ONLINE; ALTER ROLLBACK SEGMENT rb2 ONLINE; ALTER ROLLBACK SEGMENT rb3 ONLINE; ALTER ROLLBACK SEGMENT rb4 ONLINE; ALTER ROLLBACK SEGMENT rb_temp OFFLINE; DROP ROLLBACK SEGMENT rb_temp ; Example 2: ---------- connect internal startup nomount pfile=/disk00/oracle/software/7.3.4/dbs/initDB1.ora create database "DB1" maxinstances 2 maxlogfiles 32 maxdatafiles 254 characterset "US7ASCII" datafile '/disk02/oracle/oradata/DB1/system01.dbf' size 128M autoextent on next 8M maxsize 256M logfile group 1 ('/disk03/oracle/oradata/DB1/redo1a.log', '/disk04/oracle/oradata/DB1/redo1b.log') size 100M, group 2 ('/disk05/oracle/oradata/DB1/redo2a.log', ('/disk06/oracle/oradata/DB1/redo2b.log') size 100M REM * install data dictionary views @/disk00/oracle/software/7.3.4/rdbms/admin/catalog.sql @/disk00/oracle/software/7.3.4/rdbms/admin/catproq.sql create rollback segment SYSROLL tablespace system storage (initial 2M next 2M minextents 2 maxextents 255); alter rollback segment SYSROLL online; create tablespace RBS datafile '/disk01/oracle/oradata/DB1/rbs01.dbf' size 25M default storage ( initial 500K next 500K pctincrease 0 minextents 2 ); create rollback segment RBS_1 tablespace RBS1 storage (initial 512K next 512K minextents 50); create rollback segment RBS02 tablespace RBS storage (initial 500K next 500K minextents 2 optimal 1M); etc.. alter rollback segment RBS01 online; alter rollback segment RBS02 online; etc.. create tablespace DATA datafile '/disk05/oracle/oradata/DB1/data01.dbf' size 25M default storage ( initial 500K next 500K pctincrease 0 maxextends UNLIMITED ); etc.. other tablespaces you need run other scripts you need. alter user sys temporary tablespace TEMP; alter user system default tablespace TOOLS temporary tablespace TEMP; connect system/manager @/disk00/oracle/software/7.3.4/rdbms/admin/catdbsyn.sql @/disk00/oracle/software/7.3.4/rdbms/admin/pubbld.sql t.b.v. PRODUCT_USER_PROFILE, SQLPLUS_USER_PROFILE Example 3: on NT/2000 8i best example: -------------------------------------- Suppose you want a second database on a NT/2000 Server: 1. create a service with oradim oradim -new -sid -startmode -pfile 2. sqlplus /nolog (or use svrmgrl) startup nomount pfile="G:\oracle\admin\hd\pfile\init.ora" SVRMGR> CREATE DATABASE hd LOGFILE 'G:\oradata\hd\redo01.log' SIZE 2048K, 'G:\oradata\hd\redo02.log' SIZE 2048K, 'G:\oradata\hd\redo03.log' SIZE 2048K MAXLOGFILES 32 MAXLOGMEMBERS 2 MAXLOGHISTORY 1 DATAFILE 'G:\oradata\hd\system01.dbf' SIZE 264M REUSE AUTOEXTEND ON NEXT 10240K MAXDATAFILES 254 MAXINSTANCES 1 CHARACTER SET WE8ISO8859P1 NATIONAL CHARACTER SET WE8ISO8859P1; @catalog.sql @catproq.sql Oracle 9i: ---------- Example 1: ---------- CREATE DATABASE mynewdb USER SYS IDENTIFIED BY pz6r58 USER SYSTEM IDENTIFIED BY y1tz5p LOGFILE GROUP 1 ('/vobs/oracle/oradata/mynewdb/redo01.log') SIZE 100M, GROUP 2 ('/vobs/oracle/oradata/mynewdb/redo02.log') SIZE 100M, GROUP 3 ('/vobs/oracle/oradata/mynewdb/redo03.log') SIZE 100M MAXLOGFILES 5 MAXLOGMEMBERS 5 MAXLOGHISTORY 1 MAXDATAFILES 100 MAXINSTANCES 1 CHARACTER SET US7ASCII NATIONAL CHARACTER SET AL16UTF16 DATAFILE '/vobs/oracle/oradata/mynewdb/system01.dbf' SIZE 325M REUSE EXTENT MANAGEMENT LOCAL DEFAULT TEMPORARY TABLESPACE tempts1 DATAFILE '/vobs/oracle/oradata/mynewdb/temp01.dbf' SIZE 20M REUSE UNDO TABLESPACE undotbs DATAFILE '/vobs/oracle/oradata/mynewdb/undotbs01.dbf' SIZE 200M REUSE AUTOEXTEND ON NEXT 5120K MAXSIZE UNLIMITED; 9.2 Automatische start oracle bij system boot: ============================================== 9.2.1 oratab: ------------- Inhoud ORATAB in /etc of /var/opt: Voorbeeld: # $ORACLE_SID:$ORACLE_HOME:[N|Y] # ORCL:/u01/app/oracle/product/8.0.5:Y # De oracle scripts om de database te starten en te stoppen zijn: $ORACLE_HOME/bin/dbstart en dbshut, of startdb en stopdb of wat daarop lijkt. Deze kijken in ORATAB om te zien welke databases gestart moeten worden. 9.2.2 dbstart en dbshut: ------------------------ Het script dbstart zal oratab lezen en ook tests doen en om de oracle versie te bepalen. Verder bestaat de kern uit: het starten van sqldba, svrmgrl of sqlplus vervolgens doen we een connect vervolgens geven we het startup commando. Voor dbshut geldt een overeenkomstig verhaal. 9.2.3 init, sysinit, rc: ------------------------ Voor een automatische start, voeg nu de juiste entries toe in het /etc/rc2.d/S99dbstart (or equivalent) file: Tijdens het opstarten van Unix worden de scrips in de /etc/rc2.d uitgevoerd die beginnen met een 'S' en in alfabetische volgorde. De Oracle database processen zullen als (een van de) laatste processen worden gestart. Het bestAND S99oracle is gelinkt met deze directory. Inhoud S99oracle: su - oracle -c "/path/to/$ORACLE_HOME/bin/dbstart" # Start DB's su - oracle -c "/path/to/$ORACLE_HOME/bin/lsnrctl start" # Start listener su - oracle -c "/path/tp/$ORACLE_HOME/bin/namesctl start" # Start OraNames (optional) Het dbstart script is een standaard Oracle script. Het kijkt in oratab welke sid's op 'Y' staan, en zal deze databases starten. of customized via een customized startdb script: ORACLE_ADMIN=/opt/oracle/admin; export ORACLE_ADMIN su - oracle -c "$ORACLE_ADMIN/bin/startdb WPRD 1>$ORACLE_ADMIN/log/WPRD/startWPRD.$$ 2>&1" su - oracle -c "$ORACLE_ADMIN/bin/startdb WTST 1>$ORACLE_ADMIN/log/WTST/startWTST.$$ 2>&1" su - oracle -c "$ORACLE_ADMIN/bin/startdb WCUR 1>$ORACLE_ADMIN/log/WCUR/startWCUR.$$ 2>&1" 9.3 Het stoppen van Oracle in unix: ----------------------------------- Tijdens het down brengen van Unix (shutdown -i 0) worden de scrips in de directory /etc/rc2.d uitgevoerd die beginnen met een 'K' en in alfabetische volgorde. De Oracle database processen zijn een van de eerste processen die worden afgesloten. Het bestand K10oracle is gelinkt met de /etc/rc2.d/K10oracle # Configuration File: /opt/oracle/admin/bin/K10oracle ORACLE_ADMIN=/opt/oracle/admin; export ORACLE_ADMIN su - oracle -c "$ORACLE_ADMIN/bin/stopdb WPRD 1>$ORACLE_ADMIN/log/WPRD/stopWPRD.$$ 2>&1" su - oracle -c "$ORACLE_ADMIN/bin/stopdb WCUR 1>$ORACLE_ADMIN/log/WCUR/stopWCUR.$$ 2>&1" su - oracle -c "$ORACLE_ADMIN/bin/stopdb WTST 1>$ORACLE_ADMIN/log/WTST/stopWTST.$$ 2>&1" 9.4 startdb en stopdb: ---------------------- Startdb [ORACLE_SID] -------------------- Dit script is een onderdeel van het script S99Oracle. Dit script heeft 1 parameter, ORACLE_SID # Configuration File: /opt/oracle/admin/bin/startdb # Algemene omgeving zetten . $ORACLE_ADMIN/env/profile ORACLE_SID=$1 echo $ORACLE_SID # Omgeving zetten RDBMS . $ORACLE_ADMIN/env/$ORACLE_SID.env # Het starten van de database sqlplus /nolog << EOF connect / as sysdba startup EOF # Het starten van de listener lsnrctl start $ORACLE_SID # Het starten van de intelligent agent voor alle instances #lsnrctl dbsnmp_start Stopdb [ORACLE_SID] ------------------- Dit script is een onderdeel van het script K10Oracle. Dit script heeft 1 parameter, ORACLE_SID # Configuration File: /opt/oracle/admin/bin/stopdb # Algemene omgeving zetten . $ORACLE_ADMIN/env/profile ORACLE_SID=$1 export $ORACLE_SID # Settings van het RDBMS . $ORACLE_ADMIN/env/$ORACLE_SID.env # Het stoppen van de intelligent agent #lsnrctl dbsnmp_stop # Het stoppen van de listener lsnrctl stop $ORACLE_SID # Het stoppen van de database. sqlplus /nolog << EOF connect / as sysdba shutdown immediate EOF 9.5 Batches: ------------ De batches (jobs) worden gestart door het Unix proces cron # Batches (Oracle) # Configuration File: /var/spool/cron/crontabs/root # Format of lines: # min hour daymo month daywk cmd # # Dayweek 0=sunday, 1=monday... 0 9 * * 6 /sbin/sh /opt/oracle/admin/batches/bin/batches.sh >> /opt/oracle/admin/batches/log/batcheserroroutput.log 2>&1 # Configuration File: /opt/oracle/admin/batches/bin/batches.sh # Door de op de commandline ' BL_TRACE=T ; export BL_TRACE ' worden alle commando's getoond. case $BL_TRACE in T) set -x ;; esac ORACLE_ADMIN=/opt/oracle/admin; export ORACLE_ADMIN ORACLE_HOME=/opt/oracle/product/8.1.6; export ORACLE_HOME ORACLE_SID=WCUR ; export ORACLE_SID su - oracle -c ". $ORACLE_ADMIN/env/profile ; . $ORACLE_ADMIN/env/$ORACLE_SID.env; cd $ORACLE_ADMIN/batches/bin; sqlplus /NOLOG @$ORACLE_ADMIN/batches/bin/Analyse_WILLOW2K.sql 1> $ORACLE_ADMIN/batches/log/batches$ORACLE_SID.`date +"%y%m%d"` 2>&1" ORACLE_SID=WCON ; export ORACLE_SID su - oracle -c ". $ORACLE_ADMIN/env/profile ; . $ORACLE_ADMIN/env/$ORACLE_SID.env; cd $ORACLE_ADMIN/batches/bin; sqlplus /NOLOG @$ORACLE_ADMIN/batches/bin/Analyse_WILLOW2K.sql 1> $ORACLE_ADMIN/batches/log/batches$ORACLE_SID.`date +"%y%m%d"` 2>&1" 9.6 Autostart in NT/Win2K: -------------------------- 1) Older versions delete the existing instance FROM the command prompt: oradim80 -delete -sid SID recreate the instance FROM the command prompt: oradim -new -sid SID -intpwd -startmode -pfile Execute the command file FROM the command prompt: oracle_home\database\strt.cmd Check the log file generated FROM this execution: oracle_home\rdbmsxx\oradimxx.log 2) NT Registry value HKEY_LOCAL_MACHINE\SOFTWARE\ORACLE\HOME0\ORA_SID_AUTOSTART REG_EXPAND_SZ TRUE 9.7 Tools: ---------- Relink van Oracle: ------------------ info: showrev -p pkginfo -i relink: mk -f $ORACLE_HOME/rdbms/lib/ins_rdbms.mk install mk -f $ORACLE_HOME/svrmgr/lib/ins_svrmgr.mk install mk -f $ORACLE_HOME/network/lib/ins_network.mk install $ORACLE_HOME/bin relink all Relinking Oracle Background: Applications for UNIX are generally not distributed as complete executables. Oracle, like many application vendors who create products for UNIX, distribute individual object files, library archives of object files, and some source files which then get ?relinked? at the operating system level during installation to create usable executables. This guarantees a reliable integration with functions provided by the OS system libraries. Relinking occurs automatically under these circumstances: - An Oracle product has been installed with an Oracle provided installer. - An Oracle patch set has been applied via an Oracle provided installer. [Step 1] Log into the UNIX system as the Oracle software owner Typically this is the user 'oracle'. [STEP 2] Verify that your $ORACLE_HOME is set correctly: For all Oracle Versions and Platforms, perform this basic environment check first: % cd $ORACLE_HOME % pwd ...Doing this will ensure that $ORACLE_HOME is set correctly in your current environment. [Step 3] Verify and/or Configure the UNIX Environment for Proper Relinking: For all Oracle Versions and UNIX Platforms: The Platform specific environment variables LIBPATH, LD_LIBRARY_PATH, & SHLIB_PATH typically are already set to include system library locations like '/usr/lib'. In most cases, you need only check what they are set to first, then add the $ORACLE_HOME/lib directory to them where appropriate. i.e.: % setenv LD_LIBRARY_PATH ${ORACLE_HOME}/lib:${LD_LIBRARY_PATH} (see [NOTE:131207.1] How to Set UNIX Environment Variables for help with setting UNIX environment variables) If on SOLARIS (Sparc or Intel) with: Oracle 7.3.X, 8.0.X, or 8.1.X: - Ensure that /usr/ccs/bin is before /usr/ucb in $PATH % which ld ....should return '/usr/ccs/bin/ld' If using 32bit(non 9i) Oracle, - Set LD_LIBRARY_PATH=$ORACLE_HOME/lib If using 64bit(non 9i) Oracle, - Set LD_LIBRARY_PATH=$ORACLE_HOME/lib - Set LD_LIBRARY_PATH_64=$ORACLE_HOME/lib64 Oracle 9.X.X (64Bit) on Solaris (64Bit) OS - Set LD_LIBRARY_PATH=$ORACLE_HOME/lib32 - Set LD_LIBRARY_PATH_64=$ORACLE_HOME/lib Oracle 9.X.X (32Bit) on Solaris (64Bit) OS - Set LD_LIBRARY_PATH=$ORACLE_HOME/lib [Step 4] For all Oracle Versions and UNIX Platforms: Verify that you performed Step 2 correctly: % env|pg ....make sure that you see the correct absolute path for $ORACLE_HOME in the variable definitions. [Step 5] Run the OS Commands to Relink Oracle: Before relinking Oracle, shut down both the database and the listener. Oracle 8.1.X or 9.X.X ------------------------ *** NEW IN 8i AND ABOVE *** A 'relink' script is provided in the $ORACLE_HOME/bin directory. % cd $ORACLE_HOME/bin % relink ...this will display all of the command's options. usage: relink accepted values for parameter: all Every product executable that has been installed oracle Oracle Database executable only network net_client, net_server, cman client net_client, plsql client_sharedlib Client shared library interMedia ctx ctx Oracle Text utilities precomp All precompilers that have been installed utilities All utilities that have been installed oemagent oemagent Note: To give the correct permissions to the nmo and nmb executables, you must run the root.sh script after relinking oemagent. ldap ldap, oid Note: ldap option is available only from 9i. In 8i, you would have to manually relink ldap. You can relink most of the executables associated with an Oracle Server Installation by running the following command: % relink all This will not relink every single executable Oracle provides (you can discern which executables were relinked by checking their timestamp with 'ls -l' in the $ORACLE_HOME/bin directory). However, 'relink all' will recreate the shared libraries that most executables rely on and thereby resolve most issues that require a proper relink. -or- Since the 'relink' command merely calls the traditional 'make' commands, you still have the option of running the 'make' commands independently: For executables: oracle, exp, imp, sqlldr, tkprof, mig, dbv, orapwd, rman, svrmgrl, ogms, ogmsctl % cd $ORACLE_HOME/rdbms/lib % make -f ins_rdbms.mk install For executables: sqlplus % cd $ORACLE_HOME/sqlplus/lib % make -f ins_sqlplus.mk install For executables: isqlplus % cd $ORACLE_HOME/sqlplus/lib % make -f ins_sqlplus install_isqlplus For executables: dbsnmp, oemevent, oratclsh % cd $ORACLE_HOME/network/lib % make -f ins_oemagent.mk install For executables: names, namesctl % cd $ORACLE_HOME/network/lib % make -f ins_names.mk install For executables: osslogin, trcasst, trcroute, onrsd, tnsping % cd $ORACLE_HOME/network/lib % make -f ins_net_client.mk install For executables: tnslsnr, lsnrctl % cd $ORACLE_HOME/network/lib % make -f ins_net_server.mk install For executables related to ldap (for example Oracle Internet Directory): % cd $ORACLE_HOME/ldap/lib % make -f ins_ldap.mk install Note: Unix Installation/OS: RDBMS Technical Forum Displayed below are the messages of the selected thread. Thread Status: Closed From: Ray Stell 20-Apr-05 21:43 Subject: solaris upgrade RDBMS Version: 9.2.0.4 Operating System and Version: Solaris 8 Error Number (if applicable): Product (i.e. SQL*Loader, Import, etc.): Product Version: solaris upgrade I need to move a server from solaris 5.8 to 5.9. Does this require a new oracle 9.2.0 ee server install or relink or nothing at all? Thanks. -------------------------------------------------------------------------------- From: Samir Saad 21-Apr-05 03:28 Subject: Re : solaris upgrade You must relink even if you find that the databases came up after Solaris upgrade and they seem fine. As for the existing Oracle installations, they will all be fine. Samir. -------------------------------------------------------------------------------- From: Oracle, soumya anand 21-Apr-05 10:59 Subject: Re : solaris upgrade Hello Ray, As rightly pointed by Samir, after an OS upgrade it sufficient to relink the executables. Regards, Soumya Note: troubles after relink: ---------------------------- If you see on AIX something that resembles the following: P522:/home/oracle $lsnrctl exec(): 0509-036 Cannot load program lsnrctl because of the following errors: 0509-130 Symbol resolution failed for /usr/lib/libc.a[aio_64.o] because: 0509-136 Symbol kaio_rdwr64 (number 0) is not exported from dependent module /unix. 0509-136 Symbol listio64 (number 1) is not exported from dependent module /unix. 0509-136 Symbol acancel64 (number 2) is not exported from dependent module /unix. 0509-136 Symbol iosuspend64 (number 3) is not exported from dependent module /unix. 0509-136 Symbol aio_nwait (number 4) is not exported from dependent module /unix. 0509-150 Dependent module libc.a(aio_64.o) could not be loaded. 0509-026 System error: Cannot run a file that does not have a valid format. 0509-192 Examine .loader section symbols with the 'dump -Tv' command. If this occurs, you have asynchronous I/O turned off. To turn on asynchronous I/O: Run smitty chgaio and set STATE to be configured at system restart from defined to available. Press Enter. Do one of the following: Restart your system. Run smitty aio and move the cursor to Configure defined Asynchronous I/O. Then press Enter. trace: ------ truss -aef -o /tmp/trace svrmgrl To trace what a Unix process is doing enter: truss -rall -wall -p truss -p $ lsnrctl dbsnmp_start NOTE: The "truss" command works on SUN and Sequent. Use "tusc" on HP-UX, "strace" on Linux, "trace" on SCO Unix or call your system administrator to find the equivalent command on your system. Monitor your Unix system: Logfiles: --------- Unix message files record all system problems like disk errors, swap errors, NFS problems, etc. Monitor the following files on your system to detect system problems: tail -f /var/adm/SYSLOG tail -f /var/adm/messages tail -f /var/log/syslog =============== 10. CONSTRAINTS: =============== 10.1 index owner en table owner information: DBA_INDEXES ------------------------------------------- set linesize 100 SELECT DISTINCT substr(owner, 1, 10) as INDEX_OWNER, substr(index_name, 1, 40) as INDEX_NAME, substr(tablespace_name,1,40) as TABLE_SPACE, substr(index_type, 1, 10) as INDEX_TYPE, substr(table_owner, 1, 10) as TABLE_OWNER, substr(table_name, 1, 40) as TABLE_NAME, BLEVEL,NUM_ROWS,STATUS FROM DBA_INDEXES order by index_owner; SELECT DISTINCT substr(owner, 1, 10) as INDEX_OWNER, substr(index_name, 1, 40) as INDEX_NAME, substr(index_type, 1, 10) as INDEX_TYPE, substr(table_owner, 1, 10) as TABLE_OWNER, substr(table_name, 1, 40) as TABLE_NAME FROM DBA_INDEXES WHERE table_name='HEAT_CUSTOMER'; SELECT substr(owner, 1, 10) as INDEX_OWNER, substr(index_name, 1, 40) as INDEX_NAME, substr(index_type, 1, 10) as INDEX_TYPE, substr(table_owner, 1, 10) as TABLE_OWNER, substr(table_name, 1, 40) as TABLE_NAME FROM DBA_INDEXES WHERE owner<>table_owner; 10.2 PK en FK constraint relations: ---------------------------------- SELECT c.constraint_type as TYPE, SUBSTR(c.table_name, 1, 40) as TABLE_NAME, SUBSTR(c.constraint_name, 1, 40) as CONSTRAINT_NAME, SUBSTR(c.r_constraint_name, 1, 40) as REF_KEY, SUBSTR(b.column_name, 1, 40) as COLUMN_NAME FROM DBA_CONSTRAINTS c, DBA_CONS_COLUMNS b WHERE c.constraint_name=b.constraint_name AND c.OWNER in ('TRIDION_CM','TCMLOGDBUSER','VPOUSERDB') AND c.constraint_type in ('P', 'R', 'U'); SELECT c.constraint_type as TYPE, SUBSTR(c.table_name, 1, 40) as TABLE_NAME, SUBSTR(c.constraint_name, 1, 40) as CONSTRAINT_NAME, SUBSTR(c.r_constraint_name, 1, 40) as REF_KEY, SUBSTR(b.column_name, 1, 40) as COLUMN_NAME FROM DBA_CONSTRAINTS c, DBA_CONS_COLUMNS b WHERE c.constraint_name=b.constraint_name AND c.OWNER='RM_LIVE' AND c.constraint_type in ('P', 'R', 'U'); SELECT distinct c.constraint_type as TYPE, SUBSTR(c.table_name, 1, 40) as TABLE_NAME, SUBSTR(c.constraint_name, 1, 40) as CONSTRAINT_NAME, SUBSTR(c.r_constraint_name, 1, 40) as REF_KEY FROM DBA_CONSTRAINTS c, DBA_CONS_COLUMNS b WHERE c.constraint_name=b.constraint_name AND c.OWNER='RM_LIVE' AND c.constraint_type ='R'; ----------------------------------------------------------------------- create table reftables (TYPE varchar2(32), TABLE_NAME varchar2(40), CONSTRAINT_NAME varchar2(40), REF_KEY varchar2(40), REF_TABLE varchar2(40)); insert into reftables (type,table_name,constraint_name,ref_key) SELECT distinct c.constraint_type as TYPE, SUBSTR(c.table_name, 1, 40) as TABLE_NAME, SUBSTR(c.constraint_name, 1, 40) as CONSTRAINT_NAME, SUBSTR(c.r_constraint_name, 1, 40) as REF_KEY FROM DBA_CONSTRAINTS c, DBA_CONS_COLUMNS b WHERE c.constraint_name=b.constraint_name AND c.OWNER='RM_LIVE' AND c.constraint_type ='R'; update reftables set REF_TABLE=(select distinct table_name from dba_cons_columns where owner='RM_LIVE' and CONSTRAINT_NAME=REF_KEY); ---------------------------------------------------------------------- SELECT c.constraint_type as TYPE, SUBSTR(c.table_name, 1, 40) as TABLE_NAME, SUBSTR(c.constraint_name, 1, 40) as CONSTRAINT_NAME, SUBSTR(c.r_constraint_name, 1, 40) as REF_KEY FROM DBA_CONSTRAINTS c, DBA_CONS_COLUMNS b WHERE c.constraint_name=b.constraint_name AND c.OWNER='RM_LIVE' AND c.constraint_type ='R'; SELECT c.constraint_type as TYPE, SUBSTR(c.table_name, 1, 40) as TABLE_NAME, SUBSTR(c.constraint_name, 1, 40) as CONSTRAINT_NAME, SUBSTR(c.r_constraint_name, 1, 40) as REF_KEY, (select b.table_name from dba_cons_columns where b.constraint_name=c.r_constraint_name) as REF_TABLE FROM DBA_CONSTRAINTS c, DBA_CONS_COLUMNS b WHERE c.constraint_name=b.constraint_name AND c.OWNER='RM_LIVE' AND c.constraint_type ='R' or c.constraint_type ='P' ; select select c.constraint_name, c.constraint_type, c.table_name, (select table_name from c where c.r_constraint_name, o.constraint_name, o.column_name from dba_constraints c, dba_cons_columns o where c.constraint_name=o.constraint_name and c.constraint_type='R' and c.owner='BRAINS'; SELECT 'SELECT * FROM '||c.table_name||' WHERE '||b.column_name||' '||c.search_condition FROM DBA_CONSTRAINTS c, DBA_CONS_COLUMNS b WHERE c.constraint_name=b.constraint_name AND c.OWNER='BRAINS' AND c.constraint_type = 'C'; SELECT 'ALTER TABLE PROJECTS.'||table_name||' enable constraint '||constraint_name||';' FROM DBA_CONSTRAINTS WHERE owner='PROJECTS' AND constraint_type='R'; SELECT 'ALTER TABLE BRAINS.'||table_name||' disable constraint '||constraint_name||';' FROM USER_CONSTRAINTS WHERE owner='BRAINS' AND constraint_type='R'; 10.3 PK en FK constraint informatie: DBA_CONSTRAINTS ----------------------------------- -- owner and all foreign key, constraints SELECT SUBSTR(owner, 1, 10) as OWNER, constraint_type as TYPE, SUBSTR(table_name, 1, 40) as TABLE_NAME, SUBSTR(constraint_name, 1, 40) as CONSTRAINT_NAME, SUBSTR(r_constraint_name, 1, 40) as REF_KEY, DELETE_RULE as DELETE_RULE, status FROM DBA_CONSTRAINTS WHERE OWNER='BRAINS' AND constraint_type in ('R', 'P', 'U'); SELECT SUBSTR(owner, 1, 10) as OWNER, constraint_type as TYPE, SUBSTR(table_name, 1, 30) as TABLE_NAME, SUBSTR(constraint_name, 1, 30) as CONSTRAINT_NAME, SUBSTR(r_constraint_name, 1, 30) as REF_KEY, DELETE_RULE as DELETE_RULE, status FROM DBA_CONSTRAINTS WHERE OWNER='BRAINS' AND constraint_type in ('R'); -- owner en alle primary key constraints bepalen van een bepaalde user, op bepaalde objects Zelfde query: Zet OWNER='gewenste_owner' AND constraint_type='P' select owner, CONSTRAINT_NAME, CONSTRAINT_TYPE,TABLE_NAME,R_CONSTRAINT_NAME,STATUS from dba_constraints where owner='FIN_VLIEG' and constraint_type in ('P','R','U'); 10.4 opsporen bijbehorende index van een bepaalde constraint: DBA_INDEXES, DBA_CONSTRAINTS ------------------------------------------------------------ SELECT c.constraint_type as Type, substr(x.index_name, 1, 40) as INDX_NAME, substr(c.constraint_name, 1, 40) as CONSTRAINT_NAME, substr(x.tablespace_name, 1, 40) as TABLESPACE FROM DBA_CONSTRAINTS c, DBA_INDEXES x WHERE c.constraint_name=x.index_name AND c.constraint_name='UN_DEMO1'; SELECT c.constraint_type as Type, substr(x.index_name, 1, 40) as INDX_NAME, substr(c.constraint_name, 1, 40) as CONSTRAINT_NAME, substr(c.table_name, 1, 40) as TABLE_NAME, substr(c.owner, 1, 10) as OWNER FROM DBA_CONSTRAINTS c, DBA_INDEXES x WHERE c.constraint_name=x.index_name AND c.owner='JOOPLOC'; 10.5 opsporen tablespace van een constraint of constraint owner: --------------------------------------------------------------- SELECT substr(s.segment_name, 1, 40) as Segmentname, substr(c.constraint_name, 1, 40) as Constraintname, substr(s.tablespace_name, 1, 40) as Tablespace, substr(s.segment_type, 1, 10) as Type FROM DBA_SEGMENTS s, DBA_CONSTRAINTS c WHERE s.segment_name=c.constraint_name AND c.owner='PROJECTS'; 10.6 Ophalen index create statements: ------------------------------------ DBA_INDEXES DBA_IND_COLUMNS SELECT substr(i.index_name, 1, 40) as INDEX_NAME, substr(i.index_type, 1, 15) as INDEX_TYPE, substr(i.table_name, 1, 40) as TABLE_NAME, substr(c.index_owner, 1, 10) as INDEX_OWNER, substr(c.column_name, 1, 40) as COLUMN_NAME, c.column_position as POSITION FROM DBA_INDEXES i, DBA_IND_COLUMNS c WHERE i.index_name=c.index_name AND i.owner='SALES'; 10.7 Aan en uitzetten van constraints: ------------------------------------- -- aanzetten: alter table tablename enable constraint constraint_name -- uitzetten: alter table tablename disable constraint constraint_name -- voorbeeld: ALTER TABLE EMPLOYEE DISABLE CONSTRAINT FK_DEPNO; ALTER TABLE EMPLOYEE ENABLE CONSTRAINT FK_DEPNO; maar ook kan: ALTER TABLE DEMO ENABLE PRIMARY KEY; -- Alle FK constraints van een schema in een keer uitzetten: SELECT 'ALTER TABLE MIS_OWNER.'||table_name||' disable constraint '||constraint_name||';' FROM DBA_CONSTRAINTS WHERE owner='MIS_OWNER' AND constraint_type='R' AND TABLE_NAME LIKE 'MKM%'; SELECT 'ALTER TABLE MIS_OWNER.'||table_name||' enable constraint '||constraint_name||';' FROM DBA_CONSTRAINTS WHERE owner='MIS_OWNER' AND constraint_type='R' AND TABLE_NAME LIKE 'MKM%'; 10.8 Constraint aanmaken en initieel uit: ---------------------------------------- Dit kan handig zijn bij bijvoorbeeld het laden van een table waarbij mogelijk dubbele waarden voorkomen ALTER TABLE CUSTOMERS ADD CONSTRAINT PK_CUST PRIMARY KEY (custid) DISABLE; Als nu blijkt dat bij het aanzetten van de constraint, er dubbele records voorkomen, kunnen we deze dubbele records plaatsen in de EXCEPTIONS table: 1. aanmaken EXCEPTIONS table: @ORACLE_HOME\rdbms\admin\utlexcpt.sql 2. Constraint aaNzetten: ALTER TABLE CUSTOMERS ENABLE PRIMARY KEY exceptions INTO EXCEPTIONS; Nu bevat de EXCEPTIONS table de dubbele rijen. 3. Welke dubbele rijen: SELECT c.custid, c.name FROM CUSTOMERS c, EXCEPTIONS s WHERE c.rowid=s.row_id; 10.9 Gebruik PK FK constraints: ------------------------------ 10.9.1: Voorbeeld normaal gebruik met DRI: create table customers ( custid number not null, custname varchar(10), CONSTRAINT pk_cust PRIMARY KEY (custid) ); create table contacts ( contactid number not null, custid number, contactname varchar(10), CONSTRAINT pk_contactid PRIMARY KEY (contactid), CONSTRAINT fk_cust FOREIGN KEY (custid) REFERENCES customers(custid) ); Hierbij kun je dus niet zondermeer een row met een bepaald custid uit customers verwijderen, indien er een row in contacts bestaat met hetzelfde custid. 10.9.2: Voorbeeld met ON DELETE CASCADE: create table contacts ( contactid number not null, custid number, contactname varchar(10), CONSTRAINT pk_contactid PRIMARY KEY (contactid), CONSTRAINT fk_cust FOREIGN KEY (custid) REFERENCES customers(custid) ON DELETE CASCADE ); Ook de clausule "ON DELETE SET NULL" kan gebruikt worden. Nu is het wel mogelijk om in customers een row te verwijderen, terwijl in contacts een overeenkomende custid bestaat. De row in contacts wordt dan namelijk ook verwijdert. 10.10 Procedures voor insert, delete: ------------------------------------ Als voorbeeld op table customers: CREATE OR REPLACE PROCEDURE newcustomer (custid NUMBER, custname VARCHAR) IS BEGIN INSERT INTO customers values (custid,custname); commit; END; / CREATE OR REPLACE PROCEDURE delcustomer (cust NUMBER) IS BEGIN delete from customers where custid=cust; commit; END; / 10.11 User datadictonary views: ----------------------------- We hebben al gezien dat we voor constraint informatie voornamelijk de onderstaande views raadplegen: DBA_TABLES DBA_INDEXES, DBA_CONSTRAINTS, DBA_IND_COLUMNS, DBA_SEGMENTS Deze zijn echter voor de DBA. Gewone users kunnen informatie opvragen uit USER_ en ALL_ views. USER_ : in the schema van de user ALL_ : waar de user bij kan USER_TABLES, ALL_TABLES USER_INDEXES, ALL_INDEXES USER_CONSTRAINTS, ALL_CONSTRAINTS USER_VIEWS, ALL_VIEWS USER_SEQUENCES, ALL_SEQUENCES USER_CONS_COLUMNS, ALL_CONS_COLUMNS USER_TAB_COLUMNS, ALL_TAB_COLUMNS USER_SOURCE, ALL_SOURCE cat tab col dict 10.12 Create en drop index examples: ----------------------------------- CREATE UNIQUE INDEX HEATCUST0 ON HEATCUST(CUSTTYPE) TABLESPACE INDEX_SMALL PCTFREE 10 STORAGE(INITIAL 163840 NEXT 163840 PCTINCREASE 0 ); DROP INDEX indexname 10.13 Check the height of indexes: --------------------------------- Is an index rebuild neccessary ? SELECT index_name, owner, blevel, decode(blevel,0,'OK BLEVEL',1,'OK BLEVEL', 2,'OK BLEVEL',3,'OK BLEVEL',4,'OK BLEVEL','BLEVEL HIGH') OK FROM dba_indexes WHERE owner='SALES' and blevel > 3; 10.14 Make indexes unusable (before a large dataload): ----------------------------------------------------- -- Make Indexes unusable alter index HEAT_CUSTOMER_DISCON_DATE unusable; alter index HEAT_CUSTOMER_EMAIL_ADDRESS unusable; alter index HEAT_CUSTOMER_POSTAL_CODE unusable; -- Enable Indexes again alter index HEAT_CUSTOMER_DISCON_DATE rebuild; alter index HEAT_CUSTOMER_EMAIL_ADDRESS rebuild; alter index HEAT_CUSTOMER_POSTAL_CODE rebuild; ================================ 11. DBMS_JOB and scheduled Jobs: ================================ Used in Oracle 9i and lower versions. 11.1 SNP background process: ---------------------------- Scheduled jobs zijn mogelijk wanneer het SNP background process geactiveerd is. Dit kan via de init.ora: JOB_QUEUE_PROCESSES=1 aantal SNP processes (SNP0, SNP1), max 36 t.b.v. replication en jobqueue's JOB_QUEUE_INTERVAL=60 check interval 11.2 DBMS_JOB package: ---------------------- DBMS_JOB.SUBMIT() DBMS_JOB.REMOVE() DBMS_JOB.CHANGE() DBMS_JOB.WHAT() DBMS_JOB.NEXT_DATE() DBMS_JOB.INTERVAL() DBMS_JOB.RUN() 11.2.1 DBMS_JOB.SUBMIT() ----------------------- There are actually two versions SUBMIT() and ISUBMIT() PROCEDURE DBMS_JOB.SUBMIT (job OUT BINARY_INTEGER, what IN VARCHAR2, next_date IN DATE DEFAULT SYSDATE, interval IN VARCHAR2 DEFAULT 'NULL', no_parse IN BOOLEAN DEFAULT FALSE); PROCEDURE DBMS_JOB.ISUBMIT (job IN BINARY_INTEGER, what IN VARCHAR2, next_date in DATE DEFAULT SYSDATE interval IN VARCHAR2 DEFAULT 'NULL', no_parse in BOOLEAN DEFAULT FALSE); The difference between ISUBMIT and SUBMIT is that ISUBMIT specifies a job number, whereas SUBMIT returns a job number generated by the DBMS_JOB package Look for submitted jobs: ------------------------ select job, last_date, next_date, interval, substr(what, 1, 50) from dba_jobs; Submit a job: -------------- The jobnumber (if you use SUBMIT() ) will be derived from the sequence SYS.JOBSEQ Suppose you have the following procedure: create or replace procedure test1 is begin dbms_output.put_line('Hallo grapjas.'); end; / Example 1: ---------- variable jobno number; begin DBMS_JOB.SUBMIT(:jobno, 'test1;', Sysdate, 'Sysdate+1'); commit; end; / DECLARE jobno NUMBER; BEGIN DBMS_JOB.SUBMIT (job => jobno ,what => 'test1;' ,next_date => SYSDATE ,interval => 'SYSDATE+1/24'); COMMIT; END; / So suppose you submit the above job at 08.15h. Then the next, and first time, that the job will run is at 09.15h. Example 2: ---------- variable jobno number; begin DBMS_JOB.SUBMIT(:jobno, 'test1;', LAST_DAY(SYSDATE+1), 'LAST_DAY(ADD_MONTHS(LAST_DAY(SYSDATE+1),1))'); commit; end; / Example 3: ---------- VARIABLE jobno NUMBER BEGIN DBMS_JOB.SUBMIT(:jobno, 'DBMS_DDL.ANALYZE_OBJECT(''TABLE'', ''CHARLIE'', ''X1'', ''ESTIMATE'', NULL, 50);', SYSDATE, 'SYSDATE + 1'); COMMIT; END; / PRINT jobno JOBNO ---------- 14144 Example 4: this job is scheduled every hour ------------------------------------------- DECLARE jobno NUMBER; BEGIN DBMS_JOB.SUBMIT (job => jobno ,what => 'begin space_logger; end;' ,next_date => SYSDATE ,interval => 'SYSDATE+1/24'); COMMIT; END; / Example 5: Examples of intervals -------------------------------- 'SYSDATE + 7' :exactly seven days from the last execution 'SYSDATE + 1/48' :every half hour 'NEXT_DAY(TRUNC(SYSDATE), ''MONDAY'') + 15/24' :every Monday at 3PM 'NEXT_DAY(ADD_MONTHS(TRUNC(SYSDATE, ''Q''), 3), ''THURSDAY'')' :first Thursday of each quarter 'TRUNC(SYSDATE + 1)' :Every day at 12:00 midnight 'TRUNC(SYSDATE + 1) + 8/24' :Every day at 8:00 a.m. 'NEXT_DAY(TRUNC(SYSDATE ), "TUESDAY" ) + 12/24' :Every Tuesday at 12:00 noon 'TRUNC(LAST_DAY(SYSDATE ) + 1)' :First day of the month at midnight 'TRUNC(ADD_MONTHS(SYSDATE + 2/24, 3 ), 'Q' ) - 1/24' :Last day of the quarter at 11:00 p.m. NEXT_DAY(SYSDATE, "FRIDAY") ) ) + 9/24' :Every Monday, Wednesday, and Friday at 9:00 a.m. --------------------------------------------------------------------------------- Example 6: ---------- You have this testprocedure create or replace procedure test1 as id_next number; begin select max(id) into id_next from iftest; insert into iftest (id) values (id_next+1); commit; end; / Suppose on 16 juli at 9:26h you do: variable jobno number; begin DBMS_JOB.SUBMIT(:jobno, 'test1;', LAST_DAY(SYSDATE+1), 'LAST_DAY(ADD_MONTHS(LAST_DAY(SYSDATE+1),1))'); commit; end; / select job, to_char(this_date,'DD-MM-YYYY;HH24:MI'), to_char(next_date, 'DD-MM-YYYY;HH24:MI') from dba_jobs; JOB TO_CHAR(THIS_DAT TO_CHAR(NEXT_DAT ---------- ---------------- ---------------- 25 31-07-2004;09:26 Suppose on 16 juli at 9:38h you do: variable jobno number; begin DBMS_JOB.SUBMIT(:jobno, 'test1;', LAST_DAY(SYSDATE)+1, 'LAST_DAY(ADD_MONTHS(LAST_DAY(SYSDATE+1),1))'); commit; end; / JOB TO_CHAR(THIS_DAT TO_CHAR(NEXT_DAT ---------- ---------------- ---------------- 25 31-07-2004;09:26 26 01-08-2004;09:38 Suppose on 16 juli at 9:41h you do: variable jobno number; begin DBMS_JOB.SUBMIT(:jobno, 'test1;', SYSDATE, 'LAST_DAY(ADD_MONTHS(LAST_DAY(SYSDATE+1),1))'); commit; end; / JOB TO_CHAR(THIS_DAT TO_CHAR(NEXT_DAT ---------- ---------------- ---------------- 27 31-08-2004;09:41 25 31-07-2004;09:26 26 01-08-2004;09:39 Suppose on 16 juli at 9:46h you do: variable jobno number; begin DBMS_JOB.SUBMIT(:jobno, 'test1;', SYSDATE, 'TRUNC(LAST_DAY(SYSDATE + 1/24 ) )'); commit; end; / JOB TO_CHAR(THIS_DAT TO_CHAR(NEXT_DAT --------- ---------------- ---------------- 27 31-08-2004;09:41 28 31-07-2004;00:00 25 31-07-2004;09:26 29 31-07-2004;00:00 -------------------------------------------------------------------------------------- variable jobno number; begin DBMS_JOB.SUBMIT(:jobno, 'test1;', null, 'TRUNC(LAST_DAY(SYSDATE ) + 1)' ); commit; end; / In the job definition, use two single quotation marks around strings. Always include a semicolon at the end of the job definition. 11.2.2 DBMS_JOB.REMOVE() ------------------------ Removing a Job FROM the Job Queue To remove a job FROM the job queue, use the REMOVE procedure in the DBMS_JOB package. The following statements remove job number 14144 FROM the job queue: BEGIN DBMS_JOB.REMOVE(14144); END; / 11.2.3 DBMS_JOB.CHANGE() ------------------------ In this example, job number 14144 is altered to execute every three days: BEGIN DBMS_JOB.CHANGE(1, NULL, NULL, 'SYSDATE + 3'); END; / If you specify NULL for WHAT, NEXT_DATE, or INTERVAL when you call the procedure DBMS_JOB.CHANGE, the current value remains unchanged. 11.2.4 DBMS_JOB.WHAT() ---------------------- You can alter the definition of a job by calling the DBMS_JOB.WHAT procedure. The following example changes the definition for job number 14144: BEGIN DBMS_JOB.WHAT(14144, 'DBMS_DDL.ANALYZE_OBJECT(''TABLE'', ''HR'', ''DEPARTMENTS'', ''ESTIMATE'', NULL, 50);'); END; / 11.2.5 DBMS_JOB.NEXT_DATE() --------------------------- You can alter the next execution time for a job by calling the DBMS_JOB.NEXT_DATE procedure, as shown in the following example: BEGIN DBMS_JOB.NEXT_DATE(14144, SYSDATE + 4); END; / 11.2.6 DBMS_JOB.INTERVAL(): --------------------------- The following example illustrates changing the execution interval for a job by calling the DBMS_JOB.INTERVAL procedure: BEGIN DBMS_JOB.INTERVAL(14144, 'NULL'); END; / execute dbms_job.interval(,'SYSDATE+(1/48)'); In this case, the job will not run again after it successfully executes and it will be deleted FROM the job queue 11.2.7 DBMS_JOB.BROKEN(): ------------------------- A job is labeled as either broken or not broken. Oracle does not attempt to run broken jobs. Example: BEGIN DBMS_JOB.BROKEN(10, TRUE); END; / Example: The following example marks job 14144 as not broken and sets its next execution date to the following Monday: BEGIN DBMS_JOB.BROKEN(14144, FALSE, NEXT_DAY(SYSDATE, 'MONDAY')); END; / Example: exec DBMS_JOB.BROKEN( V_JOB_ID, true); Example: select JOB into V_JOB_ID from DBA_JOBS where WHAT like '%SONERA%'; DBMS_SNAPSHOT.REFRESH( 'SONERA', 'C'); DBMS_JOB.BROKEN( V_JOB_ID, false); fix broken jobs: ---------------- /* Filename on companion disk: job5.sql */* CREATE OR REPLACE PROCEDURE job_fixer AS /* || calls DBMS_JOB.BROKEN to try and set || any broken jobs to unbroken */ /* cursor selects user's broken jobs */ CURSOR broken_jobs_cur IS SELECT job FROM user_jobs WHERE broken = 'Y'; BEGIN FOR job_rec IN broken_jobs_cur LOOP DBMS_JOB.BROKEN(job_rec.job,FALSE); END LOOP; END job_fixer; 11.2.8 DBMS_JOB.RUN(): ---------------------- BEGIN DBMS_JOB.RUN(14144); END; / 11.3 DBMS_SCHEDULER: -------------------- Used in Oracle 10g. BEGIN DBMS_SCHEDULER.create_job ( job_name => 'test_self_contained_job', job_type => 'PLSQL_BLOCK', job_action => 'BEGIN DBMS_STATS.gather_schema_stats(''JOHN''); END;', start_date => SYSTIMESTAMP, repeat_interval => 'freq=hourly; byminute=0', end_date => NULL, enabled => TRUE, comments => 'Job created using the CREATE JOB procedure.'); End; / BEGIN DBMS_SCHEDULER.run_job (job_name => 'TEST_PROGRAM_SCHEDULE_JOB', use_current_session => FALSE); END; / BEGIN DBMS_SCHEDULER.stop_job (job_name => 'TEST_PROGRAM_SCHEDULE_JOB'); END; / Jobs can be deleted using the DROP_JOB procedure: BEGIN DBMS_SCHEDULER.drop_job (job_name => 'TEST_PROGRAM_SCHEDULE_JOB'); DBMS_SCHEDULER.drop_job (job_name => 'test_self_contained_job'); END; / Oracle 10g: ----------- BMS_JOB has been replaced by DBMS_SCHEDULER. Views: V_$SCHEDULER_RUNNING_JOBS GV_$SCHEDULER_RUNNING_JOBS DBA_QUEUE_SCHEDULES USER_QUEUE_SCHEDULES _DEFSCHEDULE DEFSCHEDULE AQ$SCHEDULER$_JOBQTAB_S AQ$_SCHEDULER$_JOBQTAB_F AQ$SCHEDULER$_JOBQTAB AQ$SCHEDULER$_JOBQTAB_R AQ$SCHEDULER$_EVENT_QTAB_S AQ$_SCHEDULER$_EVENT_QTAB_F AQ$SCHEDULER$_EVENT_QTAB AQ$SCHEDULER$_EVENT_QTAB_R DBA_SCHEDULER_PROGRAMS USER_SCHEDULER_PROGRAMS ALL_SCHEDULER_PROGRAMS DBA_SCHEDULER_JOBS USER_SCHEDULER_JOBS ALL_SCHEDULER_JOBS DBA_SCHEDULER_JOB_CLASSES ALL_SCHEDULER_JOB_CLASSES DBA_SCHEDULER_WINDOWS ALL_SCHEDULER_WINDOWS DBA_SCHEDULER_PROGRAM_ARGS USER_SCHEDULER_PROGRAM_ARGS ALL_SCHEDULER_PROGRAM_ARGS DBA_SCHEDULER_JOB_ARGS USER_SCHEDULER_JOB_ARGS ALL_SCHEDULER_JOB_ARGS DBA_SCHEDULER_JOB_LOG DBA_SCHEDULER_JOB_RUN_DETAILS USER_SCHEDULER_JOB_LOG USER_SCHEDULER_JOB_RUN_DETAILS ALL_SCHEDULER_JOB_LOG ALL_SCHEDULER_JOB_RUN_DETAILS DBA_SCHEDULER_WINDOW_LOG DBA_SCHEDULER_WINDOW_DETAILS ALL_SCHEDULER_WINDOW_LOG ALL_SCHEDULER_WINDOW_DETAILS DBA_SCHEDULER_WINDOW_GROUPS ALL_SCHEDULER_WINDOW_GROUPS DBA_SCHEDULER_WINGROUP_MEMBERS ALL_SCHEDULER_WINGROUP_MEMBERS DBA_SCHEDULER_SCHEDULES USER_SCHEDULER_SCHEDULES ALL_SCHEDULER_SCHEDULES DBA_SCHEDULER_RUNNING_JOBS ALL_SCHEDULER_RUNNING_JOBS USER_SCHEDULER_RUNNING_JOBS DBA_SCHEDULER_GLOBAL_ATTRIBUTE ALL_SCHEDULER_GLOBAL_ATTRIBUTE DBA_SCHEDULER_CHAINS USER_SCHEDULER_CHAINS ALL_SCHEDULER_CHAINS DBA_SCHEDULER_CHAIN_RULES USER_SCHEDULER_CHAIN_RULES ALL_SCHEDULER_CHAIN_RULES DBA_SCHEDULER_CHAIN_STEPS USER_SCHEDULER_CHAIN_STEPS ALL_SCHEDULER_CHAIN_STEPS DBA_SCHEDULER_RUNNING_CHAINS USER_SCHEDULER_RUNNING_CHAINS ALL_SCHEDULER_RUNNING_CHAINS ================== 12. Net8 / SQLNet: ================== In bijvoorbeeld sql*plus vult men in: ----------------- Username: system Password: manager Host String: XXX ----------------- NET8 bij de client kijkt in TNSNAMES.ORAnaar de eerste entry XXX= (description.. protocol..host...port.. SERVICE_NAME=Y) XXX is eigenlijk een alias en is dus willekeurig hoewel het uiteraard aansluit bij de instance name of database name waarnaar je wilt connecten. Maar het zou dus zelfs pipo mogen zijn. Wordt XXX niet gevonden, dan meld de client: ORA-12154 TNS: could not resolve SERVICE NAME Vervolgens wordt door NET8 via de connect descriptor Y contact gemaakt met de listener op de Server die luistert naar Y Is Y niet wat de listener verwacht, dan meldt de listener aan de client: TNS: listener could not resolve SERVICE_NAME in connect descriptor 12.1 sqlnet.ora voorbeeld: -------------------------- SQLNET.AUTHENTICATION_SERVICES= (NTS) NAMES.DIRECTORY_PATH= (TNSNAMES) 12.2 tnsnames.ora voorbeelden: ------------------------------ voorbeeld 1. DB1= (DESCRIPTION= (ADDRESS_LIST= (ADDRESS=(PROTOCOL=TCP)(HOST=STARBOSS)(PORT=1521) ) (CONNECT_DATA= (SERVICE_NAME=DB1.world) ) ) voorbeeld 2. DB1.world= (DESCRIPTION= (ADDRESS_LIST= (ADDRESS=(COMMUNITY=tcp.world)(PROTOCOL=TCP)(HOST=STARBOSS)(PORT=1521) ) (CONNECT_DATA=(SID=DB1) ) ) DB2.world= (... ) DB3.world= (... ) etc.. voorbeeld 3. RCAT = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = w2ktest)(PORT = 1521)) ) (CONNECT_DATA = (SERVICE_NAME = rcat.antapex) ) ) 12.3 listener.ora voorbeelden: ------------------------------ Example 1: ---------- LISTENER= (DESCRIPTION= (ADDRESS=(PROTOCOL=TCP)(HOST=STARBOSS)(PORT=1521)) ) SID_LIST_LISTENER= (SID_LIST= (SID_DESC= (GLOBAL_DBNAME=DB1.world) (ORACLE_HOME=D:\oracle8i) (SID_NAME=DB1) ) ) Example 2: ---------- ############## WPRD ##################################################### LOG_DIRECTORY_WPRD = /opt/oracle/admin/WPRD/network/log LOG_FILE_WPRD = WPRD.log TRACE_LEVEL_WPRD = OFF #ADMIN TRACE_DIRECTORY_WPRD = /opt/oracle/admin/WPRD/network/trace TRACE_FILE_WPRD = WPRD.trc WPRD = (DESCRIPTION_LIST = (DESCRIPTION = (ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=blnl01)(PORT=1521))))) SID_LIST_WPRD = (SID_LIST = (SID_DESC = (GLOBAL_DBNAME = WPRD) (ORACLE_HOME = /opt/oracle/product/8.1.6) (SID_NAME = WPRD))) ############## WTST ##################################################### LOG_DIRECTORY_WTST = /opt/oracle/admin/WTST/network/log LOG_FILE_WTST = WTST.log TRACE_LEVEL_WTST = OFF #ADMIN TRACE_DIRECTORY_WTST = /opt/oracle/admin/WTST/network/trace TRACE_FILE_WTST = WTST.trc WTST = (DESCRIPTION_LIST = (DESCRIPTION = (ADDRESS_LIST=(ADDRESS=(PROTOCOL=TCP)(HOST=blnl01)(PORT=1522))))) SID_LIST_WTST = (SID_LIST = (SID_DESC = (GLOBAL_DBNAME = WTST) (ORACLE_HOME = /opt/oracle/product/8.1.6) (SID_NAME = WTST))) Example 3: ---------- # LISTENER.ORA Network Configuration File: D:\oracle\ora901\NETWORK\ADMIN\listener.ora # Generated by Oracle configuration tools. LISTENER = (DESCRIPTION_LIST = (DESCRIPTION = (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC0)) ) (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = missrv)(PORT = 1521)) ) ) SID_LIST_LISTENER = (SID_LIST = (SID_DESC = (SID_NAME = PLSExtProc) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = extproc) ) (SID_DESC = (GLOBAL_DBNAME = o901) (ORACLE_HOME = D:\oracle\ora901) (SID_NAME = o901) ) (SID_DESC = (SID_NAME = MAST) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = hsodbc) ) (SID_DESC = (SID_NAME = NATOPS) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = hsodbc) ) (SID_DESC = (SID_NAME = VRF) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = hsodbc) ) (SID_DESC = (SID_NAME = DRILLS) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = hsodbc) ) (SID_DESC = (SID_NAME = DDS) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = hsodbc) ) (SID_DESC = (SID_NAME = IVP) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = hsodbc) (SID_DESC = (SID_NAME = ALBERT) (ORACLE_HOME = D:\oracle\ora901) (PROGRAM = hsodbc) ) ) 12.4: CONNECT TIME FAILOVER: ---------------------------- The connect-time failover feature allows clients to connect to another listener if the initial connection to the first listener fails. Multiple listener locations are specified in the clients tnsnames.ora file. If a connection attempt to the first listener fails, a connection request to the next listener in the list is attempted. This feature increases the availablity of the Oracle service should a listener location be unavailable. Here is an example of what a tnsnames.ora file looks like with connect-time failover enabled: ORCL= (DESCRIPTION= (ADDRESS_LIST= (ADDRESS=(PROTOCOL=TCP)(HOST=DBPROD)(PORT=1521)) (ADDRESS=(PROTOCOL=TCP)(HOST=DBFAIL)(PORT=1521)) ) (CONNECT_DATA=(SERVICE_NAME=PROD)(SERVER=DEDICATED) ) ) 12.5: CLIENT LOAD BALANCING: ---------------------------- Client Load Balancing is a feature that allows clients to randomly select from a list of listeners. Oracle Net moves through the list of listeners and balances the load of connection requests accross the available listeners. Here is an example of the tnsnames.ora entry that allows for load balancing: ORCL= (DESCRIPTION= (LOAD_BALANCE=ON) (ADDRESS_LIST= (ADDRESS=(PROTOCOL=TCP)(HOST=MWEISHAN-DELL)(PORT=1522)) (ADDRESS=(PROTOCOL=TCP)(HOST=MWEISHAN-DELL)(PORT=1521)) ) (CONNECT_DATA=(SERVICE_NAME=PROD)(SERVER=DEDICATED) ) ) Notice the additional parameter of LOAD_BALANCE. This enables load balancing between the two listener locations specified. 12.6: ORACLE SHARED SERVER: --------------------------- With the dedicated Server, each server process has a PGA, outside the SGA When Shared Server is used, the user program area's are in the SGA in the large pool. With a few init.ora parameters, you can configure Shared Server. 1. DISPATCHERS: The DISPATCHERS parameter defines the number of dispatchers that should start when the instance is started. For example, if you want to configure 3 TCP/IP dispatchers and to IPC dispatchers, you set the parameters as follows: DISPATCHERS="(PRO=TCP)(DIS=3)(PRO=IPC)(DIS=2)" For example, if you have 500 concurrent TCP/IP connections, and you want each dispatcher to manage 50 concurrent connections, you need 10 dispatchers. You set your DISPATCHERS parameter as follows: DISPATCHERS="(PRO=TCP)(DIS=10)" 2. SHARED_SERVER: The Shared_Servers parameter specifies the minimum number of Shared Servers to start and retain when the Oracle instance is started. View information about dispatchers and shared servers with the following commands and queries: lsnrctl services SELECT name, status, messages, idle, busy, bytes, breaks FROM v$dispatcher; 12.7: Keeping Oracle connections alive through a Firewall: ---------------------------------------------------------- Implementing keep alive packets: SQLNET.INBOUND_CONNECT_TIMEOUT Notes: ======= Note 1: ------- Doc ID: Note:274130.1 Content Type: TEXT/PLAIN Subject: SHARED SERVER CONFIGURATION Creation Date: 25-MAY-2004 Type: BULLETIN Last Revision Date: 24-JUN-2004 Status: PUBLISHED PURPOSE ------- This article discusses about the configuration of shared servers on 9i DB. SHARED SERVER CONFIGURATION: =========================== 1. Add the parameter shared_servers in the init.ora SHARED_SERVERS specifies the number of server processes that you want to create when an instance is started up. If system load decreases, this minimum number of servers is maintained. Therefore, you should take care not to set SHARED_SERVERS too high at system startup. Parameter type Integer Parameter class Dynamic: ALTER SYSTEM 2. Add the parameter DISPATCHERS in the init.ora DISPATCHERS configures dispatcher processes in the shared server architecture. USAGE: ----- DISPATCHERS = "(PROTOCOL=TCP)(DISPATCHERS=3)" 3. Save the init.ora file. 4. Change the connect string in tnsnames.ora from ORACLE.IDC.ORACLE.COM = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = xyzac)(PORT = 1521)) ) (CONNECT_DATA = (SERVER = DEDICATED) (SERVICE_NAME = oracle) ) ) to ORACLE.IDC.ORACLE.COM = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = xyzac)(PORT = 1521)) ) (CONNECT_DATA = (SERVER = SHARED) (SERVICE_NAME = Oracle) ) ) Change SERVER=SHARED. 5. Shutdown and startup the database. 6. Make a new connection to database other than SYSDBA. (NOTE: SYSDBA will always acquire dedicated connection by default.) 7. Check whether the connection is done through server server. > Select server from v$session. SERVER --------- DEDICATED DEDICATED DEDICATED SHARED DEDICATED NOTE: ==== The following parameters are optional (if not specified, Oracle selects defaults): MAX_DISPATCHERS: =============== Specifies the maximum number of dispatcher processes that can run simultaneously. SHARED_SERVERS: ============== Specifies the number of shared server processes created when an instance is started up. MAX_SHARED_SERVERS: ================== Specifies the maximum number of shared server processes that can run simultaneously. CIRCUITS: ======== Specifies the total number of virtual circuits that are available for inbound and outbound network sessions. SHARED_SERVER_SESSIONS: ====================== Specifies the total number of shared server user sessions to allow. Setting this parameter enables you to reserve user sessions for dedicated servers. Other parameters affected by shared server that may require adjustment: LARGE_POOL_SIZE: =============== Specifies the size in bytes of the large pool allocation heap. Shared server may force the default value to be set too high, causing performance problems or problems starting the database. SESSIONS: ======== Specifies the maximum number of sessions that can be created in the system. May need to be adjusted for shared server. 12.7 password for the listener: ------------------------------- Note 1: LSNRCTL> set password where is the password you want to use. To change a password, use "Change_Password" You can also designate a password when you configure the listener with the Net8 Assistant. These passwords are stored in the listener.ora file and although they will not show in the Net8 Assistant, they are readable in the listener.ora file. Note 2: The password can be set either by specifying it through the command CHANGE_PASSWORD, or through a parameter in the listener.ora file. We saw how to do that through the CHANGE_PASSWORD command earlier. If the password is changed this way, it should not be specified in the listener.ora file. The password is not displayed anywhere. When supplying the password in the listener control utility, you must supply it at the Password: prompt as shown above. You cannot specify the password in one line as shown below. LSNRCTL> set password t0p53cr3t LSNRCTL> stop Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=IPC)(KEY=EXTPROC))) TNS-01169: The listener has not recognized the password LSNRCTL> Note 3: more correct method would be to password protect the listener functions. See the net8 admin guide for info but in short -- you can: LSNRCTL> change_password Old password: New password: Reenter new password: Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=slackdog)(PORT=1521))) Password changed for LISTENER The command completed successfully LSNRCTL> set password Password: The command completed successfully LSNRCTL> save_config Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=slackdog)(PORT=1521))) Saved LISTENER configuration parameters. Listener Parameter File /d01/home/oracle8i/network/admin/listener.ora Old Parameter File /d01/home/oracle8i/network/admin/listener.bak The command completed successfully LSNRCTL> Now, you need to use a password to do various operations (such as STOP) but not others (such as STATUS) ============================================= 13. Datadictionary queries Rollback segments: ============================================= 13.1 naam, plaats en status van rollback segementen: ---------------------------------------------------- SELECT substr(segment_name, 1, 10), substr(tablespace_name, 1, 20), status, INITIAL_EXTENT, NEXT_EXTENT, MIN_EXTENTS, MAX_EXTENTS, PCT_INCREASE FROM DBA_ROLLBACK_SEGS; 13.2 indruk van aantal active transactions per rollback segment: ---------------------------------------------------------------- aantal actieve transacties: V$ROLLSTAT naam rollback segment: V$ROLLNAME SELECT n.name, s.xacts FROM V$ROLLNAME n, V$ROLLSTAT s WHERE n.usn=s.usn; (usn=undo segment number) 13.3 grootte, naam, extents, bytes van de rollback segmenten: ------------------------------------------------------------- SELECT substr(segment_name, 1, 15), bytes/1024/1024 Size_in_MB, blocks, extents, substr(tablespace_name, 1, 15) FROM DBA_SEGMENTS WHERE segment_type='ROLLBACK'; SELECT n.name, s.extents, s.rssize FROM V$ROLLNAME n, V$ROLLSTAT s WHERE n.usn=s.usn; Create Tablespace RBS datafile '/db1/oradata/oem/rbs.dbf' SIZE 200M AUTOEXTEND ON NEXT 20M MAXSIZE 500M LOGGING DEFAULT STORAGE ( INITIAL 5M NEXT 5M MINEXTENTS 2 MAXEXTENTS 100 PCTINCREASE 0 ) ONLINE PERMANENT; 13.4 De optimal parameter: -------------------------- SELECT n.name, s.optsize FROM V$ROLLNAME n, V$ROLLSTAT s WHERE n.usn=s.usn; 13.5 writes to rollback segementen: ----------------------------------- Doe de query begin meting, en bij einde meting en bekijk het verschil SELECT n.name, s.writes FROM V$ROLLNAME n, V$ROLLSTAT s WHERE n.usn=s.usn 13.6 Wie en welke processes gebruiken de rollback segs: ------------------------------------------------------- Query1: Query op v$lock, v$session, v$rollname column rr heading 'RB Segment' format a15 column us heading 'Username' format a10 column os heading 'OS user' format a10 column te heading 'Terminal' format a15 SELECT R.name rr, nvl(S.username, 'no transaction') us, S.Osuser os, S.Terminal te FROM V$LOCK L, V$SESSION S, V$ROLLNAME R WHERE L.Sid=S.Sid(+) AND trunc(L.Id1/65536)=R.usn AND L.Type='TX' AND L.Lmode=6 ORDER BY R.name / Query 2: SELECT r.name "RBS", s.sid, s.serial#, s.username "USER", t.status, t.cr_get, t.phy_io, t.used_ublk, t.noundo, substr(s.program, 1, 78) "COMMAND" FROM sys.v_$session s, sys.v_$transaction t, sys.v_$rollname r WHERE t.addr = s.taddr AND t.xidusn = r.usn ORDER BY t.cr_get, t.phy_io / 13.7 Bepaling minimum aantal rollbacksegmenten: ------------------------------------------------ Bepaal in init.ora via "show parameter transactions" transactions= a (max no of transactions, stel 100) transactions_per_rollback_segment= b (allowed no of concurrent tr/rbs, stel 10) minimum=a/b (100/10=10) 13.8 Bepaling minimale grootte rollback segmenten: -------------------------------------------------- lts=largest transaction size (normal production, niet af en toe batch loads) min_size=minimum size van rollback segment min_size= lts * 100 / (100 - (40 {%free} + 15 {iaiu} +5 {header} min_size=lts * 1.67 Stel lts=700K, dan is de startwaarde rollbacksegment=1400K ========================================================= 14. Data dictionary queries m.b.t. security, permissions: ========================================================= 14.1 user information in datadictionary --------------------------------------- SELECT username, user_id, password FROM DBA_USERS WHERE username='Kees'; 14.2 default tablespace, account_status of users ------------------------------------------------ SELECT username, default_tablespace, account_status FROM DBA_USERS; 14.3 tablespace quotas of users ------------------------------- SELECT tablespace_name, bytes, max_bytes, blocks, max_blocks FROM DBA_TS_QUOTAS WHERE username='CHARLIE'; 14.4 Systeem rechten van een user opvragen: DBA_SYS_PRIVS --------------------------------------------------------- SELECT substr(grantee, 1, 15), substr(privilege, 1, 40), admin_option FROM DBA_SYS_PRIVS WHERE grantee='CHARLIE'; SELECT * FROM dba_sys_privs WHERE grantee='Kees'; 14.5 Invalid objects in DBA_OBJECTS: ------------------------------------ SELECT substr(owner, 1, 10), substr(object_name, 1, 40), substr(object_type, 1, 40), status FROM DBA_OBJECTS WHERE status='INVALID'; 14.6 session information ------------------------ SELECT sid, serial#, substr(username, 1, 10), substr(osuser, 1, 10), substr(schemaname, 1, 10), substr(program, 1, 15), substr(module, 1, 15), status, logon_time, substr(terminal, 1, 15), substr(machine, 1, 15) FROM V$SESSION; 14.7 kill a session ------------------- alter system kill session 'SID, SERIAL#' ======================== 15. INIT.ORA parameters: ======================== 15.1 init.ora parameters en ARCHIVE MODE: ---------------------------------------- LOG_ARCHIVE_DEST=/oracle/admin/cc1/arch LOG_ARCHIVE_START=TRUE LOG_ARCHIVE_FORMAT=archcc1_%s.log 10g: LOG_ARCHIVE_DEST=c:\oracle\oradata\log' LOG_ARCHIVE_FORMAT=arch_%t_%s_%r.dbf' other: LOG_ARCHIVE_DEST_1= LOG_ARCHIVE_DEST_2= LOG_ARCHIVE_MAX_PROCESSES=2 15.2 init.ora en perfoRMANce en SGA: ------------------------------------ SORT_AREA_SIZE = 65536 (per PGA, max sort area) SORT_AREA_RETAINED_SIZE = 65536 (size after sort) PROCESSES = 100 (alle processes) DB_BLOCK_SIZE = 8192 DB_BLOCK_BUFFERS = 3400 (DB_CACHE_SIZE in Oracle 9i) SHARED_POOL_SIZE = 52428800 LOG_BUFFER = 26214400 LARGE_POOL_SIZE = DBWR_IO_SLAVES (DB_WRITER_PROCESSES) DB_WRITER_PROCESSES = 2 LGWR_IO_SLAVES= DB_FILE_MULTIBLOCK_READ_COUNT =16 (minimize io during table scans, it specifies max number of blocks in one io operation during sequential read) BUFFER_POOL_RECYCLE = BUFFER_POOL_KEEP = TIMED_STATISTICES =TRUE (statistics related to time are collected or not) OPTIMIZER_MODE =RULE, CHOOSE, FIRST_ROWS, ALL_ROWS PARALLEL_MIN_SERVERS = 2 (voor Parallel Query, en parallel recovery) PARALLEL_MAX_SERVERS = 4 RECOVERY_PARALLELISM = 2 (set parallel recovery op database niveau) SHARED_POOL_SIZE: in bytes or K or M SHARED_POOL_SIZE specifies (in bytes) the size of the shared pool. The shared pool contains shared cursors, stored procedures, control structures, and other structures. If you set PARALLEL_AUTOMATIC_TUNING to false, Oracle also allocates parallel execution message buffers from the shared pool. Larger values improve perfoRMANce in multi-user systems. Smaller values use less memory. You can monitor utilization of the shared pool by querying the view V$SGASTAT. SHARED_POOL_RESERVED_SIZE: The parameter was introduced in Oracle 7.1.5 and provides a means of reserving a portion of the shared pool for large memory allocations. The reserved area comes out of the shared pool itself. From a practical point of view one should set SHARED_POOL_RESERVED_SIZE to about 10% of SHARED_POOL_SIZE unless either the shared pool is very large OR SHARED_POOL_RESERVED_MIN_ALLOC has been set lower than the default value: 15.3 init.ora en jobs: ---------------------- JOB_QUEUE_PROCESSES=1 aantal SNP processes (SNP0, SNP1), max 36 t.b.v. replication en jobqueue's JOB_QUEUE_INTERVAL=60 check interval 15.4 instance name, sid: ------------------------ db_name = CC1 global_names = TRUE instance_name = CC1 db_domain = antapex.net 15.5 overige parameters: ------------------------ OS_AUTHENT_PREFIX = "" (stANDaard is dat OPS$) REMOTE_OS_AUTHENTICATION = TRUE or FALSE (of een OS authentication via het netwerk kan) REMOTE_LOGIN_PASSWORDFILEe = NONE or EXCLUSIVE distributed_transactions =0 or >0 (starts the RECO process) aq_tm_processes = (advanced queuing, message queues) mts_servers = (number of shared server processes in multithreaded server) mts_max_servers = audit_file_dest = /dbs01/app/oracle/admin/AMI_PRD/adump background_dump_dest = /dbs01/app/oracle/admin/AMI_PRD/bdump user_dump_dest = /dbs01/app/oracle/admin/AMI_PRD/udump core_dump_dest = /dbs01/app/oracle/admin/AMI_PRD/cdump resource_limit =true (specifies whether resource limits in profiles are in effect) license_max_sessions = (max number of concurrent user sessions) license_sessions_warning = (at this limit, warning in alert log) license_max_users = (maximum number of users that can be created in the database) are enforced) compatible = 8.1.7.0.0 control_files = /dbs04/oradata/AMI_PRD/ctrl/cc1_01.ctl control_files = /dbs05/oradata/AMI_PRD/ctrl/cc1_02.ctl control_files = /dbs06/oradata/AMI_PRD/ctrl/cc1_03.ctl db_files = 150 (max number of data files opened) java_pool_size = 0 log_checkpoint_interval = 10000 log_checkpoint_timeout = 1800 max_dump_file_size = 10240 max_enabled_roles = 40 nls_date_format = "DD-MM-YYYY" nls_language = AMERICAN nls_territory = AMERICA o7_dictionary_accessibility = TRUE open_cursors = 250 optimizer_max_permutations = 1000 optimizer_mode = CHOOSE parallel_max_servers = 5 pre_page_sga = TRUE service_names = CC1 utl_file_dir = /app01/oradata/cc1/utl_file All init.ora parameters: ------------------------- PARAMETER DESCRIPTION ------------------------------ ---------------------------------------- O7_DICTIONARY_ACCESSIBILITY Version 7 Dictionary Accessibility support [TRUE | FALSE] active_instance_count Number of active instances in the cluster database [NUMBER] aq_tm_processes Number of AQ Time Managers to start [NUMBER] archive_lag_target Maximum number of seconds of redos the standby could lose [NUMBER] asm_diskgroups Disk groups to mount automatically [CHAR] asm_diskstring Disk set locations for discovery [CHAR] asm_power_limit Number of processes for disk rebalancing [NUMBER] audit_file_dest Directory in which auditing files are to reside ['Path'] audit_sys_operations Enable sys auditing [TRUE|FALSE] audit_trail Enable system auditing [NONE|DB|DB_EXTENDED|OS] background_core_dump Core Size for Background Processes [partial | full] background_dump_dest Detached process dump directory [file_path] backup_tape_io_slaves BACKUP Tape I/O slaves [TRUE | FALSE] bitmap_merge_area_size Maximum memory allow for BITMAP MERGE [NUMBER] blank_trimming Blank trimming semantics parameter [TRUE | FALSE] buffer_pool_keep Number of database blocks/latches in keep buffer pool [CHAR: (buffers:n, latches:m)] buffer_pool_recycle Number of database blocks/latches in recycle buffer pool [CHAR: (buffers:n, latches:m)] circuits Max number of virtual circuits [NUMBER] cluster_database If TRUE startup in cluster database mode [TRUE | FALSE] cluster_database_instances Number of instances to use for sizing cluster db SGA structures [NUMBER] cluster_interconnects Interconnects for RAC use [CHAR] commit_point_strength Bias this node has toward not preparing in a two-phase commit [NUMBER (0-255)] compatible Database will be completely compatible with this software version [CHAR: 9.2.0.0.0] control_file_record_keep_time Control file record keep time in days [NUMBER] control_files Control file names list [file_path,file_path..] core_dump_dest Core dump directory [file_path] cpu_count Initial number of cpu's for this instance [NUMBER] create_bitmap_area_size Size of create bitmap buffer for bitmap index [INTEGER] cursor_sharing Cursor sharing mode [EXACT | SIMILAR | FORCE] create_stored_outlines Create stored outlines for DML statements [TRUE | FALSE | category_name] cursor_space_for_time Use more memory in order to get faster execution [TRUE | FALSE] db_16k_cache_size Size of cache for 16K buffers [bytes] db_2k_cache_size Size of cache for 2K buffers [bytes] db_32k_cache_size Size of cache for 32K buffers [bytes] db_4k_cache_size Size of cache for 4K buffers [bytes] db_8k_cache_size Size of cache for 8K buffers [bytes] db_block_buffers Number of database blocks to cache in memory [bytes: 8M or NUMBER of blocks (Ora7)] db_block_checking Data and index block checking [TRUE | FALSE] db_block_checksum Store checksum in db blocks and check during reads [TRUE | FALSE] db_block_size Size of database block [bytes] db_cache_advice Buffer cache sizing advisory [internal use only] db_cache_size Size of DEFAULT buffer pool for standard block size buffers [bytes] db_create_file_dest Default database location ['Path_to_directory'] db_create_online_log_dest_n Online log/controlfile destination (where n=1-5) ['Path'] db_domain Directory part of global database name stored with CREATE DATABASE [CHAR] * db_file_multiblock_read_count Db blocks to be read each IO [NUMBER] db_file_name_convert Datafile name convert patterns and strings for standby/clone db [, ] db_files Max allowable # db files [NUMBER] db_flashback_retention_target Maximum Flashback Database log retention time in minutes [NUMBER] db_keep_cache_size Size of KEEP buffer pool for standard block size buffers [bytes] db_name Database name specified in CREATE DATABASE [CHAR] db_recovery_file_dest Default database recovery file location [CHAR] db_recovery_file_dest_size Database recovery files size limit [bytes] db_recycle_cache_size Size of RECYCLE buffer pool for standard block size buffers [bytes] db_unique_name Database Unique Name [CHAR] db_writer_processes Number of background database writer processes to start [NUMBER] dblink_encrypt_login Enforce password for distributed login always be encrypted [TRUE | FALSE] dbwr_io_slaves DBWR I/O slaves [NUMBER] ddl_wait_for_locks Disable NOWAIT DML lock acquisitions [TRUE | FALSE] dg_broker_config_file1 Data guard broker configuration file #1 ['Path'] dg_broker_config_file2 Data guard broker configuration file #2 ['Path'] dg_broker_start Start Data Guard broker framework (DMON process) [TRUE | FALSE] disk_asynch_io Use asynch I/O for random access devices [TRUE | FALSE] dispatchers Specifications of dispatchers (MTS_dispatchers in Ora 8) [CHAR] distributed_lock_timeout Number of seconds a distributed transaction waits for a lock [Internal] dml_locks Dml locks - one for each table modified in a transaction [NUMBER] drs_start Start DG Broker monitor (DMON process)[TRUE | FALSE] enqueue_resources Resources for enqueues [NUMBER] event Debug event control - default null string [CHAR] fal_client FAL client [CHAR] fal_server FAL server list [CHAR] fast_start_io_target Upper bound on recovery reads [NUMBER] fast_start_mttr_target MTTR target of forward crash recovery in seconds [NUMBER] fast_start_parallel_rollback Max number of parallel recovery slaves that may be used [LOW | HIGH | FALSE] file_mapping Enable file mapping [TRUE | FALSE] fileio_network_adapters Network Adapters for File I/O [CHAR] filesystemio_options IO operations on filesystem files [Internal] fixed_date Fix SYSDATE value for debugging[NONE or '2000_12_30_24_59_00'] gc_files_to_locks RAC/OPS - lock granularity number of global cache locks per file (DFS) [CHAR] gcs_server_processes Number of background gcs server processes to start [NUMBER] global_context_pool_size Global Application Context Pool Size in Bytes [bytes] global_names Enforce that database links have same name as remote database [TRUE | FALSE] hash_area_size Size of in-memory hash work area (Shared Server)[bytes] hash_join_enabled Enable/disable hash join (CBO) [TRUE | FALSE] hi_shared_memory_address SGA starting address (high order 32-bits on 64-bit platforms) [NUMBER] hs_autoregister Enable automatic server DD updates in HS agent self-registration [TRUE | FALSE] ifile Include file in init.ora ['path_to_file'] instance_groups List of instance group names [CHAR] instance_name Instance name supported by the instance [CHAR] instance_number Instance number [NUMBER] instance_type Type of instance to be executed RDBMS or Automated Storage Management [RDBMS | ASM] java_max_sessionspace_size Max allowed size in bytes of a Java sessionspace [bytes] java_pool_size Size in bytes of the Java pool [bytes] java_soft_sessionspace_limit Warning limit on size in bytes of a Java sessionspace [NUMBER] job_queue_processes Number of job queue slave processes [NUMBER] large_pool_size Size in bytes of the large allocation pool [bytes] ldap_directory_access RDBMS's LDAP access option [NONE | PASSWORD | SSL] license_max_sessions Maximum number of non-system user sessions (concurrent licensing) [NUMBER] license_max_users Maximum number of named users that can be created (named user licensing) [NUMBER] license_sessions_warning Warning level for number of non-system user sessions [NUMBER] local_listener Define which listeners instances register with [CHAR] lock_name_space Used for generating lock names for standby/primary database assign each a unique name space [CHAR] lock_sga Lock entire SGA in physical memory [Internal] log_archive_config Log archive config [SEND|NOSEND] [RECEIVE|NORECEIVE] [ DG_CONFIG] log_archive_dest Archive logs destination ['path_to_directory'] log_archive_dest_n Archive logging parameters (n=1-10) Enterprise Edition [CHAR] log_archive_dest_state_n Archive logging parameter status (n=1-10) [CHAR] Enterprise Edition [CHAR] log_archive_duplex_dest Duplex archival destination ['path_to_directory'] log_archive_format Archive log filename format [CHAR: "MyApp%S.ARC"] log_archive_local_first Establish EXPEDITE attribute default value [TRUE | FALSE] log_archive_max_processes Maximum number of active ARCH processes [NUMBER] log_archive_min_succeed_dest Minimum number of archive destinations that must succeed [NUMBER] log_archive_start Start archival process on SGA initialization [TRUE | FALSE] log_archive_trace Archive log tracing level [NUMBER] log_buffer Redo circular buffer size [bytes] log_checkpoint_interval Checkpoint threshold, # redo blocks [NUMBER] log_checkpoint_timeout Checkpoint threshold, maximum time interval between checkpoints in seconds [NUMBER] log_checkpoints_to_alert Log checkpoint begin/end to alert file [TRUE | FALSE] log_file_name_convert Logfile name convert patterns and strings for standby/clone db [, ] log_parallelism Number of log buffer strands [NUMBER] logmnr_max_persistent_sessions Maximum number of threads to mine [NUMBER] max_commit_propagation_delay Max age of new snapshot in .01 seconds [NUMBER] max_dispatchers Max number of dispatchers [NUMBER] max_dump_file_size Maximum size (blocks) of dump file [UNLIMITED or bytes] max_enabled_roles Max number of roles a user can have enabled [NUMBER] max_rollback_segments Max number of rollback segments in SGA cache [NUMBER] max_shared_servers Max number of shared servers [NUMBER] mts_circuits Max number of circuits [NUMBER] mts_dispatchers Specifications of dispatchers [CHAR] mts_listener_address Address(es) of network listener [CHAR] mts_max_dispatchers Max number of dispatchers [NUMBER] mts_max_servers Max number of shared servers [NUMBER] mts_multiple_listeners Are multiple listeners enabled? [TRUE | FALSE] mts_servers Number of shared servers to start up [NUMBER] mts_service Service supported by dispatchers [CHAR] mts_sessions max number of shared server sessions [NUMBER] nls_calendar NLS calendar system name (Default=GREGORIAN) [CHAR] nls_comp NLS comparison, Enterprise Edition [BINARY | ANSI] nls_currency NLS local currency symbol [CHAR] nls_date_format NLS Oracle date format [CHAR] nls_date_language NLS date language name (Default=AMERICAN) [CHAR] nls_dual_currency Dual currency symbol [CHAR] nls_iso_currency NLS ISO currency territory name override the default set by NLS_TERRITORY [CHAR] nls_language NLS language name (session default) [CHAR] nls_length_semantics Create columns using byte or char semantics by default [BYTE | CHAR] nls_nchar_conv_excp NLS raise an exception instead of allowing implicit conversion [CHAR] nls_numeric_characters NLS numeric characters [CHAR] nls_sort Case-sensitive or insensitive sort [Language] language may be BINARY, BINARY_CI, BINARY_AI, GERMAN, GERMAN_CI, etc nls_territory NLS territory name (country settings) [CHAR] nls_time_format Time format [CHAR] nls_time_tz_format Time with timezone format [CHAR] nls_timestamp_format Time stamp format [CHAR] nls_timestamp_tz_format Timestamp with timezone format [CHAR] object_cache_max_size_percent Percentage of maximum size over optimal of the user session's ob [NUMBER] object_cache_optimal_size Optimal size of the user session's object cache in bytes [bytes] olap_page_pool_size Size of the olap page pool in bytes [bytes] open_cursors Max # cursors per session [NUMBER] open_links Max # open links per session [NUMBER] open_links_per_instance Max # open links per instance [NUMBER] optimizer_dynamic_sampling Optimizer dynamic sampling [NUMBER] optimizer_features_enable Optimizer plan compatibility (oracle version e.g. 8.1.7) [CHAR] optimizer_index_caching Optimizer index caching percent [NUMBER] optimizer_index_cost_adj Optimizer index cost adjustment [NUMBER] optimizer_max_permutations Optimizer maximum join permutations per query block [NUMBER] optimizer_mode Optimizer mode [RULE | CHOOSE | FIRST_ROWS | ALL_ROWS] oracle_trace_collection_name Oracle TRACE default collection name [CHAR] oracle_trace_collection_path Oracle TRACE collection path [CHAR] oracle_trace_collection_size Oracle TRACE collection file max. size [NUMBER] oracle_trace_enable Oracle Trace enabled/disabled [TRUE | FALSE] oracle_trace_facility_name Oracle TRACE default facility name [CHAR] oracle_trace_facility_path Oracle TRACE facility path [CHAR] os_authent_prefix Prefix for auto-logon accounts [CHAR] os_roles Retrieve roles from the operating system [TRUE | FALSE] parallel_adaptive_multi_user Enable adaptive setting of degree for multiple user streams [TRUE | FALSE] parallel_automatic_tuning Enable intelligent defaults for parallel execution parameters [TRUE | FALSE] parallel_execution_message_size Message buffer size for parallel execution [bytes] parallel_instance_group Instance group to use for all parallel operations [CHAR] parallel_max_servers Maximum parallel query servers per instance [NUMBER] parallel_min_percent Minimum percent of threads required for parallel query [NUMBER] parallel_min_servers Minimum parallel query servers per instance [NUMBER] parallel_server If TRUE startup in parallel server mode [TRUE | FALSE] parallel_server_instances Number of instances to use for sizing OPS SGA structures [NUMBER] parallel_threads_per_cpu Number of parallel execution threads per CPU [NUMBER] partition_view_enabled Enable/disable partitioned views [TRUE | FALSE] pga_aggregate_target Target size for the aggregate PGA memory consumed by the instance [bytes] plsql_code_type PL/SQL code-type [INTERPRETED | NATIVE] plsql_compiler_flags PL/SQL compiler flags [CHAR] plsql_debug PL/SQL debug [TRUE | FALSE] plsql_native_c_compiler plsql native C compiler [CHAR] plsql_native_library_dir plsql native library dir ['Path_to_directory'] plsql_native_library_subdir_count plsql native library number of subdirectories [NUMBER] plsql_native_linker plsql native linker [CHAR] plsql_native_make_file_name plsql native compilation make file [CHAR] plsql_native_make_utility plsql native compilation make utility [CHAR] plsql_optimize_level PL/SQL optimize level [NUMBER] plsql_v2_compatibility PL/SQL version 2.x compatibility flag [TRUE | FALSE] plsql_warnings PL/SQL compiler warnings settings [CHAR] See also DBMS_WARNING and DBA_PLSQL_OBJECT_SETTINGS pre_page_sga Pre-page sga for process [TRUE | FALSE] processes User processes [NUMBER] query_rewrite_enabled Allow rewrite of queries using materialized views if enabled [FORCE | TRUE | FALSE] query_rewrite_integrity Perform rewrite using materialized views with desired integrity [STALE_TOLERATED | TRUSTED | ENFORCED] rdbms_server_dn RDBMS's Distinguished Name [CHAR] read_only_open_delayed If TRUE delay opening of read only files until first access [TRUE | FALSE] recovery_parallelism Number of server processes to use for parallel recovery [NUMBER] remote_archive_enable Remote archival enable setting [RECEIVE[,SEND] | FALSE | TRUE] remote_dependencies_mode Remote-procedure-call dependencies mode parameter [TIMESTAMP | SIGNATURE] remote_listener Remote listener [CHAR] remote_login_passwordfile Use a password file [NONE | SHARED | EXCLUSIVE] remote_os_authent Allow non-secure remote clients to use auto-logon accounts [TRUE | FALSE] remote_os_roles Allow non-secure remote clients to use os roles [TRUE | FALSE] replication_dependency_tracking Tracking dependency for Replication parallel propagation [TRUE | FALSE] resource_limit Master switch for resource limit [TRUE | FALSE] resource_manager_plan Resource mgr top plan [Plan_Name] resumable_timeout Set resumable_timeout, seconds [NUMBER] rollback_segments Undo segment list [CHAR] row_locking Row-locking [ALWAYS | DEFAULT | INTENT] (Default=always) serial_reuse Reuse the frame segments [DISABLE | SELECT|DML|PLSQL|ALL|NULL] serializable Serializable [Internal] service_names Service names supported by the instance [CHAR] session_cached_cursors Number of cursors to save in the session cursor cache [NUMBER] session_max_open_files Maximum number of open files allowed per session [NUMBER] sessions User and system sessions [NUMBER] sga_max_size Max total SGA size [bytes] sga_target Target size of SGA [bytes] shadow_core_dump Core Size for Shadow Processes [PARTIAL | FULL | NONE] shared_memory_address SGA starting address (low order 32-bits on 64-bit platforms) [NUMBER] shared_pool_reserved_size Size in bytes of reserved area of shared pool [bytes] shared_pool_size Size in bytes of shared pool [bytes] shared_server_sessions Max number of shared server sessions [NUMBER] shared_servers Number of shared servers to start up [NUMBER] skip_unusable_indexes Skip unusable indexes if set to true [TRUE | FALSE] sort_area_retained_size Size of in-memory sort work area retained between fetch calls [bytes] sort_area_size Size of in-memory sort work area [bytes] smtp_out_server utl_smtp server and port configuration parameter [server_clause] spfile Server parameter file [CHAR] sp_name Service Provider Name [CHAR] sql92_security Require select privilege for searched update/delete [TRUE | FALSE] sql_trace Enable SQL trace [TRUE | FALSE] sqltune_category Category qualifier for applying hintsets [CHAR] sql_version Sql language version parameter for compatibility issues [CHAR] standby_archive_dest Standby database archivelog destination text string ['Path_to_directory'] standby_file_management If auto then files are created/dropped automatically on standby [MANUAL | AUTO] star_transformation_enabled Enable the use of star transformation [TRUE | FALSE | DISABLE_TEMP_TABLE] statistics_level Statistics level [ALL | TYPICAL | BASIC] streams_pool_size Size in bytes of the streams pool [bytes] tape_asynch_io Use asynch I/O requests for tape devices [TRUE | FALSE] thread Redo thread to mount [NUMBER] timed_os_statistics Internal os statistic gathering interval in seconds [NUMBER] timed_statistics Maintain internal timing statistics [TRUE | FALSE] trace_enabled Enable KST tracing (Internal parameter) [TRUE | FALSE] tracefile_identifier Trace file custom identifier [CHAR] transaction_auditing Transaction auditing records generated in the redo log [TRUE | FALSE] transactions Max. number of concurrent active transactions [NUMBER] transactions_per_rollback_segment Number of active transactions per rollback segment [NUMBER] undo_management Instance runs in SMU mode if TRUE, else in RBU mode [MANUAL | AUTO] undo_retention Undo retention in seconds [NUMBER] undo_suppress_errors Suppress RBU errors in SMU mode [TRUE | FALSE] undo_tablespace Use or switch undo tablespace [Undo_tbsp_name] use_indirect_data_buffers Enable indirect data buffers (very large SGA on 32-bit platforms [TRUE | FALSE] user_dump_dest User process dump directory ['Path_to_directory'] utl_file_dir utl_file accessible directories list utl_file_dir='Path1', 'Path2'.. or utl_file_dir='Path1' # Must be utl_file_dir='Path2' # consecutive entries workarea_size_policy Policy used to size SQL working areas [MANUAL | AUTO] db_file_multiblock_read_count: The db_file_multiblock_read_count initialization parameter determines the maximum number of database blocks read in one I/O operation during a full table scan. The setting of this parameter can reduce the number of I/O calls required for a full table scan, thus improving performance. 15.6 9i UNDO or ROLLBACK parameters: ------------------------------------ - UNDO_MANAGEMENT If AUTO, use automatic undo management mode. If MANUAL, use manual undo management mode. - UNDO_TABLESPACE A dynamic parameter specifying the name of an undo tablespace to use. - UNDO_RETENTION A dynamic parameter specifying the length of time to retain undo. Default is 900 seconds. - UNDO_SUPPRESS_ERRORS If TRUE, suppress error messages if manual undo management SQL statements are issued when operating in automatic undo management mode. If FALSE, issue error message. This is a dynamic parameter. If you're database is on manual, you can still use the following 8i type parameters: - ROLLBACK_SEGMENTS Specifies the rollback segments to be acquired at instance startup - TRANSACTIONS Specifies the maximum number of concurrent transactions - TRANSACTIONS_PER_ROLLBACK_SEGMENT Specifies the number of concurrent transactions that each rollback segment is expected to handle - MAX_ROLLBACK_SEGMENTS Specifies the maximum number of rollback segments that can be online for any instance 15.7 Oracle 9i init file examples: ---------------------------------= Example 1: ---------- # Cache and I/O DB_BLOCK_SIZE=4096 DB_CACHE_SIZE=20971520 # Cursors and Library Cache CURSOR_SHARING=SIMILAR OPEN_CURSORS=300 # Diagnostics and Statistics BACKGROUND_DUMP_DEST=/vobs/oracle/admin/mynewdb/bdump CORE_DUMP_DEST=/vobs/oracle/admin/mynewdb/cdump TIMED_STATISTICS=TRUE USER_DUMP_DEST=/vobs/oracle/admin/mynewdb/udump # Control File Configuration CONTROL_FILES=("/vobs/oracle/oradata/mynewdb/control01.ctl", "/vobs/oracle/oradata/mynewdb/control02.ctl", "/vobs/oracle/oradata/mynewdb/control03.ctl") # Archive LOG_ARCHIVE_DEST_1='LOCATION=/vobs/oracle/oradata/mynewdb/archive' LOG_ARCHIVE_FORMAT=%t_%s.dbf LOG_ARCHIVE_START=TRUE # Shared Server # Uncomment and use first DISPATCHES parameter below when your listener is # configured for SSL # (listener.ora and sqlnet.ora) # DISPATCHERS = "(PROTOCOL=TCPS)(SER=MODOSE)", # "(PROTOCOL=TCPS)(PRE=oracle.aurora.server.SGiopServer)" DISPATCHERS="(PROTOCOL=TCP)(SER=MODOSE)", "(PROTOCOL=TCP)(PRE=oracle.aurora.server.SGiopServer)", (PROTOCOL=TCP) # Miscellaneous COMPATIBLE=9.2.0 DB_NAME=mynewdb # Distributed, Replication and Snapshot DB_DOMAIN=us.oracle.com REMOTE_LOGIN_PASSWORDFILE=EXCLUSIVE # Network Registration INSTANCE_NAME=mynewdb # Pools JAVA_POOL_SIZE=31457280 LARGE_POOL_SIZE=1048576 SHARED_POOL_SIZE=52428800 # Processes and Sessions PROCESSES=150 # Redo Log and Recovery FAST_START_MTTR_TARGET=300 # Resource Manager RESOURCE_MANAGER_PLAN=SYSTEM_PLAN # Sort, Hash Joins, Bitmap Indexes SORT_AREA_SIZE=524288 # Automatic Undo Management UNDO_MANAGEMENT=AUTO UNDO_TABLESPACE=undotbs Example 2: ---------- ############################################################################## # Copyright (c) 1991, 2001 by Oracle Corporation ############################################################################## ########################################### # Cache and I/O ########################################### db_block_size=8192 db_cache_size=50331648 ########################################### # Cursors and Library Cache ########################################### open_cursors=300 ########################################### # Diagnostics and Statistics ########################################### background_dump_dest=D:\oracle\admin\iasdb\bdump core_dump_dest=D:\oracle\admin\iasdb\cdump timed_statistics=TRUE user_dump_dest=D:\oracle\admin\iasdb\udump ########################################### # Distributed, Replication and Snapshot ########################################### db_domain=missrv.miskm.mindef.nl remote_login_passwordfile=EXCLUSIVE ########################################### # File Configuration ########################################### control_files=("D:\oracle\oradata\iasdb\CONTROL01.CTL", "D:\oracle\oradata\iasdb\CONTROL02.CTL", "D:\oracle\oradata\iasdb\CONTROL03.CTL") ########################################### # Job Queues ########################################### job_queue_processes=4 ########################################### # MTS ########################################### dispatchers="(PROTOCOL=TCP)(PRE=oracle.aurora.server.GiopServer)", "(PROTOCOL=TCP)(PRE=oracle.aurora.server.SGiopServer)" ########################################### # Miscellaneous ########################################### aq_tm_processes=1 compatible=9.0.0 db_name=iasdb ########################################### # Network Registration ########################################### instance_name=iasdb ########################################### # Pools ########################################### java_pool_size=41943040 shared_pool_size=33554432 ########################################### # Processes and Sessions ########################################### processes=150 ########################################### # Redo Log and Recovery ########################################### fast_start_mttr_target=300 ########################################### # Sort, Hash Joins, Bitmap Indexes ########################################### pga_aggregate_target=33554432 sort_area_size=524288 ########################################### # System Managed Undo and Rollback Segments ########################################### undo_management=AUTO undo_tablespace=UNDOTBS ============== 17. Snapshots: ============== Snapshots allow you to replicate data based on column- and/or row-level subsetting, while multimaster replication requires replication of the entire table. You need a database link to implement replication. 17.1 Database link: ------------------- In de "local" database, waar de snapshot copy komt te staan, geef een statement als bijv: CREATE PUBLIC DATABASE LINK MY_LINK CONNECT TO HARRY IDENTIFIED BY password USING 'DB1'; De servicename "DB1" wordt via de tnsnames.ora geresolved in een connectdescriptor, waarin de remote Servername, protocol, en SID van de remote database bekend is geworden. Nu is het mogelijk om bijv. de table employee in de remote database "DB1" te SELECTeren: SELECT * FROM employee@MY_LINK; Ook 2PC is geimplementeerd: update employee set amount=amount-100; update employee@my_link set amount=amount+100; commit; 17.2 Snapshots: --------------- There are in general 2 styles of snapshots available Simple snapshot: One to one replication of a remote table to a local snapshot (=table). The refresh of the snapshot can be a complete refresh, with the refresh rate specified in the "create snapshot" command. Also a snapshot log can be used at the remote original table in order to replicate only the transaction data. Complex snapshot: If multiple remote tables are joined in order to create/refresh a local snapshot, it is a "complex snapshot". Only complete refreshes are possible. If joins or complex query clauses are used, like group by, one can only use a "complex snapshot". -> Example COMPLEX snapshot: On the local database: CREATE SNAPSHOT EMP_DEPT_COUNT pctfree 5 tablespace SNAP storage (initial 100K next 100K pctincrease 0) REFRESH COMPLETE START WITH SYSDATE NEXT SYSDATE+7 AS SELECT DEPTNO, COUNT(*) Dept_count FROM EMPLOYEE@MY_LINK GROUP BY Deptno; Because the records in this snapshot will not correspond one to one with the records in the master table (since the query contains a group by clause) this is a complex snapshot. Thus the snapshot will be completely recreated every time it is refreshed. -> Example SIMPLE snapshot: On the local database: CREATE SNAPSHOT EMP_DEPT_COUNT pctfree 5 tablespace SNAP storage (initial 100K next 100K pctincrease 0) REFRESH FAST START WITH SYSDATE NEXT SYSDATE+7 AS SELECT * FROM EMPLOYEE@MY_LINK In this case the refresh fast clause tells oracle to use a snapshot log to refresh the local snapshot. When a snapshotlog is used, only the changes to the master table are sent to the targets. The snapshot log must be created in the master database (WHERE the original object is) create snapshot log on employee tablespace data storage (initial 100K next 100K pctincrease 0); Snapshot groups: ---------------- A snapshot group in a replication system maintains a partial or complete copy of the objects at the target master group. Snapshot groups cannot span master group boundaries. Figure 3-7 displays the correlation between Groups A and B at the master site and Groups A and B at the snapshot site. Group A at the snapshot site (see Figure 3-7) contains only some of the objects in the corresponding Group A at the master site. Group B at the snapshot site contains all objects in Group B at the master site. Under no circumstances, however, could Group B at the snapshot site contain objects FROM Group A at the master site. As illustrated in Figure 3-7, a snapshot group has the same name as the master group on which the snapshot group is based. For example, a snapshot group based on a "PERSONNEL" master group is also named "PERSONNEL." In addition to maintaining organizational consistency between snapshot sites and master sites, snapshot groups are required for supporting updateable snapshots. If a snapshot does not belong to a snapshot group, then it must be a read-only snapshot. A snapshot group is used to organize snapshots in a logical manner. Refresh groups: --------------- If 2 or more master tables which have a PK-FK relationship, are replicated, it is possible'that the 2 cooresponding snapshots violate the referential integrety, because of different refresh times and schedules etc.. Related snapshots can be collected int refresh groups. The purpose of a refresh group is to coordinate the refresh schedules of it's members. This is achieved via the DBMS_REFRESH package. The procedures in this package are MAKE, ADD, SUBSTRACT, CHANGE, DESTROY, and REFRESH A refresh group could contain more than one snapshot groups. Types of snapshots: ------------------- Primary Key ----------- Primary key snapshots are the default type of snapshot. They are updateable if the snapshot was created as part of a snapshot group and "FOR UPDATE" was specified when defining the snapshot. Changes are propagated according to the row-level changes that have occurred, as identified by the primary key value of the row (not the ROWID). The SQL statement for creating an updateable, primary key snapshot might look like: CREATE SNAPSHOT sales.customer FOR UPDATE AS SELECT * FROM sales.customer@dbs1.acme.com; Primary key snapshots may contain a subquery so that you can create a horizontally partitioned subset of data at the remote snapshot site. This subquery may be as simple as a basic WHERE clause or as complex as a multilevel WHERE EXISTS clause. Primary key snapshots that contain a SELECTed class of subqueries can still be incrementally or fast refreshed. The following is a subquery snapshot with a WHERE clause containing a subquery: CREATE SNAPSHOT sales.orders REFRESH FAST AS SELECT * FROM sales.orders@dbs1.acme.com o WHERE EXISTS (SELECT 1 FROM sales.customer@dbs1.acme.com c WHERE o.c_id = c.c_id AND zip = 19555); ROWID ----- For backwards compatibility, Oracle supports ROWID snapshots in addition to the default primary key snapshots. A ROWID snapshot is based on the physical row identifiers (ROWIDs) of the rows in a master table. ROWID snapshots should be used only for snapshots based on master tables FROM an Oracle7 database, and should not be used when creating new snapshots based on master tables FROM Oracle release 8.0 or greater databases. CREATE SNAPSHOT sales.customer REFRESH WITH ROWID AS SELECT * FROM sales.customer@dbs1.acme.com; Complex ------- To be fast refreshed, the defining query for a snapshot must observe certain restrictions. If you require a snapshot whose defining query is more general and cannot observe the restrictions, then the snapshot is complex and cannot be fast refreshed. Specifically, a snapshot is considered complex when the defining query of the snapshot contains: A CONNECT BY clause Clauses that do not comply with the requirements detailed in Table 3-1, "Restrictions for Snapshots with Subqueries" A set operation, such as UNION, INTERSECT, or MINUS In most cases, a distinct or aggregate function, although it is possible to have a distinct or aggregate function in the defining query and still have a simple snapshot See Also: Oracle8i Data Warehousing Guide for more information about complex materialized views. "Snapshot" is synonymous with "materialized view" in Oracle documentation, and "materialized view" is used in the Oracle8i Data Warehousing Guide. The following statement is an example of a complex snapshot CREATE statement: CREATE SNAPSHOT scott.snap_employees AS SELECT emp.empno, emp.ename FROM scott.emp@dbs1.acme.com UNION ALL SELECT new_emp.empno, new_emp.ename FROM scott.new_emp@dbs1.acme.com; Read Only --------- Any of the previously described types of snapshots can be made read-only by omitting the FOR UPDATE clause or disabling the equivalent checkbox in the Replication Manager interface. Read-only snapshots use many of the same mechanisms as updateable snapshots, except that they do not need to belong to a snapshot group. Snapshot Registration at a Master Site -------------------------------------- At the master site, an Oracle database automatically registers information about a snapshots based on its master table(s). The following sections explain more about Oracle's snapshot registration mechanism. DBA_REGISTERED_SNAPSHOTS and DBA_SNAPSHOT_REFRESH_TIMES dictionary views You can query the DBA_REGISTERED_SNAPSHOTS data dictionary view to list the following information about a remote snapshot: The owner, name, and database that contains the snapshot The snapshot's defining query Other snapshot characteristics, such as its refresh method (fast or complete) You can also query the DBA_SNAPSHOT_REFRESH_TIMES view at the master site to obtain the last refresh times for each snapshot. Administrators can use this information to monitor snapshot activity FROM master sites and coordinate changes to snapshot sites if a master table needs to be dropped, altered, or relocated. Internal Mechanisms Oracle automatically registers a snapshot at its master database when you create the snapshot, and unregisters the snapshot when you drop it. Caution: Oracle cannot guarantee the registration or unregistration of a snapshot at its master site during the creation or drop of the snapshot, respectively. If Oracle cannot successfully register a snapshot during creation, Oracle completes snapshot registration during a subsequent refresh of the snapshot. If Oracle cannot successfully unregister a snapshot when you drop the snapshot, the registration information for the snapshot persists in the master database until it is manually unregistered. Complex snapshots might not be registered. Manual registration ------------------- If necessary, you can maintain registration manually. Use the REGISTER_SNAPSHOT and UNREGISTER_SNAPSHOT procedures of the DBMS_SNAPSHOT package at the master site to add, modify, or remove snapshot registration information. Snapshot Log ------------ When you create a snapshot log for a master table, Oracle creates an underlying table as the snapshot log. A snapshot log holds the primary keys and/or the ROWIDs of rows that have been updated in the master table. A snapshot log can also contain filter columns to support fast refreshes of snapshots with subqueries. The name of a snapshot log's table is MLOG$_master_table_name. The snapshot log is created in the same schema as the target master table. One snapshot log can support multiple snapshots on its master table. As described in the previous section, the internal trigger adds change information to the snapshot log whenever a DML transaction has taken place on the target master table. There are three types of snapshot logs: Primary Key: The snapshot records changes to the master table based on the primary key of the affected rows. Row ID: The snapshot records changes to the master table based on the ROWID of the affected rows. Combination: The snapshot records changes to the master table based on both the primary key and the ROWID of the affected rows. This snapshot log supports both primary key and ROWID snapshots, which is helpful for mixed environments. A combination snapshot log works in the same manner as the primary key and ROWID snapshot log, except that both the primary key and the ROWID of the affected row are recorded. Though the difference between snapshot logs based on primary keys and ROWIDs is small (one records affected rows using the primary key, while the other records affected rows using the physical ROWID), the practical impact is large. Using ROWID snapshots and snapshot logs makes reorganizing and truncating your master tables difficult because it prevents your ROWID snapshots FROM being fast refreshed. If you reorganize or truncate your master table, your ROWID snapshot must be COMPLETE refreshed because the ROWIDs of the master table have changed. To delete a snapshot log, execute the DROP SNAPSHOT LOG SQL statement in SQL*Plus. For example, the following statement deletes the snapshot log for a table named CUSTOMERS in the SALES schema: DROP SNAPSHOT LOG ON sales.customers; To delete the master table, use truncate table TABLE_NAME purge snapshot log; ============= 18. Triggers: ============= A trigger is PL/SQL code block attached and executed by an event which occurs to a database table. Triggers are implicitly invoked by DML commands. Triggers are stored as text and compiled at execute time, because of this it is wise not to include much code in them but to call out to previously stored procedures or packages as this will greatly improve perfoRMANce. You may not use COMMIT, ROLLBACK and SAVEPOINT statements within trigger blocks. Remember that triggers may be executed thousands of times for a large update - they can seriously affect SQL execution perfoRMANce. Triggers may be called BEFORE or AFTER the following events :- INSERT, UPDATE and DELETE. Triggers may be STATEMENT or ROW types. - STATEMENT triggers fire BEFORE or AFTER the execution of the statement that caused the trigger to fire. - ROW triggers fire BEFORE or AFTER any affected row is processed. An example of a statement trigger follows :- CREATE OR REPLACE TRIGGER MYTRIG1 BEFORE DELETE OR INSERT OR UPDATE ON JD11.BOOK BEGIN IF (TO_CHAR(SYSDATE,'DAY') IN ('sat','sun')) OR (TO_CHAR(SYSDATE,'hh24:mi') NOT BETWEEN '08:30' AND '18:30') THEN RAISE_APPLICATION_ERROR(-20500,'Table is secured'); END IF; END; After the CREATE OR REPLACE statement is the object identifier (TRIGGER) and the object name (MYTRIG1). This trigger specifies that before any data change event on the BOOK table this PL/SQL code block will be compiled and executed. The user will not be allowed to update the table outside of normal working hours. An example of a row trigger follows :- CREATE OR REPLACE TRIGGER MYTRIG2 AFTER DELETE OR INSERT OR UPDATE ON JD11.BOOK FOR EACH ROW BEGIN IF DELETING THEN INSERT INTO JD11.XBOOK (PREVISBN, TITLE, DELDATE) VALUES (:OLD.ISBN, :OLD.TITLE, SYSDATE); ELSIF INSERTING THEN INSERT INTO JD11.NBOOK (ISBN, TITLE, ADDDATE) VALUES (:NEW.ISBN, :NEW.TITLE, SYSDATE); ELSIF UPDATING ('ISBN) THEN INSERT INTO JD11.CBOOK (OLDISBN, NEWISBN, TITLE, UP_DATE) VALUES (:OLD.ISBN :NEW.ISBN, :NEW.TITLE, SYSDATE); ELSE /* UPDATE TO ANYTHING ELSE THAN ISBN */ INSERT INTO JD11.UBOOK (ISBN, TITLE, UP_DATE) VALUES (:OLD.ISBN :NEW.TITLE, SYSDATE); END IF END; In this case we have specified that the trigger will be executed after any data change event on any affected row. Within the PL/SQL block body we can check which update action is being performed for the currently affected row and take whatever action we feel is appropriate. Note that we can specify the old and new values of updated rows by prefixing column names with the :OLD and :NEW qualifiers. -------------------------------------------------------------------------------- The following statement creates a trigger for the Emp_tab table: CREATE OR REPLACE TRIGGER Print_salary_changes BEFORE DELETE OR INSERT OR UPDATE ON Emp_tab FOR EACH ROW WHEN (new.Empno > 0) DECLARE sal_diff number; BEGIN sal_diff := :new.sal - :old.sal; dbms_output.put('Old salary: ' || :old.sal); dbms_output.put(' New salary: ' || :new.sal); dbms_output.put_line(' Difference ' || sal_diff); END; / If you enter a SQL statement, such as the following: UPDATE Emp_tab SET sal = sal + 500.00 WHERE deptno = 10; Then, the trigger fires once for each row that is updated, and it prints the new and old salaries, and the difference. CREATE OR REPLACE TRIGGER "SALES".HENKILOROOLI_CHECK2 AFTER INSERT OR UPDATE OR DELETE ON AH_HENKILOROOLI BEGIN IF INSERTING OR DELETING THEN handle_delayed_triggers ('AH_HENKILOROOLI', 'HENKILOROOLI_CHECK'); END IF; IF INSERTING OR UPDATING OR DELETING THEN /* FE */ handle_delayed_triggers('AH_HENKILOROOLI', 'FRONTEND_FLAG'); /* FE */ END IF; /* FE */ END; A trigger is either a stored PL/SQL block or a PL/SQL, C, or Java procedure associated with a table, view, schema, or the database itself. Oracle automatically executes a trigger when a specified event takes place, which may be in the form of a system event or a DML statement being issued against the table. Triggers can be: -DML triggers on tables. -INSTEAD OF triggers on views. -System triggers on DATABASE or SCHEMA: With DATABASE, triggers fire for each event for all users; with SCHEMA, triggers fire for each event for that specific user. BEFORE and AFTER Options The BEFORE or AFTER option in the CREATE TRIGGER statement specifies exactly when to fire the trigger body in relation to the triggering statement that is being run. In a CREATE TRIGGER statement, the BEFORE or AFTER option is specified just before the triggering statement. For example, the PRINT_SALARY_CHANGES trigger in the previous example is a BEFORE trigger. INSTEAD OF Triggers The INSTEAD OF option can also be used in triggers. INSTEAD OF triggers provide a transparent way of modifying views that cannot be modified directly through UPDATE, INSERT, and DELETE statements. These triggers are called INSTEAD OF triggers because, unlike other types of triggers, Oracle fires the trigger instead of executing the triggering statement. The trigger performs UPDATE, INSERT, or DELETE operations directly on the underlying tables. CREATE TABLE Project_tab ( Prj_level NUMBER, Projno NUMBER, Resp_dept NUMBER); CREATE TABLE Emp_tab ( Empno NUMBER NOT NULL, Ename VARCHAR2(10), Job VARCHAR2(9), Mgr NUMBER(4), Hiredate DATE, Sal NUMBER(7,2), Comm NUMBER(7,2), Deptno NUMBER(2) NOT NULL); CREATE TABLE Dept_tab ( Deptno NUMBER(2) NOT NULL, Dname VARCHAR2(14), Loc VARCHAR2(13), Mgr_no NUMBER, Dept_type NUMBER); The following example shows an INSTEAD OF trigger for inserting rows into the MANAGER_INFO view. CREATE OR REPLACE VIEW manager_info AS SELECT e.ename, e.empno, d.dept_type, d.deptno, p.prj_level, p.projno FROM Emp_tab e, Dept_tab d, Project_tab p WHERE e.empno = d.mgr_no AND d.deptno = p.resp_dept; CREATE OR REPLACE TRIGGER manager_info_insert INSTEAD OF INSERT ON manager_info REFERENCING NEW AS n -- new manager information FOR EACH ROW DECLARE rowcnt number; BEGIN SELECT COUNT(*) INTO rowcnt FROM Emp_tab WHERE empno = :n.empno; IF rowcnt = 0 THEN INSERT INTO Emp_tab (empno,ename) VALUES (:n.empno, :n.ename); ELSE UPDATE Emp_tab SET Emp_tab.ename = :n.ename WHERE Emp_tab.empno = :n.empno; END IF; SELECT COUNT(*) INTO rowcnt FROM Dept_tab WHERE deptno = :n.deptno; IF rowcnt = 0 THEN INSERT INTO Dept_tab (deptno, dept_type) VALUES(:n.deptno, :n.dept_type); ELSE UPDATE Dept_tab SET Dept_tab.dept_type = :n.dept_type WHERE Dept_tab.deptno = :n.deptno; END IF; SELECT COUNT(*) INTO rowcnt FROM Project_tab WHERE Project_tab.projno = :n.projno; IF rowcnt = 0 THEN INSERT INTO Project_tab (projno, prj_level) VALUES(:n.projno, :n.prj_level); ELSE UPDATE Project_tab SET Project_tab.prj_level = :n.prj_level WHERE Project_tab.projno = :n.projno; END IF; END; FOR EACH ROW Option The FOR EACH ROW option determines whether the trigger is a row trigger or a statement trigger. If you specify FOR EACH ROW, then the trigger fires once for each row of the table that is affected by the triggering statement. The absence of the FOR EACH ROW option indicates that the trigger fires only once for each applicable statement, but not separately for each row affected by the statement. For example, you define the following trigger: -------------------------------------------------------------------------------- Note: You may need to set up the following data structures for certain examples to work: CREATE TABLE Emp_log ( Emp_id NUMBER, Log_date DATE, New_salary NUMBER, Action VARCHAR2(20)); -------------------------------------------------------------------------------- CREATE OR REPLACE TRIGGER Log_salary_increase AFTER UPDATE ON Emp_tab FOR EACH ROW WHEN (new.Sal > 1000) BEGIN INSERT INTO Emp_log (Emp_id, Log_date, New_salary, Action) VALUES (:new.Empno, SYSDATE, :new.SAL, 'NEW SAL'); END; Then, you enter the following SQL statement: UPDATE Emp_tab SET Sal = Sal + 1000.0 WHERE Deptno = 20; If there are five employees in department 20, then the trigger fires five times when this statement is entered, because five rows are affected. The following trigger fires only once for each UPDATE of the Emp_tab table: CREATE OR REPLACE TRIGGER Log_emp_update AFTER UPDATE ON Emp_tab BEGIN INSERT INTO Emp_log (Log_date, Action) VALUES (SYSDATE, 'Emp_tab COMMISSIONS CHANGED'); END; Trigger Size The size of a trigger cannot be more than 32K. Valid SQL Statements in Trigger Bodies The body of a trigger can contain DML SQL statements. It can also contain SELECT statements, but they must be SELECT... INTO... statements or the SELECT statement in the definition of a cursor. DDL statements are not allowed in the body of a trigger. Also, no transaction control statements are allowed in a trigger. ROLLBACK, COMMIT, and SAVEPOINT cannot be used.For system triggers, {CREATE/ALTER/DROP} TABLE statements and ALTER...COMPILE are allowed. Recompiling Triggers Use the ALTER TRIGGER statement to recompile a trigger manually. For example, the following statement recompiles the PRINT_SALARY_CHANGES trigger: ALTER TRIGGER Print_salary_changes COMPILE; Disable enable trigger: ALTER TRIGGER Reorder DISABLE; ALTER TRIGGER Reorder ENABLE; Or in 1 time for all triggers on a table: ALTER TABLE Inventory DISABLE ALL TRIGGERS; ALTER DATABASE rename GLOBAL_NAME TO NEW_NAME; ==================================== 19 BACKUP RECOVERY, TROUBLESHOOTING: ==================================== 19.1 SCN: -------- The Control files and all datafiles contain the last SCN (System Change Number) after: - checkpoint, for example via ALTER SYSTEM CHECKPOINT, - shutdown normal/immediate/transactional, - log switch occurs by the system - via alter system switch logfile, - alter tablespace begin backup etc.. at checkpoint the following occurs: ------------------------------------ - The database writer (DBWR) writes all modified database blocks in the buffer cache back to datafiles, - Log writer (LGWR) or Checkpoint process (CHKPT) updates both the controlfile and the datafiles to indicate when the last checkpoint occurred (SCN) Log switching causes a checkpoint, but a checkpoint does not cause a logswitch. LGWR writes logbuffers to online redo log: ------------------------------------------ - at commit - redolog buffers 1/3 full, > 1 MB changes - before DBWR writes modified blocks to datafiles LOG_CHECKPOINT_INTERVAL init.ora parameter: ------------------------------------------- The LOG_CHECKPOINT_INTERVAL init.ora parameter controls how often a checkpoint operation will be performed based upon the number of operating system blocks that have been written to the redo log. If this value is larger than the size of the redo log, then the checkpoint will only occur when Oracle performs a log switch FROM one group to another, which is preferred. NOTE: Starting with Oracle 8.1, LOG_CHECKPOINT_INTERVAL will be interpreted to mean that the incremental checkpoint should not lag the tail of the log by more than log_checkpoint_interval number of redo blocks. On most Unix systems the operating system block size is 512 bytes. This means that setting LOG_CHECKPOINT_INTERVAL to a value of 10,000 (the default setting), causes a checkpoint to occur after 5,120,000 (5M) bytes are written to the redo log. If the size of your redo log is 20M, you are taking 4 checkpoints for each log. LOG_CHECKPOINT_TIMEOUT init.ora parameter: ------------------------------------------ The LOG_CHECKPOINT_TIMEOUT init.ora parameter controls how often a checkpoint will be performed based on the number of seconds that have passed since the last checkpoint. NOTE: Starting with Oracle 8.1, LOG_CHECKPOINT_TIMEOUT will be interpreted to mean that the incremental checkpoint should be at the log position WHERE the tail of the log was LOG_CHECKPOINT_TIMEOUT seconds ago. Checkpoint frequency impacts the time required for the database to recover FROM an unexpected failure. Longer intervals between checkpoints mean that more time will be required during database recovery. LOG_CHECKPOINTS_TO_ALERT init.ora parameter: -------------------------------------------- The LOG_CHECKPOINTS_TO_ALERT init.ora parameter, when set to a value of TRUE, allows you to log checkpoint start and stop times in the alert log. This is very helpful in determining if checkpoints are occurring at the optimal frequency and gives a chronological view of checkpoints and other database activities occurring in the background. It is a misconception that setting LOG_CHECKPOINT_TIMEOUT to a given value will initiate a log switch at that interval, enabling a recovery window used for a stand-by database configuration. Log switches cause a checkpoint, but a checkpoint does not cause a log switch. The only way to cause a log switch is manually with ALTER SYSTEM SWITCH LOGFILE or resizing the redo logs to cause more FAST_START_MTTR_TARGET init.ora parameter: ------------------------------------------ FAST_START_MTTR_TARGET enables you to specify the number of seconds the database takes to perform crash recovery of a single instance. It is the number of seconds it takes to recover FROM crash recovery. The lower the value, the more often DBWR will write the blocks to disk. FAST_START_MTTR_TARGET can be overridden by either FAST_START_IO_TARGET or LOG_CHECKPOINT_INTERVAL. FAST_START_IO_TARGET init.ora paramater: ---------------------------------------- FAST_START_IO_TARGET (available only with the Oracle Enterprise Edition) specifies the number of I/Os that should be needed during crash or instance recovery. Smaller values for this parameter result in faster recovery times. This improvement in recovery perfoRMANce is achieved at the expense of additional writing activity during normal processing. ARCHIVE_LAG_TARGET init.ora parameter: -------------------------------------- The following initialization parameter setting sets the log switch interval to 30 minutes (a typical value). ARCHIVE_LAG_TARGET = 1800 Note: More on SCN: ================== >>>> thread from asktom You Asked Tom, Would you tell me what snapshot too old error. When does it happen? What's the possible causes? How to fix it? Thank you very much. Jane and we said... I think support note covers this topic very well: ORA-01555 "Snapshot too old" - Detailed Explanation =================================================== Overview ~~~~~~~~ This article will discuss the circumstances under which a query can return the Oracle error ORA-01555 "snapshot too old (rollback segment too small)". The article will then proceed to discuss actions that can be taken to avoid the error and finally will provide some simple PL/SQL scripts that illustrate the issues discussed. Terminology ~~~~~~~~~~~ It is assumed that the reader is familiar with standard Oracle terminology such as 'rollback segment' and 'SCN'. If not, the reader should first read the Oracle Server Concepts manual and related Oracle documentation. In addition to this, two key concepts are briefly covered below which help in the understanding of ORA-01555: 1. READ CONSISTENCY: ==================== This is documented in the Oracle Server Concepts manual and so will not be discussed further. However, for the purposes of this article this should be read and understood if not understood already. Oracle Server has the ability to have multi-version read consistency which is invaluable to you because it guarantees that you are seeing a consistent view of the data (no 'dirty reads'). 2. DELAYED BLOCK CLEANOUT: ========================== This is best illustrated with an example: Consider a transaction that updates a million row table. This obviously visits a large number of database blocks to make the change to the data. When the user commits the transaction Oracle does NOT go back and revisit these blocks to make the change permanent. It is left for the next transaction that visits any block affected by the update to 'tidy up' the block (hence the term 'delayed block cleanout'). Whenever Oracle changes a database block (index, table, cluster) it stores a pointer in the header of the data block which identifies the rollback segment used to hold the rollback information for the changes made by the transaction. (This is required if the user later elects to not commit the changes and wishes to 'undo' the changes made.) Upon commit, the database simply marks the relevant rollback segment header entry as committed. Now, when one of the changed blocks is revisited Oracle examines the header of the data block which indicates that it has been changed at some point. The database needs to confirm whether the change has been committed or whether it is currently uncommitted. To do this, Oracle determines the rollback segment used for the previous transaction (from the block's header) and then determines whether the rollback header indicates whether it has been committed or not. If it is found that the block is committed then the header of the data block is updated so that subsequent accesses to the block do not incur this processing. This behaviour is illustrated in a very simplified way below. Here we walk through the stages involved in updating a data block. STAGE 1 - No changes made Description: This is the starting point. At the top of the data block we have an area used to link active transactions to a rollback segment (the 'tx' part), and the rollback segment header has a table that stores information upon all the latest transactions that have used that rollback segment. In our example, we have two active transaction slots (01 and 02) and the next free slot is slot 03. (Since we are free to overwrite committed transactions.) Data Block 500 Rollback Segment Header 5 +----+--------------+ +----------------------+---------+ | tx | None | | transaction entry 01 |ACTIVE | +----+--------------+ | transaction entry 02 |ACTIVE | | row 1 | | transaction entry 03 |COMMITTED| | row 2 | | transaction entry 04 |COMMITTED| | ... .. | | ... ... .. | ... | | row n | | transaction entry nn |COMMITTED| +-------------------+ +--------------------------------+ STAGE 2 - Row 2 is updated Description: We have now updated row 2 of block 500. Note that the data block header is updated to point to the rollback segment 5, transaction slot 3 (5.3) and that it is marked uncommitted (Active). Data Block 500 Rollback Segment Header 5 +----+--------------+ +----------------------+---------+ | tx |5.3uncommitted|-+ | transaction entry 01 |ACTIVE | +----+--------------+ | | transaction entry 02 |ACTIVE | | row 1 | +-->| transaction entry 03 |ACTIVE | | row 2 *changed* | | transaction entry 04 |COMMITTED| | ... .. | | ... ... .. | ... | | row n | | transaction entry nn |COMMITTED| +------------------+ +--------------------------------+ STAGE 3 - The user issues a commit Description: Next the user hits commit. Note that all that this does is it updates the rollback segment header's corresponding transaction slot as committed. It does *nothing* to the data block. Data Block 500 Rollback Segment Header 5 +----+--------------+ +----------------------+---------+ | tx |5.3uncommitted|--+ | transaction entry 01 |ACTIVE | +----+--------------+ | | transaction entry 02 |ACTIVE | | row 1 | +--->| transaction entry 03 |COMMITTED| | row 2 *changed* | | transaction entry 04 |COMMITTED| | ... .. | | ... ... .. | ... | | row n | | transaction entry nn |COMMITTED| +------------------+ +--------------------------------+ STAGE 4 - Another user selects data block 500 Description: Some time later another user (or the same user) revisits data block 500. We can see that there is an uncommitted change in the data block according to the data block's header. Oracle then uses the data block header to look up the corresponding rollback segment transaction table slot, sees that it has been committed, and changes data block 500 to reflect the true state of the datablock. (i.e. it performs delayed cleanout). Data Block 500 Rollback Segment Header 5 +----+--------------+ +----------------------+---------+ | tx | None | | transaction entry 01 |ACTIVE | +----+--------------+ | transaction entry 02 |ACTIVE | | row 1 | | transaction entry 03 |COMMITTED| | row 2 | | transaction entry 04 |COMMITTED| | ... .. | | ... ... .. | ... | | row n | | transaction entry nn |COMMITTED| +------------------+ +--------------------------------+ ORA-01555 Explanation ~~~~~~~~~~~~~~~~~~~~~ There are two fundamental causes of the error ORA-01555 that are a result of Oracle trying to attain a 'read consistent' image. These are : o The rollback information itself is overwritten so that Oracle is unable to rollback the (committed) transaction entries to attain a sufficiently old enough version of the block. o The transaction slot in the rollback segment's transaction table (stored in the rollback segment's header) is overwritten, and Oracle cannot rollback the transaction header sufficiently to derive the original rollback segment transaction slot. Note: If the transaction of User A is not committed, the rollback segment entries will NOT be reused, but if User A commits, the entries become free for reuse, and if a query of User B takes a lot of time, and "meet" those overwritten entries, user B gets an error. Both of these situations are discussed below with the series of steps that cause the ORA-01555. In the steps, reference is made to 'QENV'. 'QENV' is short for 'Query Environment', which can be thought of as the environment that existed when a query is first started and to which Oracle is trying to attain a read consistent image. Associated with this environment is the SCN (System Change Number) at that time and hence, QENV 50 is the query environment with SCN 50. CASE 1 - ROLLBACK OVERWRITTEN This breaks down into two cases: another session overwriting the rollback that the current session requires or the case where the current session overwrites the rollback information that it requires. The latter is discussed in this article because this is usually the harder one to understand. Steps: 1. Session 1 starts query at time T1 and QENV 50 2. Session 1 selects block B1 during this query 3. Session 1 updates the block at SCN 51 4. Session 1 does some other work that generates rollback information. 5. Session 1 commits the changes made in steps '3' and '4'. (Now other transactions are free to overwrite this rollback information) 6. Session 1 revisits the same block B1 (perhaps for a different row). Now, Oracle can see from the block's header that it has been changed and it is later than the required QENV (which was 50). Therefore we need to get an image of the block as of this QENV. If an old enough version of the block can be found in the buffer cache then we will use this, otherwise we need to rollback the current block to generate another version of the block as at the required QENV. It is under this condition that Oracle may not be able to get the required rollback information because Session 1's changes have generated rollback information that has overwritten it and returns the ORA-1555 error. CASE 2 - ROLLBACK TRANSACTION SLOT OVERWRITTEN 1. Session 1 starts query at time T1 and QENV 50 2. Session 1 selects block B1 during this query 3. Session 1 updates the block at SCN 51 4. Session 1 commits the changes (Now other transactions are free to overwrite this rollback information) 5. A session (Session 1, another session or a number of other sessions) then use the same rollback segment for a series of committed transactions. These transactions each consume a slot in the rollback segment transaction table such that it eventually wraps around (the slots are written to in a circular fashion) and overwrites all the slots. Note that Oracle is free to reuse these slots since all transactions are committed. 6. Session 1's query then visits a block that has been changed since the initial QENV was established. Oracle therefore needs to derive an image of the block as at that point in time. Next Oracle attempts to lookup the rollback segment header's transaction slot pointed to by the top of the data block. It then realises that this has been overwritten and attempts to rollback the changes made to the rollback segment header to get the original transaction slot entry. If it cannot rollback the rollback segment transaction table sufficiently it will return ORA-1555 since Oracle can no longer derive the required version of the data block. It is also possible to encounter a variant of the transaction slot being overwritten when using block cleanout. This is briefly described below : Session 1 starts a query at QENV 50. After this another process updates the blocks that Session 1 will require. When Session 1 encounters these blocks it determines that the blocks have changed and have not yet been cleaned out (via delayed block cleanout). Session 1 must determine whether the rows in the block existed at QENV 50, were subsequently changed, In order to do this, Oracle must look at the relevant rollback segment transaction table slot to determine the committed SCN. If this SCN is after the QENV then Oracle must try to construct an older version of the block and if it is before then the block just needs clean out to be good enough for the QENV. If the transaction slot has been overwritten and the transaction table cannot be rolled back to a sufficiently old enough version then Oracle cannot derive the block image and will return ORA-1555. (Note: Normally Oracle can use an algorithm for determining a block's SCN during block cleanout even when the rollback segment slot has been overwritten. But in this case Oracle cannot guarantee that the version of the block has not changed since the start of the query). Solutions ~~~~~~~~~ This section lists some of the solutions that can be used to avoid the ORA-01555 problems discussed in this article. It addresses the cases where rollback segment information is overwritten by the same session and when the rollback segment transaction table entry is overwritten. It is worth highlighting that if a single session experiences the ORA-01555 and it is not one of the special cases listed at the end of this article, then the session must be using an Oracle extension whereby fetches across commits are tolerated. This does not follow the ANSI model and in the rare cases where ORA-01555 is returned one of the solutions below must be used. CASE 1 - ROLLBACK OVERWRITTEN 1. Increase size of rollback segment which will reduce the likelihood of overwriting rollback information that is needed. 2. Reduce the number of commits (same reason as 1). 3. Run the processing against a range of data rather than the whole table. (Same reason as 1). 4. Add additional rollback segments. This will allow the updates etc. to be spread across more rollback segments thereby reducing the chances of overwriting required rollback information. 5. If fetching across commits, the code can be changed so that this is not done. 6. Ensure that the outer select does not revisit the same block at different times during the processing. This can be achieved by : - Using a full table scan rather than an index lookup - Introducing a dummy sort so that we retrieve all the data, sort it and then sequentially visit these data blocks. CASE 2 - ROLLBACK TRANSACTION SLOT OVERWRITTEN 1. Use any of the methods outlined above except for '6'. This will allow transactions to spread their work across multiple rollback segments therefore reducing the likelihood or rollback segment transaction table slots being consumed. 2. If it is suspected that the block cleanout variant is the cause, then force block cleanout to occur prior to the transaction that returns the ORA-1555. This can be achieved by issuing the following in SQL*Plus, SQL*DBA or Server Manager : alter session set optimizer_goal = rule; select count(*) from table_name; If indexes are being accessed then the problem may be an index block and clean out can be forced by ensuring that all the index is traversed. Eg, if the index is on a numeric column with a minimum value of 25 then the following query will force cleanout of the index : select index_column from table_name where index_column > 24; Examples ~~~~~~~~ Listed below are some PL/SQL examples that can be used to illustrate the ORA-1555 cases given above. Before these PL/SQL examples will return this error the database must be configured as follows : o Use a small buffer cache (db_block_buffers). REASON: You do not want the session executing the script to be able to find old versions of the block in the buffer cache which can be used to satisfy a block visit without requiring the rollback information. o Use one rollback segment other than SYSTEM. REASON: You need to ensure that the work being done is generating rollback information that will overwrite the rollback information required. o Ensure that the rollback segment is small. REASON: See the reason for using one rollback segment. ROLLBACK OVERWRITTEN rem * 1555_a.sql - rem * Example of getting ora-1555 "Snapshot too old" by rem * session overwriting the rollback information required rem * by the same session. drop table bigemp; create table bigemp (a number, b varchar2(30), done char(1)); drop table dummy1; create table dummy1 (a varchar2(200)); rem * Populate the example tables. begin for i in 1..4000 loop insert into bigemp values (mod(i,20), to_char(i), 'N'); if mod(i,100) = 0 then insert into dummy1 values ('ssssssssssss'); commit; end if; end loop; commit; end; / rem * Ensure that table is 'cleaned out'. select count(*) from bigemp; declare -- Must use a predicate so that we revisit a changed block at a different -- time. -- If another tx is updating the table then we may not need the predicate cursor c1 is select rowid, bigemp.* from bigemp where a < 20; begin for c1rec in c1 loop update dummy1 set a = 'aaaaaaaa'; update dummy1 set a = 'bbbbbbbb'; update dummy1 set a = 'cccccccc'; update bigemp set done='Y' where c1rec.rowid = rowid; commit; end loop; end; / ROLLBACK TRANSACTION SLOT OVERWRITTEN rem * 1555_b.sql - Example of getting ora-1555 "Snapshot too old" by rem * overwriting the transaction slot in the rollback rem * segment header. This just uses one session. drop table bigemp; create table bigemp (a number, b varchar2(30), done char(1)); rem * Populate demo table. begin for i in 1..200 loop insert into bigemp values (mod(i,20), to_char(i), 'N'); if mod(i,100) = 0 then commit; end if; end loop; commit; end; / drop table mydual; create table mydual (a number); insert into mydual values (1); commit; rem * Cleanout demo table. select count(*) from bigemp; declare cursor c1 is select * from bigemp; begin -- The following update is required to illustrate the problem if block -- cleanout has been done on 'bigemp'. If the cleanout (above) is commented -- out then the update and commit statements can be commented and the -- script will fail with ORA-1555 for the block cleanout variant. update bigemp set b = 'aaaaa'; commit; for c1rec in c1 loop for i in 1..20 loop update mydual set a=a; commit; end loop; end loop; end; / Special Cases ~~~~~~~~~~~~~ There are other special cases that may result in an ORA-01555. These are given below but are rare and so not discussed in this article : o Trusted Oracle can return this if configured in OS MAC mode. Decreasing LOG_CHECKPOINT_INTERVAL on the secondary database may overcome the problem. o If a query visits a data block that has been changed by using the Oracle discrete transaction facility then it will return ORA-01555. o It is feasible that a rollback segment created with the OPTIMAL clause maycause a query to return ORA-01555 if it has shrunk during the life of the query causing rollback segment information required to generate consistent read versions of blocks to be lost. Summary ~~~~~~~ This article has discussed the reasons behind the error ORA-01555 "Snapshot too old", has provided a list of possible methods to avoid the error when it is encountered, and has provided simple PL/SQL scripts that illustrate the cases discussed. >>>>> thread about SCN Do It Yourself (DIY) Oracle replication Here's a demonstration. First I create a simple table, called TBL_SRC. This is the table on which we want to perform change-data-capture (CDC). create table tbl_src ( x number primary key, y number ); Next, I show a couple of CDC tables, and the trigger on TBL_SRC that will load the CDC tables. create table trx ( trx_id varchar2(25) primary key, SCN number, username varchar2(30) ); create table trx_detail ( trx_id varchar(25) , step_id number , step_tms date , old_x number , old_y number , new_x number , new_y number , operation char(1) ); alter table trx_detail add constraint xp_trx_detail primary key ( trx_id, step_id ); create or replace trigger b4_src before insert or update or delete on tbl_src for each row DECLARE l_trx_id VARCHAR2(25); l_step_id NUMBER; BEGIN BEGIN l_trx_id := dbms_transaction.local_transaction_id; l_step_id := dbms_transaction.step_id; INSERT INTO trx VALUES (l_trx_id, userenv('COMMITSCN'), USER); EXCEPTION WHEN dup_val_on_index THEN NULL; END; INSERT INTO trx_detail (trx_id, step_id, step_tms, old_x, old_y, new_x, new_y) VALUES (l_trx_id, l_step_id, SYSDATE, :OLD.x, :OLD.y, :NEW.x, :NEW.y); END; / Let's see the magic in action. I'll insert a record. We'll see the 'provisional' SCN in the TRX table. Then we'll commit, and see the 'true'/post-commit SCN: insert into tbl_src values ( 1, 1 ); 1 row created. select * from trx; TRX_ID SCN USERNAME ------------------------- ---------- ------------------- 3.4.33402 3732931665 CIDW commit; Commit complete. select * from trx; TRX_ID SCN USERNAME ------------------------- ---------- ------------------- 3.4.33402 3732931668 CIDW Notice how the SCN "changed" from 3732931665 to 3732931668. Oracle was doing some background transactions in between. And we can look at the details of the transaction: column step_id format 999,999,999,999,999,999,999; / TRX_ID STEP_ID STEP_TMS OLD_X OLD_Y NEW_X NEW_Y O ------------------------- ---------------------------- --------- ---------- ---------- ---------- ---------- - 3.4.33402 4,366,162,821,393,448 11-NOV-06 1 1 This approach works back to at least Oracle 7.3.4. Not perfect, because it only captures DML. A TRUNCATE is DDL, and that's not captured. For the actual implementation, I stored the before and after values as CSV strings. For 9i or later, I'd use built-in Oracle functionality. 19.2 init.ora parameters and ARCHIVE MODE: ---------------------------------------- LOG_ARCHIVE_DEST=/oracle/admin/cc1/arch LOG_ARCHIVE_DEST_1=d:\oracle\oradata\arc LOG_ARCHIVE_START=TRUE LOG_ARCHIVE_FORMAT=arc_%s.log LOG_ARCHIVE_DEST_1= LOG_ARCHIVE_DEST_2= LOG_ARCHIVE_MAX_PROCESSES=2 19.3 Enabling or disabling archive mode: ---------------------------------- ALTER DATABASE ARCHIVELOG (mounted, niet open) ALTER DATABASE NOARCHIVELOG (mounted, niet open) 19.4 Implementation backup in archive mode via OS script: -------------------------------------------------------- 19.4.1 OS backup script in unix ------------------------------ ############################################### # Example archive log backup script in UNIX: # ############################################### # Set up the environment to point to the correct database ORACLE_SID=CC1; export ORACLE_SID ORAENV_ASK=NO; export ORAENV_ASK .oraenv # Backup the tablespaces svrmgrl <= first_change#; sequence# : geeft het nummer aan van de archived redo log first_change# : eerste scn in archived redo log next_change# : laatste scn in archived redo log, en de eerste scn van de volgende log checkpoint_change# : laatste actuele SCN FUZZY : Y/N, indien YES dan bevat de file changes die later zijn dan de scn in de header A datafile that contains a block whose SCN is more recent than the SCN of its header is called a fuzzy datafile. 19.8 Archived redo logs nodig voor recovery: ------------------------------------------- In V$RECOVERY_LOG staan die archived logs vermeld die nodig zijn bij een recovery. Je kunt ook V$RECOVER_FILE gebruiken om te bepalen welke files moeten recoveren. SELECT * FROM v$recover_file; Hier vind je de FILE# en deze kun je weer gebruiken met v$datafile en v$tablespace: SELECT d.name, t.name FROM v$datafile d, v$tablespace t WHERE t.ts# = d.ts# AND d.file# in (14,15,21); # use values obtained FROM V$RECOVER_FILE query 19.9 voorbeeld recovery 1 datafile: ---------------------------------- Stel 1 datafile is corrupt. Nu behoeft slechts die ene file te worden teruggezet en daarna recovery toe te passen. SVRMGRL>alter database datafile '/u01/db1/users01.dbf' offline; $ cp /stage/users01.dbf /u01/db1 SVRMGRL>recover datafile '/u01/db1/users01.dbf'; en oracle komt met een suggestie van het toepassen van archived logfiles SVRMGRL>alter database datafile '/u01/db1/users01.dbf' online; 19.10 voorbeeld recovery database: --------------------------------- Stel meerdere datafiles zijn verloren. Zet nu backup files terug. SVRMGRL>startup mount; SVRMGRL>recover database; en oracle zal de archived redo logfiles toepassen. media recovery complete SVRMGRL>alter database open; 19.11 restore naar ANDere disks: ------------------------------- - alter database backup controlfile to trace; - restore files naar nieuwe lokatie: - edit control file met nieuwe lokatie files - save dit als .sql script en voer het uit: SVRMGRL>@new.sql controlfile: startup nomount create controlfile reuse database "brdb" noresetlogs archivelog maxlogfiles 16 maxlogmembers 2 maxdatafiles 100 maxinstances 1 maxloghistory 226 logfile group 1 ('/disk03/db1/redo/redo01a.dbf', '/disk04/db1/redo/redo01b.dbf') size 2M, group 2 ('/disk03/db1/redo/redo02a.dbf', '/disk04/db1/redo/redo02b.dbf') size 2M datafile '/disk04/oracle/db1/sys01.dbf', '/disk05/oracle/db1/rbs01.dbf', '/disk06/oracle/db1/data01.dbf', '/disk04/oracle/db1/index01.dbf', character set 'us7ascii' ; RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE; ALTER DATABASE OPEN RESETLOGS; 19.12 Copy van database naar ANDere Server: ------------------------------------------ 1. kopieer alle files precies van ene lokatie naar ANDere 2. source server: alter database backup controlfile to trace 3. Maak een juiste init.ora met references nieuwe server 4. edit de ascii versie controlfile uit stap 2 waarbij alle schijflokaties verwijzen naar de target STARTUP NOMOUNT CREATE CONTROLFILE REUSE SET DATABASE "FSYS" RESETLOGS noARCHIVELOG MAXLOGFILES 8 MAXLOGMEMBERS 4 etc.. ALTER DATABASE OPEN resetlogs; of CREATE CONTROLFILE REUSE SET DATABASE "TEST" RESETLOGS ARCHIVELOG .. #RECOVER DATABASE ALTER DATABASE OPEN RESETLOGS; ALTER DATABASE OPEN RESETLOGS; CREATE CONTROLFILE REUSE DATABASE "PROD" NORESETLOGS ARCHIVELOG .. .. RECOVER DATABASE # All logs need archiving AND a log switch is needed. ALTER SYSTEM ARCHIVE LOG ALL; # Database can now be opened normally. ALTER DATABASE OPEN; 5. SVRMGRL>@script bij probleem: delete originele controlfiles en geen reuse. Voorbeeld create controlfile: ----------------------------- If you want another database name use CREATE CONTROLFILE SET DATABASE STARTUP NOMOUNT CREATE CONTROLFILE REUSE DATABASE "O901" RESETLOGS NOARCHIVELOG MAXLOGFILES 50 MAXLOGMEMBERS 5 MAXDATAFILES 100 MAXINSTANCES 1 MAXLOGHISTORY 113 LOGFILE GROUP 1 'D:\ORACLE\ORADATA\O901\REDO01.LOG' SIZE 100M, GROUP 2 'D:\ORACLE\ORADATA\O901\REDO02.LOG' SIZE 100M, GROUP 3 'D:\ORACLE\ORADATA\O901\REDO03.LOG' SIZE 100M DATAFILE 'D:\ORACLE\ORADATA\O901\SYSTEM01.DBF', 'D:\ORACLE\ORADATA\O901\UNDOTBS01.DBF', 'D:\ORACLE\ORADATA\O901\CWMLITE01.DBF', 'D:\ORACLE\ORADATA\O901\DRSYS01.DBF', 'D:\ORACLE\ORADATA\O901\EXAMPLE01.DBF', 'D:\ORACLE\ORADATA\O901\INDX01.DBF', 'D:\ORACLE\ORADATA\O901\TOOLS01.DBF', 'D:\ORACLE\ORADATA\O901\USERS01.DBF' CHARACTER SET UTF8 ; Voorbeeld controlfile: ---------------------- STARTUP NOMOUNT CREATE CONTROLFILE REUSE DATABASE "SALES" NORESETLOGS ARCHIVELOG MAXLOGFILES 5 MAXLOGMEMBERS 2 MAXDATAFILES 255 MAXINSTANCES 2 MAXLOGHISTORY 1363 LOGFILE GROUP 1 ( '/oradata/system/log/log1.log', '/oradata/dump/log/log1.log' ) SIZE 100M, GROUP 2 ( '/oradata/system/log/log2.log', '/oradata/dump/log/log2.log' ) SIZE 100M DATAFILE '/oradata/system/system.dbf', '/oradata/rbs/rollback.dbf', '/oradata/rbs/rollbig.dbf', '/oradata/system/users.dbf', '/oradata/temp/temp.dbf', '/oradata/data_big/ahp_lkt_data_small.dbf', '/oradata/data_small/ahp_lkt_data_big.dbf', '/oradata/data_big/ahp_lkt_index_small.dbf', '/oradata/index_small/ahp_lkt_index_big.dbf', '/oradata/data_small/maniin_ah_data_small.dbf', '/oradata/index_small/maniin_ah_data_big.dbf', '/oradata/index_big/maniin_ah_index_small.dbf', '/oradata/index_big/maniin_ah_index_big.dbf', '/oradata/index_big/fe_heat_data_big.dbf', '/oradata/data_small/fe_heat_index_big.dbf', '/oradata/data_small/eksa_data_small.dbf', '/oradata/data_big/eksa_data_big.dbf', '/oradata/index_small/eksa_index_small.dbf', '/oradata/index_big/eksa_index_big.dbf', '/oradata/data_small/provisioning_data_small.dbf', '/oradata/data_small/softplan_data_small.dbf', '/oradata/index_small/provisioning_index_small.dbf', '/oradata/system/tools.dbf', '/oradata/index_small/fe_heat_index_small.dbf', '/oradata/data_small/softplan_data_big.dbf', '/oradata/index_small/softplan_index_small.dbf', '/oradata/index_small/softplan_index_big.dbf', '/oradata/data_small/fe_heat_data_small.dbf' ; # Recovery is required if any of the datafiles are restored backups, # or if the last shutdown was not normal or immediate. RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE; ALTER DATABASE OPEN RESETLOGS; 19.13 PROBLEMS DURING RECOVERY: ------------------------------- BEGIN BACKUP END BACKUP normal business | system=453 switch logfile | users=455 | | CRASH tools=459 | | | | | | | ------------------------------------------------------------------------------ t=t0 t=t1 t=t2 t=t3 ORA-01194, ORA-01195: --------------------- ------- Note 1: ------- Suppose the system comes with: ORA-01194: file 1 needs more recovery to be consistent ORA-01110: data file 1: '/u03/oradata/tstc/dbsyst01.dbf' Either you had the database in archive mode or in non archive mode: archive mode RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE; ALTER DATABASE OPEN RESETLOGS; non-archive mode: # RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE; ALTER DATABASE OPEN RESETLOGS; If you have checked that the scn's of all files are the samed number, you might try in the init.ora file: _allow_resetlogs_corruption = true ------- Note 2: ------- Problem Description ------------------- You restored your hot backup and you are trying to do a point-in-time recovery. When you tried to open your database you received the following error: ORA-01195: online backup of file needs more recovery to be consistent Cause: An incomplete recovery session was started, but an insufficient number of redo logs were applied to make the file consistent. The reported file is an online backup that must be recovered to the time the backup ended. Action: Either apply more redo logs until the file is consistent or restore the file from an older backup and repeat the recovery. For more information about online backup, see the index entry "online backups" in the . This is assuming that the hot backup completed error free. Solution Description -------------------- Continue to apply the requested logs until you are able to open the database. Explanation ----------- When you perform hot backups on a file, the file header is frozen. For example, datafile01 may have a file header frozen at SCN #456. When you backup the next datafile the SCN # may be differnet. For example the file header for datafile02 may be frozen with SCN #457. Therefore, you must apply archive logs until you reach the SCN # of the last file that was backed up. Usually, applying one or two more archive logs will solve the problem, unless there was alot of activity on the database during the backup. ------- Note 3: ------- ORA-01194: file 1 needs more recovery to be consistent I am working with a test server, I can load it again but I would like to know if this kind of problem could be solved or not. Just to let you know, that I am new in Oracle Database Administration. I ran a hot backup script, which deleted the old ARCHIVE, logs at the end. After checking the script's log, I realized that the hot backup was not successful and it deleted the Archives. I tried to startup the database and an error occurred; "ORA-01589: must use RESETLOGS or NORESETLOGS option for database open" I tried to open it with the RESETLOGS option then another error occurred; "ORA-01195: online backup of file 1 needs more recovery to be consistent" Just because, it was a test environment, I have never taken any cold backups. I still have hot backups. I don't know how to recover from those. If anyone can tell me how to do it from SQLPLUS (SVRMGRL is not loaded), I would really appreciate it. Thanks, Hi Hima, The following might help. You now have a database that is operating like it's in noarchive mode since the logs are gone. 1. Mount the database. 2. Issue the following query: SELECT V1.GROUP#, MEMBER, SEQUENCE#, FIRST_CHANGE# FROM V$LOG V1, V$LOGFILE V2 WHERE V1.GROUP# = V2.GROUP# ; This will list all your online redolog files and their respective sequence and first change numbers. 3. If the database is in NOARCHIVELOG mode, issue the query: SELECT FILE#, CHANGE# FROM V$RECOVER_FILE; If the CHANGE# is GREATER than the minimum FIRST_CHANGE# of your logs, the datafile can be recovered. 4. Recover the datafile, after taking offline, you cannot take system offline which is the file in error in your case. RECOVER DATAFILE '' 5. Confirm each of the logs that you are prompted for until you receive the message "Media recovery complete". If you are prompted for a non-existing archived log, Oracle probably needs one or more of the online logs to proceed with the recovery. Compare the sequence number referenced in the ORA-280 message with the sequence numbers of your online logs. Then enter the full path name of one of the members of the redo group whose sequence number matches the one you are being asked for. Keep entering online logs as requested until you receive the message "Media recovery complete". 6. Bring the datafile online. No need for system. 7. If the database is at mount point, open it Perform a full closed backup of the existing database ------- Note 4: ------- Recover until time using backup controlfile Hi, I am trying to perform an incomplete recovery to an arbitrary point in time in the past. Eg. I want to go back five minutes. I have a hot backup of my database. (Tablespaces into hotbackup mode, copy files, tablespaces out of hotbackup mode, archive current log, backup controlfile to a file and also to a trace). (yep im in archivelog mode as well) I shutdown the current database and blow the datafiles,online redo logs,controlfiles away. I restore my backup copy of the database - (just the datafiles) startup nomount and then run an edited controlfile trace backup (with resetlogs). I then RECOVER DATABSE UNTIL TIME 'whenever' USING BACKUP CONTROLFILE. I'm prompted for logs in the usual way but the recovery ends with an ORA-1547 - Recover succeeded but open resetlogs would give the following error. The next error is that datafile 1 (system ts) - would need more recovery. Now metalink tells me that this is usually due to backups being restored that are older than the archive redo logs - this isn't the case. I have all the archive redo logs I need to cover the time the backup was taken up to the present. The time specified in the recovery is after the backup as well. What am I missing here? Its driving me nuts. I'm off back to the docs again! Thanks in advance Tim -------------------------------------------------------------------------------- From: Anand Devaraj 15-Aug-02 15:15 Subject: Re : Recover until time using backup controlfile The error indicates that Oracle requires a few more scns to get all the datafiles in sync. It is quite possible that those scns are present in the online redo logfiles which were lost. In such cases when Oracle asks for a non-existent archive log, you should provide the complete path of the online log file for the recovery to succeed. Since you dont have an online log file you should use RECOVER DATABASE UNTIL CANCEL USING BACKUP CONTROLFILE. In this case when you exhaust all the archive log files, you issue the cancel command which will automatically rollback all the incomplete transactions and get all the datafile headers in sync with the controlfile. To do an incomplete recovery using time,you usually require the online logfiles to be present. Anand -------------------------------------------------------------------------------- From: Radhakrishnan paramukurup 15-Aug-02 16:19 Subject: Re : Recover until time using backup controlfile I am not sure whether you have missed this step or just missed in the note. You need to also to switch the log at the end of the back up (I do as a matter of practice else you need the next log which is not sure to be available in case of a failure). Otherwise some of the changes to reach a consistant state is still in the online log and you can never open untill you reach a consistent state. Hope this helps ........ -------------------------------------------------------------------------------- From: Mark Gokman 15-Aug-02 16:41 Subject: Re : Recover until time using backup controlfile To successfully perform incomplete recovery, you need a full db backup that was completed prior to the point to which you want to recover, plus you need all archive logs containing all SCNs up to the point to which you want to recover. Applying these rules to your case, I have two questions: - are you recovering to the point in time AFTER the time the successful full backup was copleted? - is there an archive log that was generated AFTER the time you specify in until time? If both answers are yes, then you should have no problems. I actually recently performed such a recovery several times. -------------------------------------------------------------------------------- From: Tim Palmer 15-Aug-02 18:02 Subject: Re : Re : Recover until time using backup controlfile Thanks Guys! I think Mark has hit the nail on the head here. I was being an idiot! Ive ran this exercise a few more times (with success) and I am convinced that what I was doing was trying to recover to a point in time that basically was before the latest scn of any one file in the hot backup set I was using - convinced myself that I wasnt - but I must have been..... perhaps I need a holiday! Thanks again Tim -------------------------------------------------------------------------------- From: Oracle, Rowena Serna 16-Aug-02 15:44 Subject: Re : Recover until time using backup controlfile Thanks to mark for his input for helping you out. ------- Note 5: ------- ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get error below ORA-01152: file 2 was not restored from a sufficiently old backup ORA-01110: data file 2: 'D:\ORACLE\ORADATA\\UNDOTBS01.DBF' File number, name and directory may vary depending on Oracle configuration Details: Undo tablespace data description In an Oracle database, Undo tablespace data is an image or snapshot of the original contents of a row (or rows) in a table. This data is stored in Undo segments (formerly Rollback segments in earlier releases of Oracle) in the Undo tablespace. When a user begins to make a change to the data in a row in an Oracle table, the original data is first written to Undo segments in the Undo tablespace. The entire process (including the creation of the Undo data) is recorded in Redo logs before the change is completed and written in the Database Buffer Cache, and then the data files via the database writer (DBWn) process. If the transaction does not complete due to some error or should there be a user decision to reverse (rollback) the change, this Undo data is critical for the ability to roll back or undo the changes that were made. Undo data also ensures a way to provide read consistency in the database. Read consistency means that if there is a data change in a row of data that is not yet committed, a new query of this same row or table will not display any of the uncommitted data to other users, but will use the information from the Undo segments in the Undo tablespace to actually construct and present a consistent view of the data that only includes committed transactions or information. During recovery, Oracle uses its Redo logs to play forward through transactions in a database so that all lost transactions (data changes and their Undo data generation) are replayed into the database. Then, once all the Redo data is applied to the data files, Oracle uses the information in the Undo segments to undo or roll back all uncommitted transactions. Once recovery is complete, all data in the database is committed data, the System Change Numbers (SCN) on all data files and the control_files match, and the database is considered consistent. As for Oracle 9i, the default method of Undo management is no longer manual, but automatic; there are no Rollback segments in individual user tablespaces, and all Undo management is processed by the Oracle server, using the Undo tablespace as the container to maintain the Undo segments for the user tablespaces in the database. The tablespace that still maintains its own Rollback segments is the System tablespace, but this behavior is by design and irrelevant to the discussion here. If this configuration is left as the default for the database, and the 5.022 or 5.025 version of the VERITAS Backup Exec (tm) Oracle Agent is used to perform Oracle backups, the Undo tablespace will not be backed up. If Automatic Undo Management is disabled and the database administrator (DBA) has modified the locations for the Undo segments (if the Undo data is no longer in the Undo tablespace), this data may be located elsewhere, and the issues addressed by this TechNote may not affect the ability to fully recover the database, although it is still recommended that the upgrade to the 5.026 Oracle Agent be performed. Scenario 1 The first scenario would be a recovery of the entire database to a previous point-in-time. This type of recovery would utilize the RECOVER DATABASE USING BACKUP CONTROLFILE statement and its customizations to restore the entire database to a point before the entry of improper or corrupt data or to roll back to a point before the accidental deletion of critical data. In this type of situation, the most common procedure for the restore is to just restore the entire online backup over the existing Oracle files with the database shutdown. (See the Related Documents section for the appropriate instructions on how to restore and recover an Oracle database to a point-in-time using an online backup.) In this scenario, where the entire database would be rolled back in time, an offline restore would include all data files, archived log files, and the backup control_file from the tape or backup media. Once the RECOVER DATABASE USING BACKUP CONTROLFILE command was executed, Oracle would begin the recovery process to roll forward through the Redo log transactions, and it would then roll back or undo uncommitted transactions. At the point when the recovery process started on the actual Undo tablespace, Oracle would see that the SCN of that tablespace was too high (in relation to the record in the control_file). This would happen simply because the Undo tablespace wasn't on the tape or backup media that was restored, so the original Undo tablespace wouldn't have been overwritten, as were the other data files, during the restore operation. The failure would occur because the Undo tablespace would still be at its SCN before the restore from backup (an SCN in the future as related to the restored backup control_file). All other tablespaces and control_files would be back at their older SCNs (not necessarily consistent yet), and the Oracle server would respond with the following error messages: ORA-01547: warning: RECOVER succeeded but OPEN RESETLOGS would get error below ORA-01152: file 2 was not restored from a sufficiently old backup ORA-01110: data file 2: 'D:\ORACLE\ORADATA\\UNDOTBS01.DBF' At this point, the database cannot be opened with the RESETLOGS option, nor in a normal mode. Any attempt to do so yields the error referenced above. SQL> alter database open resetlogs; alter database open resetlogs * Error at line 1: ORA-01152: file 2 was not restored from a sufficiently old backup ORA-01110: data file 2: 'D:\ORACLE\ORADATA\DRTEST\UNDOTBS01.DBF' The only recourse here is to recover or restore an older backup that contains an Undo tablespace, whether from an older online backup, or from a closed or offline backup or copy of the database. Without this ability to acquire an older Undo tablespace to rerun the recovery operation, it will not be possible to start the database. At this point, Oracle Technical Support must be contacted. Scenario 2 The second scenario would involve the actual corruption or loss of the Undo tablespace's data files. If the Undo tablespace data is lost or corrupted due to media failure or other internal logical error or user error, this data/tablespace must be recovered. Oracle 9i does offer the ability to create a new Undo tablespace and to alter the Oracle Instance to use this new tablespace when deemed necessary by the DBA. One of the requirements to accomplish this change, though, is that there cannot be any active transactions in the Undo segments of the tablespace when it is time to actually drop it. In the case of data file corruption, uncommitted transactions in the database that have data in Undo segments can be extremely troublesome because the existence of any uncommitted transactions will lock the Undo segments holding the data so that they cannot be dropped. This will be evidenced by an "ORA-01548" error if this is attempted. This error, in turn, prevents the drop and recreation of the Undo tablespace, and thus prevents the successful recovery of the database. To overcome this problem, the transaction tables of the Undo segments can be traced to provide details on transactions that Oracle is trying to recover via rollback and these traces will also identify the objects that Oracle is trying to apply the undo to. Oracle Doc ID: 94114.1 may be referenced to set up a trace on the database startup so that the actual transactions that are locking the Undo segments can be identified and dropped. Dropping objects that contain uncommitted transactions that are holding locks on Undo segments does entail data loss, and the amount of loss depends on how much uncommitted data was in the Undo segments at the point of failure. When utilized, this trace is actually monitoring or dumping data from the transaction tables in the headers of the Undo segments (where the records that track the data in the Undo segments are located), but if the Undo tablespace's data file is actually missing, has been offline dropped, or if these Undo segment headers have been corrupted, even the ability to dump the transaction table data is lost and the only recourse at this point may be to open the database, export, and rebuild. At this point, Oracle Technical Support must be contacted. Backup Exec Agent for Oracle 5.022 and 5.025 should be upgraded to 5.026 When using the 5.022 or 5.025 version of the Backup Exec for Windows Servers Oracle Agent (see the Related Documents section for the appropriate instructions on how to identify the version of the Oracle Agent in use), the Oracle Undo tablespace is not available for backup because the Undo tablespace falls into the type category of Undo, and only tablespaces with a content type of PERMANENT are located and made available for backup. Normal full backups with all Oracle components selected will run without error and will complete with a successful status since the Undo tablespace is not actually flagged as a selection. In most Oracle recovery situations, this absence of the Undo tablespace data for restore would not cause any problem because the original Undo tablespace is still available on the database server. Restores of User tablespaces, which do not require a rollback in time, would proceed normally since lost data or changes would be replayed back into the database, and Undo data would be available to roll back uncommitted transactions to leave the database in a consistent state and ready for user access. However, in certain recovery scenarios, (in which a rollback in time or full database recovery is attempted, or in the case of damaged or missing Undo tablespace data files) this missing Undo data can result in the inability to properly recover tablespaces back to a point-in-time, and could potentially render the database unrecoverable without an offline backup or the assistance of Oracle Technical Support. The scenarios in this TechNote describe two examples (this does not necessarily imply that these are the only scenarios) of how this absence of the Undo tablespace on tape or backup media, and thus its inability to be restored, can result in failure of the database to open and can result in actual data loss. The only solution to the problems referenced within this TechNote is to upgrade the Backup Exec for Windows Servers Oracle Agent to version 5.026, and to take new offline (closed database) and then new online (running database) backups of the entire Oracle 9i database as per the Oracle Agent documentation in the Backup Exec 9.0 for Windows Servers Administrator's Guide. Oracle 9i database backups made with the 5.022 and 5.025 Agent that shipped with Backup Exec 9.0 for Windows Servers build 4367 or build 4454 should be considered suspect in the context of the information provided in this TechNote. Note: The 5.022, 5.025, and 5.026 versions of the Oracle Agent are compatible with Backup Exec 8.6 for Windows NT and Windows 2000, which includes support for Oracle 9i, as well as Backup Exec 9.0 for Windows Servers. See the Related Documents section for instructions on how to identify the version of the Oracle Agent in use. ------- Note 6: ------- - Backup a) Consistent backups A consistent backup means that all data files and control files are consistent to a point in time. I.e. they have the same SCN. This is the only method of backup when the database is in NO Archive log mode. b) Inconsistent backups An Inconsistent backup is possible only when the database is in Archivelog mode and proper Oracle aware software is used. Most default backup software can not backup open files. Special precautions need to be used and testing needs to be done. You must apply redo logs to the data files, in order to restore the database to a consistent state. c) Database Archive mode The database can run in either Archivelog mode or noarchivelog mode. When you first create the database, you specify if it is to be in Archivelog mode. Then in the init.ora file you set the parameter log_archive_start=true so that archiving will start automatically on startup. If the database has not been created with Archivelog mode enabled, you can issue the command whilst the database is mounted, not open. SVRMGR> alter database Archivelog;. SVRMGR> log archive start SVRMGR> alter database open SVRMGR> archive log list This command will show you the log mode and if automatic archival is set. d) Backup Methods Essentially, there are two backup methods, hot and cold, also known as online and offline, respectively. A cold backup is one taken when the database is shutdown. A hot backup is on taken when the database is running. Commands for a hot backup: 1. Svrmgr>alter database Archivelog Svrmgr> log archive start Svrmgr> alter database open 2. Svrmgr> archive log list --This will show what the oldest online log sequence is. As a precaution, always keep the all archived log files starting from the oldest online log sequence. 3. Svrmgr> Alter tablespace tablespace_name BEGIN BACKUP 4. --Using an OS command, backup the datafile(s) of this tablespace. 5. Svrmgr> Alter tablespace tablespace_name END BACKUP --- repeat step 3, 4, 5 for each tablespace. 6. Svrmgr> archive log list ---do this again to obtain the current log sequence. You will want to make sure you have a copy of this redo log file. 7. So to force an archived log, issue Svrmgr> ALTER SYSTEM SWITCH LOGFILE A better way to force this would be: svrmgr> alter system archive log current; 8. Svrmgr> archive log list This is done again to check if the log file had been archived and to find the latest archived sequence number. 9. Backup all archived log files determined from steps 2 and 8. Do not backup the online redo logs. These will contain the end-of-backup marker and can cause corruption if use doing recovery. 10. Back up the control file: Svrmgr> Alter database backup controlfile to 'filename' e) Incremental backups These are backups that are taken on blocks that have been modified since the last backup. These are useful as they don't take up as much space and time. There are two kinds of incremental backups Cumulative and Non cumulative. Cumulative incremental backups include all blocks that were changed since the last backup at a lower level. This one reduces the work during restoration as only one backup contains all the changed blocks. Noncumulative only includes blocks that were changed since the previous backup at the same or lower level. Using rman, you issue the command "backup incremental level n" f) Support scenarios When the database crashes, you now have a backup. You restore the backup and then recover the database. Also, don't forget to take a backup of the control file whenever there is a schema change. RECOVERY ========= There are several kinds of recovery you can perform, depending on the type of failure and the kind of backup you have. Essentially, if you are not running in archive log mode, then you can only recover the cold backup of the database and you will lose any new data and changes made since that backup was taken. If, however, the database is in Archivelog mode you will be able to restore the database up to the time of failure. There are three basic types of recovery: 1. Online Block Recovery. This is performed automatically by Oracle.(pmon) Occurs when a process dies while changing a buffer. Oracle will reconstruct the buffer using the online redo logs and writes it to disk. 2. Thread Recovery. This is also performed automatically by Oracle. Occurs when an instance crashes while having the database open. Oracle applies all the redo changes in the thread that occurred since the last time the thread was checkpointed. 3. Media Recovery. This is required when a data file is restored from backup. The checkpoint count in the data files here are not equal to the check point count in the control file. This is also required when a file was offlined without checkpoint and when using a backup control file. Now let's explain a little about Redo vs Rollback. Redo information is recorded so that all commands that took place can be repeated during recovery. Rollback information is recorded so that you can undo changes made by the current transaction but were not committed. The Redo Logs are used to Roll Forward the changes made, both committed and non- committed changes. Then from the Rollback segments, the undo information is used to rollback the uncommitted changes. Media Failure and Recovery in Noarchivelog Mode In this case, your only option is to restore a backup of your Oracle files. The files you need are all datafiles, and control files. You only need to restore the password file or parameter files if they are lost or are corrupted. Media Failure and Recovery in Archivelog Mode In this case, there are several kinds of recovery you can perform, depending on what has been lost. The three basic kinds of recovery are: 1. Recover database - here you use the recover database command and the database must be closed and mounted. Oracle will recover all datafiles that are online. 2. Recover tablespace - use the recover tablespace command. The database can be open but the tablespace must be offline. 3. Recover datafile - use the recover datafile command. The database can be open but the specified datafile must be offline. Note: You must have all archived logs since the backup you restored from, or else you will not have a complete recovery. a) Point in Time recovery: A typical scenario is that you dropped a table at say noon, and want to recover it. You will have to restore the appropriate datafiles and do a point-in-time recovery to a time just before noon. Note: you will lose any transactions that occurred after noon. After you have recovered until noon, you must open the database with resetlogs. This is necessary to reset the log numbers, which will protect the database from having the redo logs that weren't used be applied. The four incomplete recovery scenarios all work the same: Recover database until time '1999-12-01:12:00:00'; Recover database until cancel; (you type in cancel to stop) Recover database until change n; Recover database until cancel using backup controlfile; Note: When performing an incomplete recovery, the datafiles must be online. Do a select name, status from v$datafile to find out if there are any files which are offline. If you were to perform a recovery on a database which has tablespaces offline, and they had not been taken offline in a normal state, you will lose them when you issue the open resetlogs command. This is because the data file needs recovery from a point before the resetlogs option was used. b) Recovery without control file If you have lost the current control file, or the current control file is inconsistent with files that you need to recover, you need to recover either by using a backup control file command or create a new control file. You can also recreate the control file based on the current one using the 'backup control file to trace' command which will create a script for you to run to create a new one. Recover database using backup control file command must be used when using a control file other that the current. The database must then be opened with resetlogs option. c) Recovery of missing datafile with rollback segment The tricky part here is if you are performing online recovery. Otherwise you can just use the recover datafile command. Now, if you are performing an online recovery, you must first ensure that in the init.ora file, you remove the parameter rollback_segments. Otherwise, oracle will want to use those rollback segments when opening the database, but can't find them and wont open. Until you recover the datafiles that contain the rollback segments, you need to create some temporary rollback segments in order for new transactions to work. Even if other rollback segments are ok, they will have to be taken offline. So, all the rollback segments that belong to the datafile need to be recovered. If all the datafiles belonging to the tablespace rollback_data were lost, you can now issue a recover tablespace rollback_data. Next bring the tablespace online and check the status of the rollback segments by doing a select segment_name, status from dba_rollback_segs; You will see the list of rollback segments that are in status Need Recovery. Simply issue alter rollback segment online command to complete. Don't forget to reset the rollback_segments parameter in the init.ora. d) Recovery of missing datafile without rollback segment There are three ways to recover in this scenario, as mentioned above. 1. recover database 2. recover datafile 'c:\orant\database\usr1orcl.ora' 3. recover tablespace user_data e) Recovery with missing online redo logs Missing online redo logs means that somehow you have lost your redo logs before they had a chance to archived. This means that crash recovery cannot be performed, so media recovery is required instead. All datafiles will need to berestored and rolled forwarded until the last available archived log file is applied. This is thus an incomplete recovery, and as such, the recover database command is necessary. (i.e. you cannot do a datafile or tablespace recovery). As always, when an incomplete recovery is performed, you must open the database with resetlogs. Note: the best way to avoid this kind of a loss, is to mirror your online log files. f) Recovery with missing archived redo logs If your archives are missing, the only way to recover the database is to restore from your latest backup. You will have lost any uncommitted transactions which were recorded in the archived redo logs. Again, this is why Oracle strongly suggests mirroring your online redo logs and duplicating copies of the archives. g) Recovery with resetlogs option Reset log option should be the last resort, however, as we have seen from above, it may be required due to incomplete recoveries. (recover using a backup control file, or a point in time recovery). It is imperative that you backup up the database immediately after you have opened the database with reset logs. The reason is that oracle updates the control file and resets log numbers, and you will not be able to recover from the old logs. The next concern will be if the database crashes after you have opened the database with resetlogs, but have not had time to backup the database. How to recover? Shut down the database Backup all the datafiles and the control file Startup mount Alter database open resetlogs This will work, because you have a copy of a control file after the resetlogs point. Media failure before a backup after resetlogs. If a media failure should occur before a backup was made after you opened the database using resetlogs, you will most likely lose data. The reason is because restoring a lost datafile from a backup prior to the resetlogs will give an error that the file is from a point in time earlier, and you don't have its backup log anymore. h) Recovery with corrupted/missing rollback segments. If a rollback segment is missing or corrupted, you will not be able to open the database. The first step is to find out what object is causing the rollback to appear corrupted. If we can determine that, we can drop that object. If we can't we will need to log an iTar to engage support. So, how do we find out if it's actually a bad object? 1. Make sure that all tablespaces are online and all datafiles are online. This can be checked through v$datafile, under the status column. For tablespaces associated with the datafiles, look in dba_tablespaces. If this doesn't show us anything, i.e., all are online, then 2. Put the following in the init.ora: event = "10015 trace name context forever, level 10" This event will generate a trace file that will reveal information about the transaction Oracle is trying to roll back and most importantly, what object Oracle is trying to apply the undo to. Stop and start the database. 3. Check in the directory that is specified by the user_dump_dest parameter (in the init.ora or show parameter command) for a trace file that was generated at startup time. 4. In the trace file, there should be a message similar to: error recovery tx(#,#) object #. TX(#,#) refers to transaction information. The object # is the same as the object_id in sys.dba_objects. 5. Use the following query to find out what object Oracle is trying to perform recovery on. select owner, object_name, object_type, status from dba_objects where object_id = ; 6. Drop the offending object so the undo can be released. An export or relying on a backup may be necessary to restore the object after the corrupted rollback segment goes away. 7. After dropping the object, put the rollback segment back in the init.ora parameter rollback_segments, remove the event, and shutdown and startup the database. In most cases, the above steps will resolve the problematic rollback segment. If this still does not resolve the problem, it may be likely that the corruption is in the actual rollback segment. If in fact the rollback segment itself is corrupted, we should see if we can restore from a backup. However, that isn't always possible, there may not be a recent backup etc. In this case, we have to force the database open with the unsupported, hidden parameters, you will need to log an iTar to engage support. Please note, that this is potentially dangerous! When these are used, transaction tables are not read on opening of the database Because of this, the typical safeguards associated with the rollback segment are disabled. Their status is 'offline' in dba_rollback_segs. Consequently, there is no check for active transactions before dropping the rollback segment. If you drop a rollback segment which contains active transactions then you will have logical corruption. Possibly this corruption will be in the data dictionary. If the rollback segment datafile is physically missing, has been offlined dropped, or the rollback segment header itself is corrupt, there is no way to dump the transaction table to check for active transactions. So the only thing to do is get the database open, export and rebuild. Log an iTar to engage support to help with this process. If you cannot get the database open, there is no other alternative than restoring from a backup. i) Recovery with System Clock change. You can end up with duplicate timestamps in the datafiles when a system clock changes. A solution here is to recover the database until time 'yyyy-mm-dd:00:00:00', and set the time to be later than the when the problem occurred. That way it will roll forward through the records that were actually performed later, but have an earlier time stamp due to the system clock change. Performing a complete recovery is optimal, as all transactions will be applied. j) Recovery with missing System tablespace. The only option is to restore from a backup. k) Media Recovery of offline tablespace When a tablespace is offline, you cannot recover datafiles belonging to this tablespace using recover database command. The reason is because a recover database command will only recover online datafiles. Since the tablespace is offline, it thinks the datafiles are offline as well, so even if you recover database and roll forward, the datafiles in this tablespace will not be touched. Instead, you need to perform a recover tablespace command. Alternatively, you could restored the datafiles from a cold backup, mount the database and select from the v$datafile view to see if any of the datafiles are offline. If they are, bring them online, and then you can perform a recover database command. l) Recovery of Read-Only tablespaces If you have a current control file, then recovery of read only tablespaces is no different than recovering read-write files. The issues with read-only tablespaces arise if you have to use a backup control file. If the tablespace is in read-only mode, and hasn't changed to read-write since the last backup, then you will be able to media recovery using a backup control file by taking the tablespace offline. The reason here is that when you are using the backup control file, you must open the database with resetlogs. And we know that Oracle wont let you read files from before a resetlogs was done. However, there is an exception with read-only tablespaces. You will be able to take the datafiles online after you have opened the database. When you have tablespaces that switch modes and you don't have a current control file, you should use a backup control file that recognizes the tablespace in read-write mode. If you don't have a backup control file, you can create a new one using the create controlfile command. Basically, the point here is that you should take a backup of the control file every time you switch a tablespaces mod ORA-01547: ORA-01110: ORA-01588 ORA-00205: ---------- ------- NOTE 7: ------- PURPOSE ------- To consolidate the common reasons & solutions for the ORA-1113 error. SCOPE & APPLICATION -------------------- Customers facing ORA-1113 and analysts requiring information on known issues with ORA-1113 errors. ORA-1113 ======== An ORA-1113 occurs when a datafile needs recovery. Error Explanation: ------------------ 01113, 00000, "file %s needs media recovery" Cause: An attempt was made to online or open a database with a file that is in need of media recovery. Action: First apply media recovery to the file. This error is usually followed with ORA-1110 error which will indicate the name of the datafile that needs media recovery. Eg: ORA-01113: file 28 needs media recovery ORA-01110: data file 28: '/h04/usupport/app/oracle/oradata/v817/nar.dbf' This error message indicates that a datafile that is not up-to-date with respect to the controlfile and other datafiles. Oracle's architecture is tightly coupled in the sense that all database files i.e., datafiles, redolog files, and controlfiles -- must be in sync when the database is opened or at the end of a checkpoint. This implies that the checkpoint SCN (System Commit Number) of all datafiles must be the same. If that is not the case for a particular datafile, an ORA-1113 error will be generated. For example, when you put a tablespace in hot backup mode, the checkpoint SCN of all its datafiles is frozen at the current value until you issue the corresponding end backup. If the database crashes during a hot backup and you try to restart it without doing recovery, you will likely get ORA-1113 for at least one of the datafiles in the tablespace that was being backed up, since its SCN will probably be lower than that of the controlfile and the datafiles in other tablespaces. Likewise, offlining a datafile causes its checkpoint SCN to freeze. If you simply attempt to online the file without recovering it first, its SCN will likely be much older than that of the online datafiles, and thus an ORA-1113 will result. ********************************************** Before Starting these actions do the following: ********************************************** Note : If you are using Oracle9i, use SQL*Plus, instead of Server Manager to execute the mentioned commands, since Server Manager is not available in Oracle9i. Query the V$LOG and V$LOGFILE. 1. If the database is down, you need to mount it first. SVRMGR> STARTUP MOUNT PFILE=; 2. Then connect internal Server Manager and issue the query: SVRMGR> CONNECT INTERNAL; or SQL> connect / as sysdba (for Oracle9i) SVRMGR> SELECT V1.GROUP#, MEMBER, SEQUENCE#, FIRST_CHANGE# FROM V$LOG V1, V$LOGFILE V2 WHERE V1.GROUP# = V2.GROUP# ; This will list all your online redolog files and their respective sequence and first change numbers. The steps to take next depend on the scenario in which the ORA-1113 was issued. This is discussed in the following sections. POSSIBLE CAUSES AND SOLUTIONS SUMMARY: ===================================== I. AT STARTUP AFTER CRASH WITH TABLESPACE(S) IN HOT BACKUP II. AT STARTUP AFTER RESTORING A DATAFILE OR TABLESPACE FROM A BACKUP III. TRYING TO ONLINE A DATAFILE OR TABLESPACE IV. WHEN RECOVERING ' USING BACKUP CONTROLFILE' OPTION TO DO INCOMPLETE RECOVERY I. AT STARTUP AFTER CRASH WITH TABLESPACE(S) IN HOT BACKUP ********************************************************** A. WITH ORACLE 7.1 OR LOWER 1. Mount the database.(If the database is NOT already mounted) SVRMGR> STARTUP MOUNT PFILE=; 2. Apply media recovery to the database. SVRMGR> RECOVER DATABASE; 3. Confirm each of the archived logs that you are prompted for until you receive the message "Media recovery complete". If you are prompted for an archived log that does not exist, Oracle probably needs one or more of the online logs to proceed with the recovery. Compare the sequence number referenced in the ORA-280 message with the sequence numbers of your online logs. Then enter the full path name of one of the members of the redo group whose sequence number matches the one you are being asked for. 4. Open the database. SVRMGR> ALTER DATABASE OPEN; B. WITH ORACLE 7.2 OR HIGHER 1. Mount the database. SVRMGR> STARTUP MOUNT; 2. Find out which datafiles were in hot backup mode when the database crashed or was shutdown abort or the machine was rebooted by running the query: SVRMGR> SELECT V1.FILE#, NAME FROM V$BACKUP V1, V$DATAFILE V2 WHERE V1.STATUS = 'ACTIVE' AND V1.FILE# = V2.FILE# ; 3. For each of the files returned by the above query, issue the command: SVRMGR> ALTER DATABASE DATAFILE '' END BACKUP; 4. Open the database. SVRMGR> ALTER DATABASE OPEN; II. AT STARTUP AFTER RESTORING A DATAFILE OR TABLESPACE FROM A BACKUP ********************************************************************* A. WITH THE DATABASE IN ARCHIVELOG MODE 1. Mount the database. SVRMGR> STARTUP MOUNT; 2. Recover the datafile: SVRMGR> RECOVER DATAFILE ''; If recovering more than one datafile in a tablepace issue a SVRMGR> RECOVER TABLESPACE; If recovering more than one tablespace issue a SVRMGR> RECOVER DATABASE; 3. Confirm each of the archived logs that you are prompted for until you receive the message "Media recovery complete". If you are prompted for an archived log that does not exist, Oracle probably needs one or more of the online logs to proceed with the recovery. Compare the sequence number referenced in the ORA-280 message with the sequence numbers of your online logs. Then enter the full path name of one of the members of the redo group whose sequence number matches the one you are being asked for. 4. Open the database. SVRMGR> ALTER DATABASE OPEN; B. WITH THE DATABASE IN NOARCHIVELOG MODE In this case, you will only succeed in recovering the datafile or tablespace if the redo to be applied to it is within the range of your online logs. Issue the query: SVRMGR> SELECT FILE#, CHANGE# FROM V$RECOVER_FILE; Compare the change number you obtain with the FIRST_CHANGE# of your online logs. If the CHANGE# is GREATER than the minimum FIRST_CHANGE# of your logs, the datafile can be recovered. In this case, the procedure to be followed is analogous to that of scenario II.A above, except that you must always enter the appropriate online log when prompted, until recovery is finished. If the CHANGE# is LESS than the minimum FIRST_CHANGE# of your logs, the file cannot be recovered.Your options at this point include: - If the datafile is in a temporary or index tablespace, you may drop it with an ALTER DATABASE DATAFILE '' OFFLINE DROP statement and then open the database. Once the database is up, you must drop the tablespace to which the datafile belongs and recreate it. - If the datafile is in the SYSTEM or in a rollback tablespace, restore an up-to-date copy of the datafile (if available) or your most recent full backup.In case you do not have either of this, then it might not be possible to recover the database fully. For more details or to assist you in your decision, please contact Oracle Customer Support. For all other cases in this scenario, you must weigh the cost of going to a backup versus the cost of recreating the tablespace involved, as described in the two previous cases.For more details or to assist you in your decision, please contact Oracle Customer Support. III. TRYING TO ONLINE A DATAFILE OR TABLESPACE ********************************************** 1. Recover the datafile: SVRMGRL> RECOVER DATAFILE ''; If recovering a tablespace, do SVRMGRL> RECOVER TABLESPACE ; If recovering a database, do SVRMGRL> RECOVER DATABASE; 2. Confirm each of the archived logs that you are prompted for until you receive the message "Media recovery complete". If you are prompted for an archived log that does not exist, Oracle probably needs one or more of the online logs to proceed with the recovery. Compare the sequence number referenced in the ORA-280 message with the sequence numbers of your online logs. Then enter the full path name of one of the members of the redo group whose sequence number matches the one you are being asked for. 3. Open the database. SVRMGR> ALTER DATABASE OPEN; IV. WHEN RECOVERING ' USING BACKUP CONTROLFILE' OPTION TO DO INCOMPLETE RECOVERY ******************************************************************************* If the database is recovered with the "RECOVER DATABASE USING BACKUP CONTROLFILE;" option without specifying the "UNTIL CANCEL" option, then upon "ALTER DATABASE OPEN RESETLOGS;" you will encounter the ORA-1113 error. Steps to workaround this issue: 1. Recover database again using: SVRMGR> RECOVER DATABASE USING BACKUP CONTROLFILE UNTIL CANCEL; 2. Cancel recovery by issuing the "CANCEL" command. 3. Open the database using: SVRMGR> ALTER DATABASE OPEN RESETLOGS; Notes that explain other possible recovery scenarios involving ORA-1113: ------------------------------------------------------------------------ Note 116374.1 -- "ORA-1113 on Datafile After Moving Datafile Using USFDMP Utility" ORA-1113 on Datafile After Moving Datafile Using USFDMP Note 1079626.6 -- "VMS: Mount Phase of Database Startup Results in ORA-01113, ORA-01186 & ORA-01122" ORA-1113 due to datafile getting locked on OpenVMS Note 1020262.102 -- "ORA-01113, ORA-01110: TRYING TO STARTUP DATABASE AFTER INCOMPLETE RECOVERY" Another scenario of ORA-1113 Note 168115.1 -- "ORA-01113 ORA-01110 on Database Startup after Write Disk failure" Datafile header contains different SCN comparing to other database files due to a write disk failure Note 146039.1 -- "Database Startup Fails with ORA-01113, ORA-01110" Dropping the datafile while ORA-1113, if the datafile is not required Some usefull queries: set pagesize 20000 set linesize 1000 set pause off set serveroutput on set feedback on set echo on set numformat 999999999999999 Spool recover.lst show parameter pfile; archive log list; select * from v$backup; select file#, status, substr(name, 1, 70) from v$datafile; select distinct checkpoint_change# from v$datafile_header; select status,resetlogs_change#,resetlogs_time,checkpoint_change#, to_char(checkpoint_time, 'DD-MON-YYYY HH24:MI:SS') as checkpoint_time,count(*) from v$datafile_header group by status, resetlogs_change#, resetlogs_time, checkpoint_change#, checkpoint_time order by status, checkpoint_change#, checkpoint_time ; select substr(name,1,60), recover, fuzzy, checkpoint_change#, resetlogs_change#, resetlogs_time from v$datafile_header; select name, open_mode, checkpoint_change#, ARCHIVE_CHANGE# from v$database; select GROUP#,THREAD#,SEQUENCE#,MEMBERS,ARCHIVED,STATUS,FIRST_CHANGE# from v$log; select GROUP#,substr(member,1,60) from v$logfile; select * from v$log_history; select * from v$recover_file; select * from v$recovery_log; select HXFIL File_num,substr(HXFNM,1,70) File_name,FHTYP Type,HXERR Validity, FHSCN SCN, FHTNM TABLESPACE_NAME,FHSTA status ,FHRBA_SEQ Sequence from X$KCVFH; select hxfil FileNo,FHSTA status from x$kcvfhall; spool off ------- NOTE 8: ------- Subject: How to recover and open the database if the archivelog required for recovery is either missing, lost or corrupted? Doc ID: Note:465478.1 Type: HOWTO Last Revision Date: 28-MAY-2008 Status: PUBLISHED In this Document Goal Solution References -------------------------------------------------------------------------------- Applies to: Oracle Server - Enterprise Edition - Version: 8.1.7.4 to 11.1 Information in this document applies to any platform. Goal How to recover and open the database if the archivelog required for recovery is either missing, lost or corrupted? Solution The assumption here is that we have exhausted all possible locations to find another good and valid copy or backup of the archivelog that we are looking for, which could be in one of the following: directories defined in the LOG_ARCHIVE_DEST_n another directory in the same server or another server standby database RMAN backup OS backup If the archivelog is not found in any of the above mentioned locations, then the approach and strategy on how to recover and open the database depends on the SCN (System Change Number) of the datafiles, as well as, whether the log sequence# required for the recovery is still available in the online redologs. For the SCN of the datafiles, it is important to know the mode of the database when the datafiles are backed up. That is whether the database is open, mounted or shutdown (normally) when the backup is taken. If the datafiles are restored from an online or hot backup, which means that the database is open when the backup is taken, then we must apply at least the archivelog(s) or redolog(s) whose log sequence# are generated from the beginning and until the completion of the said backup that was used to restore the datafiles. However, if the datafiles are restored from an offline or cold backup, and the database is cleanly shutdown before the backup is taken, that means that the database is either not open, is in nomount mode or mounted when the backup is taken, then the datafiles are already synchronized in terms of their SCN. In this situation, we can immediately open the database without even applying archivelogs, because the datafiles are already in a consistent state, except if there is a requirement to roll the database forward to a point-in-time after the said backup is taken. The critical key thing here is to ensure that all of the online datafiles are synchronized in terms of their SCN before we can normally open the database. So, run the following SQL statement, as shown below, to determine whether the datafiles are synchronized or not. Take note that we query the V$DATAFILE_HEADER, because we want to know the SCN recorded in the header of the physical datafile, and not the V$DATAFILE, which derives the information from the controlfile. select status, checkpoint_change#, to_char(checkpoint_time, 'DD-MON-YYYY HH24:MI:SS') as checkpoint_time, count(*) from v$datafile_header group by status, checkpoint_change#, checkpoint_time order by status, checkpoint_change#, checkpoint_time; The results of the above query must return one and only one row for the online datafiles, which means that they are already synchronized in terms of their SCN. Otherwise, if the results return more than one row for the online datafiles, then the datafiles are still not synchronized yet. In this case, we need to apply archivelog(s) or redolog(s) to synchronize all of the online datafiles. By the way, take note of the CHECKPOINT_TIME in the V$DATAFILE_HEADER, which indicates the date and time how far the datafiles have been recovered. The results of the query above may return some offline datafiles. So, ensure that all of the required datafiles are online, because we may not be able to recover later the offline datafile once we open the database in resetlogs. Even though we can recover the database beyond resetlogs for the Oracle database starting from 10g and later versions due to the introduction of the format "%R" in the LOG_ARCHIVE_FORMAT, it is recommended that you online the required datafiles now than after the database is open in resetlogs to avoid any possible problems. However, in some cases, we intentionally offline the datafile(s), because we are doing a partial database restore, or perhaps we don't need the contents of the said datafile. You may run the following query to determine the offline datafiles: select file#, name from v$datafile where file# in (select file# from v$datafile_header where status='OFFLINE'); You may issue the following SQL statement to change the status of the required datafile(s) from "OFFLINE" to "ONLINE": alter database datafile online; If we are lucky that the required log sequence# is still available in the online redologs and the corresponding redolog member is still physically existing on disk, then we may apply them instead of the archivelog. To confirm, issue the following query, as shown below, that is to determine the redolog member(s) that you can apply to recover the database: set echo on feedback on pagesize 100 numwidth 16 alter session set nls_date_format = 'DD-MON-YYYY HH24:MI:SS'; select LF.member, L.group#, L.thread#, L.sequence#, L.status, L.first_change#, L.first_time, DF.min_checkpoint_change# from v$log L, v$logfile LF, (select min(checkpoint_change#) min_checkpoint_change# from v$datafile_header where status='ONLINE') DF where LF.group# = L.group# and L.first_change# >= DF.min_checkpoint_change#; If the above query returns no rows, because the V$DATABASE.CONTROLFILE_TYPE has a value of "BACKUP", then try to apply each of the redolog membes one at a time during the recovery. You may run the following query to determine the redolog members: select * from v$logfile; If you have tried to apply all of the online redolog members instead of an archivelog during the recovery, but you always received the ORA-00310 error, as shown in the example below, then the log sequence# required for recovery is no longer available in the online redolog. ORA-00279: change 189189555 generated at 11/03/2007 09:27:46 needed for thread 1 ORA-00289: suggestion : +BACKUP ORA-00280: change 189189555 for thread 1 is in sequence #428 Specify log: {=suggested | filename | AUTO | CANCEL} +BACKUP/prmy/onlinelog/group_2.258.603422107 ORA-00310: archived log contains sequence 503; sequence 428 required ORA-00334: archived log: '+BACKUP/prmy/onlinelog/group_2.258.603422107' After trying all of the possible solutions mentioned above, but you still cannot open the database, because the archivelog required for recovery is either missing, lost or corrupted, or the corresponding log sequence# is no longer available in the online redolog, since they are already overwritten during the redolog switches, then we cannot normally open the database, since the datafiles are in an inconsistent state. So, the following are the 3 options available to allow you to open the database: Option#1: Force open the database by setting the _ALLOW_RESETLOGS_CORRUPTION=TRUE in the init.ora. But there is no 100% guarantee that this will open the database. However, once the database is opened, then we must immediately rebuild the database. Database rebuild means doing the following, namely: (1) perform a full-database export, (2) create a brand new and separate database, and finally (3) import the recent export dump. This option can be tedious and time consuming, but once we successfully open the new database, then we expect minimal or perhaps no data loss at all. Before you try this option, ensure that you have a good and valid backup of the current database. Option#2: If you have a good and valid backup of the database, then restore the database from the said backup, and recover the database by applying up to the last available archivelog. In this option, we will only recover the database up to the last archivelog that is applied, and any data after that are lost. If no archivelogs are applied at all, then we can only recover the database from the backup that is restored. However, if we restored from an online or hot backup, then we may not be able to open the database, because we still need to apply the archivelogs generated during the said backup in order to synchronize the SCN of the datafiles before we can normally open the database. Option#3: Manually extract the data using the Oracle's Data Unloader (DUL), which is performed by Oracle Field Support at the customer site on the next business day and for an extra charge. If the customer wants to pursue this approach, we need the complete name, phone# and email address of the person who has the authority to sign the work order in behalf of the customer. References OTHER ERRORS: ============= 1. Control file missing ORA-00202: controlfile: 'g:\oradata\airm\control03.ctl' ORA-27041: unable to open file OSD-04002: unable to open file O/S-Error: (OS 2) The system cannot find the file specified. Sat May 24 20:02:40 2003 ORA-205 signalled during: alter database airm mount... Solution: just copy one of the present to the missing one ORA=00214 --------- 1. one Control file is different version Solution: just copy one of the present to the different one 19.13 recovery FROM ------------------ alter system disable distributed recovery ORA-2019 ORA-2058 ORA-2068 ORA-2050: FAILED DISTRIBUTED TRANSACTIONS for step by step instructions on how to proceed. The above errors indicates that there is a failed distributed transaction that needs to be manually cleaned up. See In some cases, the instance may crash before the solutions are implemented. If this is the case, issue an 'alter system disable distributed recovery' immediately after the database starts to allow the database to run without having reco terminate the instance. 19.14 get a tablespace out of backup mode: -------------------------------------- SVRMGR> connect internal SVRMGR> startup mount SVRMGR> SELECT df.name,bk.time FROM v$datafile df,v$backup bk 2> WHERE df.file# = bk.file# and bk.status = 'ACTIVE'; Shows the datafiles currently in a hot backup state. SVRMGR> alter database datafile 2> '/u03/oradata/PROD/devlPROD_1.dbf' end backup; Do an "end backup" on those listed hot backup datafiles. SVRMGR> alter database open; 19.15 Disk full, corrupt archive log --------------------------------- Archive mandatory in log_archive_dest is unavailable and it's impossible to make a full recovery. Workaround Configure log_archive_min_succeed_dest = 2 Do not use log_archive_duplex_dest 19.16 ORA-1578 ORACLE data block corrupted (file # %s, block # %s) --------------------------------------------------------------- SELECT segment_name , segment_type , owner , tablespace_name FROM sys.dba_extents WHERE file_id = &bad_file_id AND &bad_block_id BETWEEN block_id and block_id + blocks -1 19.17 Database does not start (1) SGADEF.DBF LK.DBF -------------------------------------------------- Note:1034037.6 Subject: ORA-01102: WHEN STARTING THE DATABASE Type: PROBLEM Status: PUBLISHED Content Type: TEXT/PLAIN Creation Date: 25-JUL-1997 Last Revision Date: 10-FEB-2000 Problem Description: ==================== You are trying to startup the database and you receive the following error: ORA-01102: cannot mount database in EXCLUSIVE mode Cause: Some other instance has the database mounted exclusive or shared. Action: Shutdown other instance or mount in a compatible mode. or scumnt: failed to lock /opt/oracle/product/8.0.6/dbs/lkSALES Fri Sep 13 14:29:19 2002 ORA-09968: scumnt: unable to lock file SVR4 Error: 11: Resource temporarily unavailable Fri Sep 13 14:29:19 2002 ORA-1102 signalled during: alter database mount... Fri Sep 13 14:35:20 2002 Shutting down instance (abort) Problem Explanation: ==================== A database is started in EXCLUSIVE mode by default. Therefore, the ORA-01102 error is misleading and may have occurred due to one of the following reasons: - there is still an "sgadef.dbf" file in the "ORACLE_HOME/dbs" directory - the processes for Oracle (pmon, smon, lgwr and dbwr) still exist - shared memory segments and semaphores still exist even though the database has been shutdown - there is a "ORACLE_HOME/dbs/lk" file Search Words: ============= ORA-1102, crash, immediate, abort, fail, fails, migration Solution Description: ===================== Verify that the database was shutdown cleanly by doing the following: 1. Verify that there is not a "sgadef.dbf" file in the directory "ORACLE_HOME/dbs". % ls $ORACLE_HOME/dbs/sgadef.dbf If this file does exist, remove it. % rm $ORACLE_HOME/dbs/sgadef.dbf 2. Verify that there are no background processes owned by "oracle" % ps -ef | grep ora_ | grep $ORACLE_SID If background processes exist, remove them by using the Unix command "kill". For example: % kill -9 3. Verify that no shared memory segments and semaphores that are owned by "oracle" still exist % ipcs -b If there are shared memory segments and semaphores owned by "oracle", remove the shared memory segments % ipcrm -m and remove the semaphores % ipcrm -s NOTE: The example shown above assumes that you only have one database on this machine. If you have more than one database, you will need to shutdown all other databases before proceeding with Step 4. 4. Verify that the "$ORACLE_HOME/dbs/lk" file does not exist 5. Startup the instance Solution Explanation: ===================== The "lk" and "sgadef.dbf" files are used for locking shared memory. It seems that even though no memory is allocated, Oracle thinks memory is still locked. By removing the "sgadef" and "lk" files you remove any knowledge oracle has of shared memory that is in use. Now the database can start. . 19.18 Rollback segment missing, active transactions ------------------------------------------------ Note:1013221.6 Subject: RECOVERING FROM A LOST DATAFILE IN A ROLLBACK TABLESPACE Type: PROBLEM Status: PUBLISHED Content Type: TEXT/PLAIN Creation Date: 16-OCT-1995 Last Revision Date: 18-JUN-2002 Solution 1: --------------- Error scenario: 1. set transaction use rollback segment rb1; 2. INSERTS into's... 3. SHUTDOWN ABORT; (simulate Media errors) 4. Delete file rb1.ora (Tablespace RB1 with segment rb1 ); 5. Restore a backup of the file Recover: 1. comment out INIT.ORA ROLLBACK_SEGMENT parameter , so ORACLE does not try to find the incorrect segment rb1 2. STARTUP MOUNT 3. ALTER DATABASE DATAFILE 'rb1.ora' OFFLINE; 4. ALTER DATABASE OPEN # now we are in business 5. CREATE ROLLBACK SEGMENT rbtemp TABLESPACE SYSTEM; # We need Temporary RBS for further steps; 6. ALTER ROLLBACK SEGMENT rbtemp ONLINE; 7. RECOVER TABLESPACE RB1; 8. ALTER TABLESPACE RB1 ONLINE; 9. ALTER ROLLBACK SEGMENT rb1 ONLINE; 10. ALTER ROLLBACK SEGMENT rbtemp OFFLINE; 11. DROP ROLLBACK SEGMENT rbtemp; Result: Successfully rollback uncommitted Transactions, no suspect instance. Solution 2: --------------- INTRODUCTION ------------ Rollback segments can be monitored through the data dictionary view, dba_rollback_segs. There is a status column that describes what state the rollback segment is currently in. Normal states are either online or offline. Occasionally, the status of "needs recovery" will appear. When a rollback segment is in this state, bringing the rollback segment offline or online either through the alter rollback segment command or removing it FROM the rollback_segments parameter in the init.ora usually has no effect. UNDERSTANDING ------------- A rollback segment falls into this status of needs recovery whenever Oracle tries to roll back an uncommitted transaction in its transaction table and fails. Here are some examples of why a transaction may need to rollback: 1-A user may do a dml transaction and decides to issue rollback 2-A shutdown abort occurs and the database needs to do an instance recovery in which case, Oracle has to roll back all uncommitted transactions. When a rollback of a transaction occurs, undo must be applied to the data block the modified row/s are in. If for whatever reason, that data block is unavailable, the undo cannot be applied. The result is a 'corrupted' rollback segment with the status of needs recovery. What could be some reasons a datablock is unaccessible for undo? 1-If a tablespace or a datafile is offline or missing. 2-If the object the datablock belongs to is corrupted. 3-If the datablock that is corrupt is actually in the rollback segment itself rather than the object. HOW TO RESOLVE IT ----------------- 1-MAKE sure that all tablespaces are online and all datafiles are online. This can be checked through v$datafile, under the status column. For tablespaces associated with the datafiles, look in dba_tablespaces. If that still does not resolve the problem then 2-PUT the following in the init.ora- event = "10015 trace name context forever, level 10" Setting this event will generate a trace file that will reveal the necessary information about the transaction Oracle is trying to roll back and most importantly, what object Oracle is trying to apply the undo to. 3-SHUTDOWN the database (if normal does not work, immediate, if that does not work, abort) and bring it back up. Note: An ora-1545 may be encountered, or other errors. If the database cannot startup, contact customer support at this point. 4-CHECK in the directory that is specified by the user_dump_dest parameter (in the init.ora or show parameter command) for a trace file that was generated at startup time. 5-IN the trace file, there should be a message similar to- error recovery tx(#,#) object #. TX(#,#) refers to transaction information. The object # is the same as the object_id in sys.dba_objects. 6-USE the following query to find out what object Oracle is trying to perform recovery on. SELECT owner, object_name, object_type, status FROM dba_objects WHERE object_id = ; 7-THIS object must be dropped so the undo can be released. An export or relying on a backup may be necessary to restore the object after the corrupted rollback segment goes away. 8-AFTER dropping the object, put the rollback segment back in the init.ora parameter rollback_segments, removed the event, and shutdown and startup the database. In most cases, the above steps will resolve the problematic rollback segment. If this still does not resolve the problem, it may be likely that the corruption is in the actual rollback segment. At this point, if the problem has not been resolved, please contact customer support. Solution 3: --------------- Recovery FROM the loss of a Rollback segment datafile containing active transactions How do I recover the datafile containing rollback segments having active transactions and if the backup is done with RMAN without using catalog. I have tried the case study FROM the Oracle recovery handbook,but when i tried to open the database after offlining the Rollback segment file I got the following errors ORA-00604: error occurred at recursive SQL level 2 ORA-00376: file 2 cannot be read at this time ORA-01110:data file 2: '/orabackup/CCD1prod/oradata/rbs01CCD1prod.dbf' the status of the datafile was "Recover". Anyhow shutting down and starup mounting the database allows for the database or the datafile recovery, but this was done through SVRMGRL. Here is whats happening. simulate the loss of datafile by removing FROM the os and shut down abort the database. mount the database so RMAN can restore the file, at this point offlining the file succeeds but you cannot open the database. so the question is can we offline a rollback segment datafile containing active transactions and open the database ? How to perform recovery in such case using an RMAN backup without using the catalog. I appreciate for any insight and tips into this issue. Madhukar FROM: Oracle, Tom Villane 01-May-02 21:04 Subject: Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi, The only supported way to recover FROM the loss of a rollback segment datafile containing a rollback segment with a potentially active data dictionary transaction is to restore the datafile FROM backup and roll forward to a point in time prior to the loss of the datafile (assuming archivelog mode). Tom Villane Oracle Support Metalink Analyst FROM: Madhukar Yedulapuram 02-May-02 06:46 Subject: Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi Tom, What does Rollforward upto a time prior to the loss of the datafile got to do with the recovery, are you suggesting this so that active transaction is not lost,is it possible ? Because during the recovery the rollforward is followed by rollback and all the active transactions FROM the rollback segment's transaction table will be rolled back isnt it ? My question is if I have a active transaction in a rollback segment and the file containing that rollback segment is lost and the database crashed or did a shutdown abort can we open the database after offlining the datafile and commenting out the rollback_segments parameter in the init.ora parameter, I tried to do it and got the errors which I mentioned earlier. So in this case I have to do offline recovery only or what ? Thanks, madhukar FROM: Oracle, Tom Villane 02-May-02 16:24 Subject: Re : Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi, You won't be able to open the database if you lose a rollback segment datafile that contains an active transaction. You will have to: Restore a good backup of the file RECOVER DATAFILE '' ALTER DATABASE DATAFILE '' ONLINE; The only way you would be able to open the database is if the status of the rollback were OFFLINE, any other status requires that you recover as noted before. As recovering FROM rollback corruption needs to be done properly, you may want to log an iTAR if you have additional questions. Regards Tom Villane Oracle Support Metalink Analyst FROM: Madhukar Yedulapuram 03-May-02 07:22 Subject: Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi Tom, Thank you for the reply.you said that the only way the database can be opened is if the status of the rollback segment was offline,but what happens to an active transaction which was using this rollback segment, once the database is opened and the media recovery performed on the datafile,the database will show values which were part of an active transaction and not committed,isnt this the logical corruption? madhukar FROM: Madhukar Yedulapuram 05-May-02 08:14 Subject: Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Tom, Can I get some reponse to my questions. Thank You, Madhukar FROM: Oracle, Tom Villane 07-May-02 13:53 Subject: Re : Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi, Sorry for the confusion, I should not have said "rolling forward to a point in time..." in my previous reply. No, there won't be corruption or inconsistency. The redo logs will contain the information for both committed and uncommitted transactions. Since this includes changes made to rollback segment blocks, it follows that rollback data is also (indirectly) recorded in the redo log. To recover FROM a loss of Datafiles in the SYSTEM tablespace or datafiles with active rollback segments. You must perform closed database recovery. -Shutdown the database -Restore the file FROM backup -Recover the datafile -Open the database. References: Oracle8i Backup and Recovery Guide, chapter 6 under "Losing Datafiles in ARCHIVELOG Mode ". Regards Tom Villane Oracle Support Metalink Analyst FROM: Madhukar Yedulapuram 07-May-02 22:23 Subject: Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi Tom, After offlining the rollback segment containing active transaction you can open the database and do the recovery and after that any active transactions should be rolled back and the data should not show up, but I performed the following test and Oracle is showing logical corruption by showing data which was never committed. SVRMGR> create tablespace test_rbs datafile '/orabackup/CCD1prod/oradata/test_rbs01.dbf' size 10M 2> default storage (initial 1M next 1M minextents 1 maxextents 1024); Statement processed. SVRMGR> create rollback segment test_rbs tablespace test_rbs; Statement processed. SVRMGR> create table case5 (c1 number) tablespace tools; Statement processed. SVRMGR> set transaction use rollback segment test_rbs; ORA-01598: rollback segment 'TEST_RBS' is not online SVRMGR> alter rollback segment test_rbs online; Statement processed. SVRMGR> set transaction use rollback segment test_rbs; Statement processed. SVRMGR> insert into case5 values (5); 1 row processed. SVRMGR> alter rollback segment test_rbs offline; Statement processed. SVRMGR> shutdown abort ORACLE instance shut down. SVRMGR> startup mount ORACLE instance started. Total System Global Area 145981600 bytes Fixed Size 73888 bytes Variable Size 98705408 bytes Database Buffers 26214400 bytes Redo Buffers 20987904 bytes Database mounted. SVRMGR> alter database datafile '/orabackup/CCD1prod/oradata/test_rbs01.dbf' offline; Statement processed. SVRMGR> alter database open; Statement processed. SVRMGR> recover tablespace test_rbs; Media recovery complete. SVRMGR> alter tablespace test_rbs online; Statement processed. SVRMGR> SELECT * FROM case5; C1 ---------- 5 1 row SELECTed. SVRMGR> alter rollback segment test_rbs online; Statement processed. SVRMGR> SELECT * FROM case5; C1 ---------- 5 1 row SELECTed. SVRMGR> drop rollback segment test_rbs; drop rollback segment test_rbs * ORA-01545: rollback segment 'TEST_RBS' specified not available SVRMGR> SELECT segment_name,status FROM dba_rollback_segs; SEGMENT_NAME STATUS ------------------------------ ---------------- SYSTEM ONLINE R0 OFFLINE R01 OFFLINE R02 OFFLINE R03 OFFLINE R04 OFFLINE R05 OFFLINE R06 OFFLINE R07 OFFLINE R08 OFFLINE R09 OFFLINE R10 OFFLINE R11 OFFLINE R12 OFFLINE BIG_RB OFFLINE TEST_RBS ONLINE 16 rows SELECTed. SVRMGR> drop rollback segment test_rbs; drop rollback segment test_rbs * ORA-01545: rollback segment 'TEST_RBS' specified not available Here I have to bring the rollback segment offline to dropt it. Can this be explained or is this a bug,because this caused logical corruption. FROM: Oracle, Tom Villane 10-May-02 13:19 Subject: Re : Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi, What you are showing is expected and normal, and not corruption. At the time that you issue the "alter rollback segment test_rbs online;" Oracle does an implicit commit becuase any "ALTER" statement is considered DDL and Oracle issues an implicit COMMIT before and after any data definition language (DDL)statement. Regards Tom Villane Oracle Support Metalink Analyst -------------------------------------------------------------------------------- FROM: Madhukar Yedulapuram 14-May-02 20:12 Subject: Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi Tom, So what you are saying is the moment I say Alter rollback segment RBS# online,oracle will issue an implicit commit,but if you look at my test just after performing the tablespace recovery (had only one datafile in the RBS tablespace which was offlined before opening the database and doing the recovery), I brought the tablespace online and did a SELECT FROM the table which was having the active transaction in one of the rollback segments,so this statement has issued an implicit commit and I could see the data which was never actually committed,doesnt this contradict the Oracle's stance that only that data will be shown which shown which is committed, I think this statement is true for Intance and Crash recovery,not for media recovery as the case in point proves,but still if you say Oracle issues an implicit commit,then the stance of oracle is consistent. madhukar FROM: Oracle, Tom Villane 15-May-02 18:30 Subject: Re : Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi, A slight correction to what I posted, I should have said the implicit commit happened when the rollback segment was altered offline. Whether it's an implicit commit (before and after a DDL statement like CREATE, DROP, RENAME, ALTER) or if the user did the commit, or if the user exits the application (forces a commit). All of the above are considered commits and the data will be saved. Regards Tom Villane Oracle Support Metalink Analyst FROM: Madhukar Yedulapuram 16-May-02 23:17 Subject: Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi Tom, Thank You very much,so the moment i brought the RBS offline,the transaction was committed and the data saved in the table,is that what you are saying. So the data was committed even before performing the recovery,so recovery is essentially not applying anything in this case. madhukar FROM: Oracle, Tom Villane 17-May-02 12:18 Subject: Re : Re : Recovery FROM the loss of a Rollback segment datafile containing active transactions Hi, Yes, that is what happened. Regards Tom Villane Oracle Support Metalink Analyst 19.19 After backup you increase a datafile. ------------------------------------------ problem 2: "the backed up datafile size is smaller, and Oracle won't accept it for recovery." isn't a problem because we most certainly will accept that file. As a test you can do this (i just did) o create a small 1m tablespace with a datafile. o alter it and begin backup. o copy the datafile o alter it and end backup. o alter the datafile and "autoextend on next 1m" it. o create a table with initial 2m initial extent. This will grow the datafile. o offline the tablespace o copy the 1m original file back. o try to online it -- it'll tell you the file that needs recovery (its already accepted the smaller file at this point) o alter database recover datafile 'that file'; o alter the tablespace online again -- all is well. As for the questions: 1) There is such a command -- "alter database create datafile". Here is an example I just ran through: tkyte@TKYTE816> alter tablespace t begin backup; Tablespace altered. I copied the single datafile that is in T at this point tkyte@TKYTE816> alter tablespace t end backup; Tablespace altered. tkyte@TKYTE816> alter tablespace t add datafile 'c:\temp\t2.dbf' size 1m; Tablespace altered. So, I added a datafile AFTER the backup... tkyte@TKYTE816> alter tablespace t offline; Tablespace altered. At this point, I went out and erased the two datafiles associated with T. I moved the copy of the one datafile in place... tkyte@TKYTE816> alter tablespace t online; alter tablespace t online * ERROR at line 1: ORA-01113: file 9 needs media recovery ORA-01110: data file 9: 'C:\TEMP\T.DBF' So, it sees the copy is out of sync... tkyte@TKYTE816> recover tablespace t; ORA-00283: recovery session canceled due to errors ORA-01157: cannot identify/lock data file 10 - see DBWR trace file ORA-01110: data file 10: 'C:\TEMP\T2.DBF' and now it tells of the missing datafile -- all we need do at this point is: tkyte@TKYTE816> alter database create datafile 'c:\temp\t2.dbf'; Database altered. tkyte@TKYTE816> recover tablespace t; Media recovery complete. tkyte@TKYTE816> alter tablespace t online; Tablespace altered. and we are back in business.... 19.22 Setting Trace Events ------------------------- database level via init.ora EVENT="604 TRACE NAME ERRORSTACK FOREVER" EVENT="10210 TRACE NAME CONTEXT FOREVER, LEVEL 10" session level ALTER SESSION SET EVENTS 'IMMEDIATE TRACE NAME BLOCKDUMP LEVEL 67109037'; ALTER SESSION SET EVENTS 'IMMEDIATE TRACE NAME CONTROLF LEVEL 10'; system trace dump file ALTER SESSION SET EVENTS 'IMMEDIATE TRACE NAME SYSTEMSTATE LEVEL 10'; 19.23 DROP TEMP DATAFILE ----------------------- SVRMGRL>startup mount SVRMGRL>alter database open; ora-01157 cannot identify datafile 4 - file not found ora-01110 data file 4 '/oradata/temp/temp.dbf' SVRMGRL>alter database datafile '/oradata/temp/temp.dbf' offline drop; SVRMGRL>alter database open; SVRMGRL>drop tablespace temp including contents; SVRMGRL>create tablespace temp datafile '.... 19.24 SYSTEM DATAFILE RECOVERY ----------------------------- - a normal datafile can be taken offline and the database started up. - the system file can be taken offline but the database cannot start - restore a backup copy of the system file - recover the file 19.25 Strange processes=.. and database does not start ----------------------------------------------------- Does the PROCESSES initialization parameter of init.ora depend on some other parameter ? We were getting the error as maximum no of process (50) exceeded..... The value was initially set to 50, so when the value was....changed to 200, and the database was restarted, it gave an error of "end-of-file on communication channel" The value was reduced to 150 & 100 and the same error was encountered.... when it was set back to 50, the database started.... Can anyone clear ? check out ur semaphore settings in /etc/system. try increasing seminfo_semmns 19.26 ORA-00600 -------------- I work with ORACLE DB ver.8.0.5 and recieved an error in alert.log ksedmp: internal or fatal error ORA-00600: internal error code, arguments: [12700], [3383], [41957137], [44], [], [], [], [] oerr ora 600 00600, 00000, "internal error code, arguments: [%s], [%s], [%s], [%s], [%s], [%s], [%s], [%s]" Cause: This is the generic internal error number for Oracle program exceptions. This indicates that a process has encountered an exceptional condition. Action: Report as a bug - the first argument is the internal error number Number [12700] indicates "invalid NLS parameter value (%s)" Cause: An invalid or unknown NLS configuration parameter was specified. 19.27 segment has reached it's max_extents ----------------------------------------- oracle later than 7.3.x Version 7.3 and later: You can set the MAXEXTENTS storage parameter value to UNLIMITED for any object. Rollback Segment ================ ALTER ROLLBACK SEGMENT rollback_segment STORAGE ( MAXEXTENTS UNLIMITED); Temporary Segment ================= ALTER TABLESPACE tablespace DEFAULT STORAGE ( MAXEXTENTS UNLIMITED); Table Segment ============= ALTER TABLE MANIIN_ASIAKAS STORAGE ( MAXEXTENTS UNLIMITED); ALTER TABLE MANIIN_ASIAKAS STORAGE ( NEXT 5M ); Index Segment ============= ALTER INDEX index STORAGE ( MAXEXTENTS UNLIMITED); Table Partition Segment ======================= ALTER TABLE table MODIFY PARTITION partition STORAGE (MAXEXTENTS UNLIMITED); 19.28 max logs -------------- Problem Description ------------------- In the "alert.log", you find the following warning messages: kccrsz: denied expansion of controlfile section 9 by 65535 record(s) the number of records is already at maximum value (65535) krcpwnc: following controlfile record written over: RECID #520891 Recno 53663 Record timestamp ... kccrsz: denied expansion of controlfile section 9 by 65535 record(s) the number of records is already at maximum value (65535) krcpwnc: following controlfile record written over: RECID #520892 Recno 53664 Record timestamp The database is still running. The CONTROL_FILE_RECORD_KEEP_TIME init parameter is set to 7. If you display the records used in the LOG HISTORY section 9 of the controlfile: SQL> SELECT * FROM v$controlfile_record_section WHERE type='LOG HISTORY' ; TYPE RECORDS_TOTAL RECORDS_USED FIRST_INDEX LAST_INDEX LAST_RECID ------------- ------------- ------------ ----------- ---------- ---------- LOG HISTORY 65535 65535 33864 33863 520892 The number of RECORDS_USED has reached the maximum allowed in RECORDS_TOTAL. Solution Description -------------------- Set the CONTROL_FILE_RECORD_KEEP_TIME to 0: * Insert the parameter CONTROL_FILE_RECORD_KEEP_TIME = 0 IN "INIT.ORA" -OR- * Set it momentarily if you cannot shut the database down now: SQL> alter system set control_file_record_keep_time=0; Explanation ----------- The default value for * the CONTROL_FILE_RECORD_KEEP_TIME is 7 days. SELECT value FROM v$parameter WHERE name='control_file_record_keep_time'; VALUE ----- 7 * the MAXLOGHISTORY database parameter has already reached the maximum of 65535 and it cannot be increased anymore. SQL> alter database backup controlfile to trace; => in the trace file, MAXLOGHISTORY is 65535 The MAXLOGHISTORY increases dynamically when the CONTROL_FILE_RECORD_KEEP_TIME is set to a value different FROM 0, but does not exceed 65535. Once reached, the message appears in the alert.log warning you that a controlfile record is written over. 19.29 ORA-470 maxloghistory -------------------------- Problem Description: ==================== Instance cannot be started because of ORA-470. LGWR has also died creating a trace file with an ORA-204 error. It is possible that the maxloghistory limit of 65535 as specified in the controlfile has been reached. Diagnostic Required: ==================== The following information should be requested for diagnostics: 1. LGWR trace file produced 2. Dump of the control file - using the command: ALTER SESSION SET EVENTS 'immediate trace name controlf level 10' 3. Controlfile contents, using the command: ALTER DATABASE BACKUP CONTROLFILE TO TRACE; Diagnostic Analysis: ==================== The following observations will indicate that we have the maxloghistory limit of 65535: 1. The Lgwr trace file should show the following stack trace: - in 8.0.3 and 8.0.4, OSD skgfdisp returns ORA-27069, stack: kcrfds -> kcrrlh -> krcpwnc -> kccroc -> kccfrd -> kccrbl -> kccrbp - in 8.0.5 kccrbl causes SEGV before the call to skgfdisp with wrong block number. stack: kcrfds -> kcrrlh -> krcpwnc -> kccwnc -> kccfrd -> kccrbl 2. FROM the 'dump of the controlfile': ... ... numerous lines omittted ... LOG FILE HISTORY RECORDS: (blkno = 0x13, size = 36, max = 65535, in-use = 65535, last-recid= 188706) ... the max value of 65535 reconfirms that the limit has been reached. 3. Further confirmation can be seen FROM the controlfile trace: CREATE CONTROLFILE REUSE DATABASE "ORCL" NORESETLOGS NOARCHIVELOG MAXLOGFILES 16 MAXLOGMEMBERS 2 MAXDATAFILES 50 MAXINSTANCES 1 MAXLOGHISTORY 65535 ... Diagnostic Solution: =================== 1. Set control_file_record_keep_time = 0 in the init.ora. This parameter specifies the minimum age of a log history record in days before it can be reused. With the parameter set to 0, reusable sections never expand and records are reused immediately as required. [NOTE:1063567.6] gives a good description on the use of this parameter. 2. Mount the database and retrieve details of online redo log files for use in step 6. Because the recovery will need to roll forward through current online redo logs, a list of online log details is required to indicate which redo log is current. This can be obtained using the following command: startup mount SELECT * FROM v$logfile; 3. Open the database. This is a very important step. Although the startup will fail, it is a very important step before recreating the controlfile in step 5 and hense, enabling crash recovery to repair any incomplete log switch. Without this step it may be impossible to recover the database. alter database open 4. Shutdown the database, if it did not already crash in step 3. 5. Using the backup controlfile trace, recreate the controlfile with a smaller maxloghistory value. The MAXLOGHISTORY section of the current control file cannot be extended beyond 65536 entries. The value should reflect the amount of log history that you wish to maintain. An ORA-219 may be returned when the size of the controlfile, based on the values of the MAX- parameters, is higher then the maximum allowable size. [NOTE:1012929.6] gives a good step-by-step guide to recreating the control file. 6. Recover the database. The database will automatically be mounted due to the recreation of the controlfile in step 5 : Recover database using backup controlfile; At the recovery prompt apply the online logs in sequence by typing the unquoted full path and file name of the online redo log to apply, as noted in step 2. After applying the current redo log, you will receive the message 'Media Recovery Complete'. 7. Once media recovery is complete, open the database as follows: alter database open resetlogs; Note: keep recurring "Control file resized from" > /dbms/tdbaplay/playroca/admin/dump/udump/playroca_ora_1548438.trc > Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production > With the Partitioning, OLAP and Data Mining options > ORACLE_HOME = /dbms/tdbaplay/ora10g/home > System name: AIX > Node name: pl003 > Release: 3 > Version: 5 > Machine: 00CB560D4C00 > Instance name: playroca > Redo thread mounted by this instance: 1 > Oracle process number: 28 > Unix process pid: 1548438, image: oracle@pl003 (TNS V1-V3) > > *** 2008-02-21 12:51:57.587 > *** ACTION NAME:(0000010 FINISHED67) 2008-02-21 12:51:57.583 > *** SERVICE NAME:(SYS$USERS) 2008-02-21 12:51:57.583 > *** SESSION ID:(518.643) 2008-02-21 12:51:57.583 > Control file resized from 454 to 470 blocks > kccrsd_append: rectype = 28, lbn = 227, recs = 1128 19.30 Compatible init.ora change: -------------------------------- Database files have the COMPATIBLE version in the file header. If you set the parameter to a higher value, all the headers will be updated at next database startup. This means that if you shutdown your database, downgrade the COMPATIBLE parameter, and try to restart your database, you'll receive an error message something like: ORA-00201: control file version 7.3.2.0.0 incompatible with ORACLE version 7.0.12.0.0 ORA-00202: control file: '/usr2/oracle/dbs/V73A/ctrl1V73A.ctl' In the above case, database was running with COMPATIBLE 7.3.2.0. I commented out the parameter in init.ora, that is; kernel uses default 7.0.12.0 and returns an error before mounting since kernel cannot read the controlfile header. - You may only change the value of COMPATIBLE after a COLD Backup. - You may only change the value of COMPATIBLE if the database has been shutdown in NORMAL/IMMEDIATE mode. This parameter allows you to use a new release, while at the same time guaranteeing backward compatibility with an earlier release (in case it becomes necessary to revert to the earlier release). This parameter specifies the release with which Oracle7 Server must maintain compatibility. Some features of the current release may be restricted. For example, if you are running release 7.2.2.0 with compatibility set to 7.1.0.0 in order to guarantee compatibility, you will not be able to use 7.2 features. When using the standby database and feature, this parameter must have the same value on the primary and standby databases, and the value must be 7.3.0.0.0 or higher. This parameter allows you to immediately take advantage of the maintenance improvements of a new release in your production systems without testing the new functionality in your environment. The default value is the earliest release with which compatibility can be guaranteed. Ie: It is not possible to set COMPATIBLE to 7.3 on an Oracle8 database. ----------------- Hi Tom, Just installed DB9.0.1, I tried to modify parameter in init.ora file: compatible=9.0.0(default) to 8.1.0. After I restarted the 901 DB, I got error below when I login to sqlplus: ERROR: ORA-01033: ORACLE initialization or shutdown in progress Anything wrong with that? If I change back, everything is ok. The database could not start up. If you start the database manually, from the command line -- you would discover this. For example: idle> startup pfile=initora920.ora ORACLE instance started. Total System Global Area 143725064 bytes Fixed Size 451080 bytes Variable Size 109051904 bytes Database Buffers 33554432 bytes Redo Buffers 667648 bytes Database mounted. ORA-00402: database changes by release 9.2.0.0.0 cannot be used by release 8.1.0.0.0 ORA-00405: compatibility type "Locally Managed SYSTEM tablespace" ..... Generally, compatible cannot be set DOWN as you are already using new features many times that are not compatible with the older release. You would have had to of created the database with 8.1 file formats (compatible set to 8.1 from the very beginning) ------------------------------ 19.31 ORA-27044: unable to write the header block of file: --------------------------------------------------------- Problem Description: ==================== When you manually switch redo logs, or when the log buffer causes the redo threads to switch, you see errors similar to the following in your alert log: ... Fri Apr 24 13:42:00 1998 Thread 1 advanced to log sequence 170 Current log# 4 seq# 170 mem# 0: /.../rdlACPT04.rdl Fri Apr 24 13:42:04 1998 Errors in file /.../acpt_arch_15973.trc: ORA-202: controlfile: '/.../ctlACPT01.dbf' ORA-27044: unable to write the header block of file SVR4 Error: 48: Operation not supported Additional information: 3 Fri Apr 24 13:42:04 1998 kccexpd: controlfile resize from 356 to 368 block(s) denied by OS ... Note: The particular SVR4 error observed may differ in your case and is irrelevant here. ORA-00202: "controlfile: '%s'" Cause: This message reports the name file involved in other messages. Action: See associated error messages for a description of the problem. ORA-27044: "unable to write the header block of file" Cause: write system call failed, additional information indicates which function encountered the error Action: check errno Solution Description: ===================== To workaround this problem you can: 1. Use a database blocksize smaller than 16k. This may not be practical in all cases, and to change the db_block_size of a database you must rebuild the database. - OR - 2. Set the init.ora parameter CONTROL_FILE_RECORD_KEEP_TIME equal to zero. This can be done by adding the following line to your init.ora file: CONTROL_FILE_RECORD_KEEP_TIME = 0 The database must be shut down and restarted to have the changed init.ora file read. Explanation: ============ This is [BUG:663726] , which is fixed in release 8.0.6. The write of a 16K buffer to a control file seems to fail during an implicit resize operation on the controlfile that came as a result of adding log history records (V$LOG_HISTORY) when archiving an online redo log after a log switch. Starting with Oracle8 the control file can grow to a much larger size than it was able to in Oracle7. Bug 663726 is only reproducible when the control file needs to grow AND when the db_block_size = 16k. This has been tested on instances with a smaller database block size and the problem has not been able to be reproduced. Records in some sections in the control file are circularly reusable while records in other sections are never reused. CONTROL_FILE_RECORD_KEEP_TIME applies to reusable sections. It specifies the minimum age in days that a record must have before it can be reused. In the event a new record needs to be added to a reusable section and the oldest record has not aged enough, the record section expands. If CONTROL_FILE_RECORD_KEEP_TIME is set to 0, then reusable sections never expand and records are reused as needed. 19.32 ORA-04031 error shared_pool: --------------------------------- DIAGNOSING AND RESOLVING ORA-04031 ERROR For most applications, shared pool size is critical to Oracle perfoRMANce. The shared pool holds both the d ata dictionary cache and the fully parsed or compiled representations of PL/SQL blocks and SQL statements. When any attempt to allocate a large piece of contiguous memory in the shared pool fails Oracle first flushes all objects that are not currently in use from the pool and the resulting free memory chunks are merged. If there is still not a single chunk large enough to satisfy the request ORA-04031 is returned. The message that you will get when this error appears is the following: Error: ORA 4031 Text: unable to allocate %s bytes of shared memory (%s,%s,%s) The ORA-04031 error is usually due to fragmentation in the library cache or shared pool reserved space. Before of increasing the shared pool size consider to tune the application to use shared sql and tune SHARED_POOL_SIZE, SHARED_POOL_RESERVED_SIZE, and SHARED_POOL_RESERVED_MIN_ALLOC. First determine if the ORA-04031 was a result of fragmentation in the library cache or in the shared pool reserved space by issuing the following query: SELECT free_space, avg_free_size, used_space, avg_used_size, request_failures, last_failure_size FROM v$shared_pool_reserved; The ORA-04031 is a result of lack of contiguous space in the shared pool reserved space if: REQUEST_FAILURES is > 0 and LAST_FAILURE_SIZE is > SHARED_POOL_RESERVED_MIN_ALLOC. To resolve this consider increasing SHARED_POOL_RESERVED_MIN_ALLOC to lower the number of objects being cached into the shared pool reserved space and increase SHARED_POOL_RESERVED_SIZE and SHARED_POOL_SIZE to increase the available memory in the shared pool reserved space. The ORA-04031 is a result of lack of contiguous space in the library cache if: REQUEST_FAILURES is > 0 and LAST_FAILURE_SIZE is < SHARED_POOL_RESERVED_MIN_ALLOC or REQUEST_FAILURES is 0 and LAST_FAILURE_SIZE is < SHARED_POOL_RESERVED_MIN_ALLOC The first step would be to consider lowering SHARED_POOL_RESERVED_MIN_ALLOC to put more objects into the shared pool reserved space and increase SHARED_POOL_SIZE. This view keeps information of every SQL statement and PL/SQL block executed in the database. The following SQL can show you statements with literal values or candidates to include bind variables: SELECT substr(sql_text,1,40) "SQL", count(*) , sum(executions) "TotExecs" FROM v$sqlarea WHERE executions < 5 GROUP BY substr(sql_text,1,40) HAVING count(*) > 30 ORDER BY 2; 19.33 ORA-4030 Out of memory: ---------------------------- Possibly no memory left in Oracle, or the OS does not grant more memory. Also inspect the size of any swap file. The errors is also reported if execute permissions are not in place on some procedure. 19.34 wrong permissions on oracle: ---------------------------------- Hi, I am under very confusing situation. I'm running database (8.1.7) My oracle is installed under ownership of userid "oracle" when i login with unix id "TEST" and give oracle_sid,oracle_home,PATH variables and then do sqlplus sys after logging in when i give "select file#,error from v$datafile_header;" for some file# i get error as "CAN NOT READ HEADER" but when i login through other unix id and do the same thing. I'm not getting any error.. This seems very very confusing, Could you tell me the reason behind this?? Thank & Regards, Atul Followup: sounds like you did not run the root.sh during the install and the permissions on the oracle binaries are wrong. what does ls -l $ORACLE_HOME/bin/oracle look like. it should look like this: $ ls -l $ORACLE_HOME/bin/oracle -rwsr-s--x 1 ora920 ora920 51766646 Mar 31 13:03 /usr/oracle/ora920/bin/oracle with the "s" bits set. rwsr-s--x 1 oracle dba 494456 Dec 7 1999 lsnrctl regardless of who I log in as, when you have a setuid program as the oracle binary is, it'll be running "as the owner" tell me, what does ipcs -a show you, who is the owner of the shared memory segments associated with the SGA. If that is not Oracle -- you are "getting confused" somewhere for the s bit would ensure that Oracle was the owner. Some connection troubleshooting: -------------------------------- 19.35: ====== ORA-12545: ---------- This one is probaly due to the fact the IP or HOSTNAME in tnsnames is wrong. ORA-12514: ---------- This one is probaly due to the fact the SERVICE_NAME in tnsnames is wrong or should be fully qualified with domain name. ORA-12154: ---------- This one is probaly due to the fact the alias you have used in the logon dialogbox is wrong. fully qualified with domain name. ORA-12535: ---------- The TNS-12535 or ORA-12535 error is normally a timeout error associated with Firewalls or slow Networks. + It can also be an incorrect listener.ora parameter setting for the CONNECT_TIMEOUT_ value specified. + In essence, the ORA-12535/TNS-12535 is a timing issue between the client and server. ORA-12505: ---------- TNS:listener does not currently know of SID given in connect descriptor Note 1: ------- Symptom: When trying to connect to Oracle the following error is generated: ORA-12224: TNS: listener could not resolve SID given in connection description. Cause: The SID specified in the connection was not found in the listener’s tables. This error will be returned if the database instance has not registered with the listener. Possible Remedy: Check to make sure that the SID is correct. The SIDs that are currently registered with the listener can be obtained by typing: LSNRCTL SERVICES These SIDs correspond to SID_NAMEs in TNSNAMES.ORA or DB_NAME in the initialisation file. Note 2: ------- ORA-12505: TNS:listener could not resolve SID given in connect descriptor You are trying to connect to a database, but the SID is not known. Although it is possible that a tnsping command succeeds, there might still a problem with the SID parameter of the connection string. eg. C:>tnsping ora920 TNS Ping Utility for 32-bit Windows: Version 9.2.0.7.0 - Production Copyright (c) 1997 Oracle Corporation. All rights reserved. Used parameter files: c:\oracle\ora920\network\admin\sqlnet.ora Used TNSNAMES adapter to resolve the alias Attempting to contact (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = DEV01)(PORT = 2491))) (CONNECT_DATA = (SID = UNKNOWN) (SERVER = DEDICATED))) OK (20 msec) As one can see, this is the connection information stored in a tnsnames.ora file: ORA920.EU.DBMOTIVE.COM = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = DEV01)(PORT = 2491)) ) (CONNECT_DATA = (SID = UNKNOWN) (SERVER = DEDICATED) ) ) However, the SID UNKNOWN is not known by the listener at the database server side. In order to test the known services by a listener, we can issue following command at the database server side: C:>lsnrctl services LSNRCTL for 32-bit Windows: Version 10.1.0.2.0 - Production Copyright (c) 1991, 2004, Oracle. All rights reserved. Connecting to (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=DEV01)(PORT=1521))) Services Summary... Service "ORA10G.eu.dbmotive.com" has 1 instance(s). Instance "ORA10G", status UNKNOWN, has 1 handler(s) for this service... Handler(s): "DEDICATED" established:0 refused:0 LOCAL SERVER Service "ORA920.eu.dbmotive.com" has 2 instance(s). Instance "ORA920", status UNKNOWN, has 1 handler(s) for this service... Handler(s): "DEDICATED" established:0 refused:0 LOCAL SERVER Instance "ORA920", status READY, has 1 handler(s) for this service... Handler(s): "DEDICATED" established:2 refused:0 state:ready LOCAL SERVER The command completed successfully Know services are ORA10G and ORA920. Changing the SID in our tnsnames.ora to a known service by the listener (ORA920) solved the problem. 19.36 ORA-12560 --------------- Note 1: ------- Oracle classify this as a ‘generic protocol adapter error’. In my experience it indicates that Oracle client does not know what instance to connect to or what TNS alias to use. Set the correct ORACLE_HOME ans ORACLE_SID variables. Note 2: ------- Doc ID: Note:73399.1 Subject: WINNT: ORA-12560 DB Start via SVRMGRL or SQL*PLUS ORACLE_SID is set correctly Type: BULLETIN Status: PUBLISHED Content Type: TEXT/PLAIN Creation Date: 28-JUL-1999 Last Revision Date: 14-JAN-2004 PURPOSE To assist in resolving ORA-12560 errors on Oracle8i. SCOPE & APPLICATION Support Analysts and customers. RELATED DOCUMENTS PR:1070749.6 NOTE:1016454.102 TNS 12560 DB CREATE VIA INSTALLATION OR CONFIGURATION ASSISTANT FAILS BUG:948671 ORADIM SUCCSSFULLY CREATES AN UNUSABLE SID WITH NON-ALPHANUMERIC CHARACTER BUG:892253 ORA-12560 CREATING DATABASE WITH DB CONFIGURATION ASSISTANT IF SID HAS NON-ALPHA If you encounter an ORA-12560 error when you try to start Server Manager or SQL*Plus locally on your Windows NT server, you should first check the ORACLE_SID value. Make sure the SID is correctly set, either in the Windows NT registry or in your environment (with a set command). Also, you must verify that the service is running. See the entries above for more details. If you have verified that ORACLE_SID is properly set, and the service is running, yet you still get an ORA-12560, then it is possible that you have created an instance with a non-alphanumeric character. The Getting Started Guide for Oracle8i on Windows NT documents that SID names can contain only alphanumerics, however if you attempt to create a SID with an underscore or a dash on Oracle8i you are not prevented from doing so. The service will be created and started successfully, but attempts to connect will fail with an ORA-12560. You must delete the instance and recreate it with no special characters - only alphanumerics are allowed in the SID name. See BUG#948671, which was logged against 8.1.5 on Windows NT for this issue. Note 3: ------- Doc ID : Note:119008.1 Content Type: TEXT/PLAIN Subject: ORA-12560 Connecting to the Server on Unix - Troubleshooting Creation Date: 04-SEP-2000 Type: PROBLEM Last Revision Date: 20-MAR-2003 Status: PUBLISHED PURPOSE ------- This note describes some of the possible reasons for ORA-12560 errors connecting to server on Unix Box. The list below shows some of the causes, the symptoms and the action to take. It is possible you will hit a cause not described here, in that case the information above should allow it to be identified. SCOPE & APPLICATION ------------------- Support Analysts and customers alike. ORA-12560 CONNECTING TO THE SERVER ON UNIX - TROUBLESHOOTING ------------------------------------------------------------ ORA-12560: TNS:protocol adapter error Cause: A generic protocol adapter error occurred. Action: Check addresses used for proper protocol specification. Before reporting this error, look at the error stack and check for lower level transport errors. For further details, turn on tracing and re execute the operation. Turn off tracing when the operation is complete. This is a high-level error just reporting an error occurred in the actual transport layer. Look at the next error down the stack and process that. 1. ORA-12500 ORA-12560 MAKING MULTIPLE CONNECTIONS TO DATABASE Problem: Trying to connect to the database via listener and the ORA-12500 are prompted. You may see in the listener.log ORA-12500 and ORA-12560: ORA-12500: TNS:listener failed to start a dedicated server process Cause: The process of starting up a dedicated server process failed. The executable could not be found or the environment maybe set up incorrectly. Action: Turn on tracing at the ADMIN level and re execute the operation. Verify that the ORACLE Server executable is present and has execute permissions enabled. Ensure that the ORACLE environment is specified correctly in LISTENER.ORA. If error persists, contact Worldwide Customer Support. In many cases the error ORA-12500 is caused due to leak of resources in the Unix Box, if you are enable to connect to database and randomly you get the error your operating system is reached the maximum values for some resources. Otherwise, if you get the error in first connection the problem may be in the configuration of the system. Solution: Finding the resource which is been reached is difficult, the note 2064862.102 indicates some suggestion to solve the problems. 2. ORA-12538/ORA-12560 connecting to the database via SQL*Net Problem: Trying to connect to database via SQL*Net the error the error ORA-12538 is prompted. In the trace file you can see: nscall: error exit nioqper: error from nscall nioqper: nr err code: 0 nioqper: ns main err code: 12538 nioqper: ns (2) err code: 12560 nioqper: nt main err code: 508 nioqper: nt (2) err code: 0 nioqper: nt OS err code: 0 Solution: - Check the protocol used in the TNSNAMES.ORA by the connection string - Ensure that the TNSNAMES.ORA you check is the one that is actually being used by Oracle. Define the TNS_ADMIN environment variable to point to the TNSNAMES directory. - Using the $ORACLE_HOME/bin/adapters command, ensure the protocol is installed. Run the command without parameters to check if the protocol is installed, then run the command with parameters to see whether a particular tool/application contains the protocol symbols e.g.: 1. $ORACLE_HOME/bin/adapters 2. $ORACLE_HOME/bin/adapters $ORACLE_HOME/bin/oracle $ORACLE_HOME/bin/adapters $ORACLE_HOME/bin/sqlplus Explanation: If the protocol is not installed every connection attempting to use it will fail with ORA-12538 because the executable doesn't contain the required protocol symbol/s. Error ORA-12538 may also be caused by an issue with the '$ORACLE_HOME/bin/relink all' command. 'Relink All' does not relink the sqlplus executable. If you receive error ORA-12538 when making a sqlplus connection, it may be for this reason. To relink sqlplus manually: $ su - oracle $ cd $ORACLE_HOME/sqlplus/lib $ make -f ins_sqlplus.mk install $ ls -l $ORACLE_HOME/bin/sqlplus --> should show a current date/time stamp 3. ORA-12546 ORA-12560 connecting locally to the database Problem: Trying to connect to database locally with a different account to the software owner, the error the error ORA-12546 is prompted. In the trace file you can see: nioqper: error from nscall nioqper: nr err code: 0 nioqper: ns main err code: 12546 nioqper: ns (2) err code: 12560 nioqper: nt main err code: 516 nioqper: nt (2) err code: 13 nioqper: nt OS err code: 0 Solution: Make sure the permissions of oracle executable are correct, this should be: 52224 -rwsr-sr-x 1 oracle dba 53431665 Aug 10 11:07 oracle Explanation: The problem occurs due to an incorrect setting on the oracle executable. 4. ORA-12541 ORA-12560 TRYING TO CONNECT TO A DATABASE Problem: You are trying to connect to a database using SQL*Net and receive the following error ORA-12541 ORA-12560 after change the TCP/IP port in the listener.ora and you are using PARAMETER USE_CKPFILE_LISTENER in listener.ora. The following error struct appears in the SQLNET.LOG: nr err code: 12203 TNS-12203: TNS:unable to connect to destination ns main err code: 12541 TNS-12541: TNS:no listener ns secondary err code: 12560 nt main err code: 511 TNS-00511: No listener nt secondary err code: 239 nt OS err code: 0 Solution: Check [NOTE:1061927.6] to resolve the problem. Explanation: If TCP protocol is listed in the Listener.ora's ADDRESS_LIST section and the parameter USE_CKPFILE_LISTENER = TRUE, the Listener ignores the TCP port number defined in the ADDRESS section and listens on a random port. RELATED DOCUMENTS ----------------- Note:39774.1 LOG & TRACE Facilities on NET . Note:45878.1 SQL*Net Common Errors & Diagnostic Worksheet Net8i Admin/Ch.11 Troubleshooting Net8 / Resolving the Most Common Error Messages 19.37 ORA-12637 --------------- Packet received failed. A process was unable to receive a packet from another process. Possible causes are: 1. The other process was terminated. 2. The machine on which the other process is running went down. 3. Some other communications error occurred. Note 1: Just edit the file sqlnet.ora and search for the string SQLNET.AUTHENTICATION_SERVICES. When it exists it’s set to = (TNS), change this to = (NONE). When it doesn’t exist, add the string SQLNET.AUTHENTICATION_SERVICES = (NONE) Note 2: What does SQLNET.AUTHENTICATION_SERVICES do? SQLNET.AUTHENTICATION_SERVICES Purpose Use the parameter SQLNET.AUTHENTICATION_SERVICES to enable one or more authentication services. If authentication has been installed, it is recommended that this parameter be set to either none or to one of the authentication methods. Default None Values Authentication Methods Available with Oracle Net Services: none for no authentication methods. A valid username and password can be used to access the database. all for all authentication methods nts for Windows NT native authentication Authentication Methods Available with Oracle Advanced Security: kerberos5 for Kerberos authentication cybersafe for Cybersafe authentication radius for RADIUS authentication dcegssapi for DCE GSSAPI authentication See Also: Oracle Advanced Security Administrator's Guide Example SQLNET.AUTHENTICATION_SERVICES=(kerberos5, cybersafe) Note 3: ORA-12637 for members of one NT group, using OPS$ login Being "identified externally", users can work fine until the user is added to a "wwwauthor" NT group to allow them to publish documents on Microsoft IIS (intranet) -- then they get ORA-12637 starting the Oracle c/s application (document management system). The environment is: Oracle 9.2.0.1.0 on Windows 2000 Advanced Server w. SP4, Windows 2003 domain controllers in W2K compatible mode, client workstations with W2K and Win XP. Any hint will be appreciated. Problem solved. Specific NT group (wwwauthor) which caused problems had existed already with specific permissions, then it was dropped and created again with exactly the same name (but, of course, with different internal ID). This situation have been identified as causing some kind of mess. A completely new group with different name has been created. Note 4: ORA-12637 packet receive failure I added a second instance to the Oracle server. Since then, on the server and all clients, I get ORA-12637 packet receive failure when I try to connect to this database. Why is this? Hello Try commenting out the SQLNET.CRYPTO_SEED and SQLNET.AUTHENTICATION_SERVICES in the server's SQLNET.ORA and on the client sqlnet file if they exist. Please also verify that the server's LISTENER.ORA file contains the following parameter: CONNECT_TIMEOUT_LISTENER=0 Note 5: Workaround is to turn off prespawned server processes in "listener.ora". In the "listener.ora", comment out or delete the prespawn parameters, ie: SID_LIST_LISTENER = (SID_LIST = (SID_DESC = (SID_NAME = prd) (ORACLE_HOME = /raid/app/oracle/product/7.3.4) # (PRESPAWN_MAX = 99) # (PRESPAWN_LIST = # (PRESPAWN_DESC = (PROTOCOL = TCP) (POOL_SIZE = 1) (TIMEOUT = 30)) # ) ) ) Note 6: Problem Description ------------------- Connections to Oracle 9.2 using a Cybersafe authenticated user fails on Solaris 2.6 with ORA-12637 and a core dump is generated. Solution Description -------------------- 1) Shutdown Oracle, the listener and any clients. 2) In $ORACLE_HOME/lib take a backup copy of the file sysliblist 3) Edit sysliblist. Move the -lthread entry to the beginning. So change from, -lnsl -lsocket -lgen -ldl -lsched -lthread To, -lthread -lnsl -lsocket -lgen -ldl -lsched 4) Do $ORACLE_HOME/bin/relink all Note 7: fact: Oracle Server - Personal Edition 8.1 fact: MS Windows symptom: Starting Server Manager (Svrmgrl) Fails symptom: ORA-12637: Packet Receive Failed cause: Oracle's installer will set the authentication to (NTS) by default. However, if the Windows machine is not in a Domain where there is a Windows Domain Controller, it will not be able to contact the KDC (Key Distribtion Centre) needed for Authentication. fix: Comment out SQLNET.AUTHENTICATION_SERVICES=(NTS) in sqlnet.ora 19.38 ORA 02058: ================ dba_2pc_pending: Lists all in-doubt distributed transactions. The view is empty until populated by an in-doubt transaction. After the transaction is resolved, the view is purged. SQL> SELECT LOCAL_TRAN_ID, GLOBAL_TRAN_ID, STATE, MIXED, HOST, COMMIT# 2 FROM DBA_2PC_PENDING 3 / LOCAL_TRAN_ID GLOBAL_TRAN_ID ---------------------- ---------------------------------------------------------- 6.31.5950 1145324612.10D447310B5FCE408A296417959EBEEC00000000 SQL> select STATE, MIXED, HOST, COMMIT# 2 FROM DBA_2PC_PENDING 3 / STATE MIX HOST ---------------- --- ------------------------------------------------------------ forced rollback no REBV\PGSS-TST-TCM SQL> select * from dba_2pc_neighbors; LOCAL_TRAN_ID IN_ DATABASE ---------------------- --- -------------------------------------------------- 6.31.5950 in O SQL> select state, tran_comment, advice from dba_2pc_pending; STATE TRAN_COMMENT ---------------- ------------------------------------------------------------ prepared SQL> rollback force '6.31.5950'; Rollback complete. SQL> commit; Doc ID: Note:290405.1 Subject: ORA-30019 When Executing Dbms_transaction.Purge_lost_db_entry Type: PROBLEM Status: MODERATED Content Type: TEXT/X-HTML Creation Date: 11-NOV-2004 Last Revision Date: 16-NOV-2004 The information in this document applies to: Oracle Server - Enterprise Edition - Version: 9.2.0.5 This problem can occur on any platform. Errors ORA-30019 Illegal rollback Segment operation in Automatic Undo mode Symptoms Attempting to clean up the pending transaction using DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY, getting ora-30019: ORA-30019: Illegal rollback Segment operation in Automatic Undo mode Changes AUTO UNDO MANAGEMENT is running Cause DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY is not supported in AUTO UNDO MANAGEMENT This is due to fact that "set transaction use rollback segment.." cannot be done in AUM. Fix 1.) alter session set "_smu_debug_mode" = 4; 2.) execute DBMS_TRANSACTION.PURGE_LOST_DB_ENTRY('local_tran_id'); 19.39. ORA-600 [12850]: ======================= Doc ID : Note:1064436.6 Content Type: TEXT/PLAIN Subject: ORA-00600 [12850], AND ORA-00600 [15265]: WHEN SELECT OR DESCRIBE ON TABLE Creation Date: 14-JAN-1999 Type: PROBLEM Last Revision Date: 29-FEB-2000 Status: PUBLISHED Problem Description: --------------------- You are doing a describe or select on a table and receive: ORA-600 [12850]: Meaning: 12850 occurs when it can't find the user who owns the object from the dictionary. If you try to delete the table, you receive: ORA-600 [15625]: Meaning: The arguement 15625 is occuring because some index entry for the table is not found in obj$. Problem Explanation: -------------------- The data dictionary is corrupt. You cannot drop the tables in question because the data dictionary doesn't know they exist. Search Words: ------------- ORA-600 [12850] ORA-600 [15625] describe delete table Solution Description: --------------------- You need to rebuild the database. Solution Explanation: --------------------- Since the table(s) cannot be accessed or dropped because of the data dictionary corruption, rebuilding the database is the only option. 19.40 ORA-01092: ================ ------------------------------------------------------------------------------------------- Doc ID : Note:222132.1 Content Type: TEXT/PLAIN Subject: ORA-01599 and ORA-01092 while starting database Creation Date: 03-DEC-2002 Type: PROBLEM Last Revision Date: 07-AUG-2003 Status: PUBLISHED PURPOSE ------- The purpose of this Note is to fix errors ORA-01599 & ORA-01092 when recieved at startup. SCOPE & APPLICATION ------------------- All DBAs, Support Analyst. Symptom(s) ~~~~~~~~~~ Starting the database gives errors similar to: ORA-01599: failed to acquire rollback segment (20), cache space is full (currently has (19) entries) ORA-01092: ORACLE instance terminated Change(s) ~~~~~~~~~~ Increased shared_pool_size parameter. Increased processes and/or sessions parameters. Cause ~~~~~~~ Low value for max_rollback_segments The above changes changed the value for max_rollback_segments internally. Fix ~~~~ The value for max_rollback_segments which is to be calculated as follows: max_rollback_segments = transactions/transactions_per_rollback_segment or 30 whichever is greater. transactions = session * 1.1; sessions = (processes * 1.1) + 5; The default value for transactions_per_rollback_segment = 5; 1. Use these calculations and find out the value for max_rollback_segments. 2. Set it to this value or 30 whichever is greater. 3. Startup database after this correct setting. Reference info ~~~~~~~~~~~~~~ [BUG:2233336] - RDBMS ERRORS AT STARTUP CAN CAUSE ODMA TO OMIT CLEANUP ACTIONS [NOTE:30764.1] - Init.ora Parameter "MAX_ROLLBACK_SEGMENTS" Reference Note -------------------------------------------------------------------------------------------- Doc ID : Note:1038418.6 Content Type: TEXT/PLAIN Subject: ORA-01092 STARTING UP ORACLE RDBMS DATABASE Creation Date: 17-NOV-1997 Type: PROBLEM Last Revision Date: 06-JUL-1999 Status: PUBLISHED Problem Summary: ================ ORA-01092 starting up Oracle RDBMS database. Problem Description: ==================== When you startup your Oracle RDBMS database, you receive the following error: ORA-01092: ORACLE instance terminated. Disconnection forced. Problem Explanation: ==================== Oracle cannot write to the alert_.log file because the ownership and/or permissions on the BACKGROUND_DUMP_DEST directory are incorrect. Solution Summary: ================= Modify the ownership and permissions of directory BACKGROUND_DUMP_DEST. Solution Description: ===================== To allow oracle to write to the BACKGROUND_DUMP_DEST directory (contains alert_.log), modify the ownership of directory BACKGROUND_DUMP_DEST so that the oracle user (software owner) is the owner and make the permissions on directory BACKGROUND_DUMP_DEST 755. Follow these steps: 1. Determine the location of the BACKGROUND_DUMP_DEST parameter defined in the init.ora or config.ora files. 2. Login as root. 3. Change directory to the location of BACKGROUND_DUMP_DEST. 4. Change the owner of all the files and the directory to the software owner. For example: % chown oracle * 5. Change the permissions on the directory to 755. % chmod 755 . Solution Explanation: ===================== Changing the ownership and permissions of the BACKGROUND_DUMP_DEST directory, enables oracle to write to the alert_.log file. --------------------------------------------------------------------------- Doc ID : Note:273413.1 Content Type: TEXT/X-HTML Subject: Database Does not Start, Ora-00604 Ora-25153 Ora-00604 Ora-1092 Creation Date: 19-MAY-2004 Type: PROBLEM Last Revision Date: 04-OCT-2004 Status: MODERATED The information in this article applies to: Oracle Server - Enterprise Edition - Version: 8.1.7.4 to 10.1.0.4 This problem can occur on any platform. Errors ORA-1092 Oracle instance terminated. ORA-25153 Temporary Tablespace is Empty ORA-604 error occurred at recursive SQL level Symptoms The database is not opening and in the alert.log the following errors are reported: ORA-00604: error occurred at recursive SQL level 1 ORA-25153: Temporary Tablespace is Empty Error 604 happened during db open, shutting down database USER: terminating instance due to error 604 Instance terminated by USER, pid = xxxxx ORA-1092 signalled during: alter database open... You might find SQL in the trace file like: select distinct d.p_obj#,d.p_timestamp from sys.dependency$ d, obj$ o where d.p_obj#>=:1 and d.d_obj#=o.obj# and o.status!=5 Cause In the case where there's locally managed temp tablespace in the database,after controlfile is re-created using the statement generated by "alter database backup controlfile to trace", the database can't be opened again because it complains that temp tablespace is empty. However no tempfiles can be added to the temp tablespace, nor can the temp tablespace be dropped because the database is not yet open. The query failed because of inadequate sort space(memory + disk) Fix We can increase the sort_area_size and sort_area_retained_size to a very high value so that the query completes. Then DB will open and we can take care of the TEMP tablespace If the error still persists after increasing the sort_area_size and sort_area_retained_size to a high vale, then the only option remains is to restore and recover. ------------------------------------------------------------------------------- Displayed below are the messages of the selected thread. Thread Status: Active From: Ronald Shaffer 17-Mar-05 19:23 Subject: Deleted OUTLN and now I get ORA-1092 and ORA-18008 RDBMS Version: 10G Operating System and Version: RedHat ES 3 Error Number (if applicable): ORA-1092 and ORA-18008 Product (i.e. SQL*Loader, Import, etc.): Product Version: Deleted OUTLN and now I get ORA-1092 and ORA-18008 One of our DBAs dropped the OUTLN user in 10G and now the instance will not start. We get an ORA-18008 specifying the schema is missing and an ORA-1092 when it attempts to OPEN. Startup mount is as far as we can get. Any experience with this issue out there? Thanks... From: Fairlie Rego 23-Mar-05 01:26 Subject: Re : Deleted OUTLN and now I get ORA-1092 and ORA-18008 Hi Ronald, You are hitting bug 3786479 AFTER DROPPING THE OUTLN USER/SCHEMA, DB WILL NO LONGER OPEN.ORA-18008 http://metalink.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=BUG&p_id=3786479 If this is still an issue file a Tar and get a backport. Regards, Fairlie Rego ---------------------------------------------------------------------------------- Displayed below are the messages of the selected thread. Thread Status: Closed From: Henry Lau 06-Mar-03 10:38 Subject: ORA-01092 while alter datbase open RDBMS Version: 9.0.1.3 Operating System and Version: Linux Redhat 7.1 Error Number (if applicable): ORA-01092 Product (i.e. SQL*Loader, Import, etc.): ORACLE DATABASE Product Version: 9.0.1.3 ORA-01092 while alter datbase open Hi, Since our undotbs is very large and we try to follow the Doc ID: 157278.1, we are trying to change the undotbs to a new one We try to 1. Create UNDO tablespace undotb2 datafile $ORACLE_HOME/oradata/undotb2.dbf size 300M 2. ALTER SYSTEM SET undo_tablespace=undotb2; 3. Change undo = undotb2; 4. Restart the database; 5. alter tablespace undotbs offline; 6. when we restart the database, it shows the following error. SQL> startup mount pfile=$ORACLE_HOME/admin/TEST/pfile/init.ora ORACLE instance started. Total System Global Area 386688540 bytes Fixed Size 280092 bytes Variable Size 318767104 bytes Database Buffers 67108864 bytes Redo Buffers 532480 bytes Database mounted. SQL> alter database nomount; alter database nomount * ERROR at line 1: ORA-02231: missing or invalid option to ALTER DATABASE SQL> alter database open; alter database open * ERROR at line 1: ORA-01092: ORACLE instance terminated. Disconnection forced I have checked the Log file as follow: SQL> /u01/oracle/product/9.0.1/admin/TEST/udump/ora_29151.trc Oracle9i Release 9.0.1.3.0 - Production JServer Release 9.0.1.3.0 - Production ORACLE_HOME = /u01/oracle/product/9.0.1 System name: Linux Node name: utxrho01.unitex.com.hk Release: 2.4.2-2smp Version: #1 SMP Sun Apr 8 20:21:34 EDT 2001 Machine: i686 Instance name: TEST Redo thread mounted by this instance: 1 Oracle process number: 9 Unix process pid: 29151, image: oracle@utxrho01.unitex.com.hk (TNS V1-V3) *** SESSION ID:(8.3) 2003-03-06 17:25:38.615 Evaluating checkpoint for thread 1 sequence 8 block 2 ORA-00376: file 2 cannot be read at this time ORA-01110: data file 2: '/u01/oracle/product/9.0.1/oradata/TEST/undotbs01.dbf' ~ ~ ~ ~ Please help to check what the problem is ?? Thank you !! Regards, Henry From: Oracle, Pravin Sheth 07-Mar-03 09:31 Subject: Re : ORA-01092 while alter datbase open Hi Henry, What you are seeing is bug 2360088, which is fixed in Oracle 9.2.0.2. I suggest that you log an iSR (formerly iTAR) for a quicker solution for the problem. Regards Pravin ----------------------------------------------------------------------------------- 19.41 ORA-600 [qerfxFetch_01] ============================= Note 1: ------- Doc ID: Note:255881.1 Subject: ORA-600 [qerfxFetch_01] Type: REFERENCE Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 10-NOV-2003 Last Revision Date: 12-NOV-2004 This note contains information that has not yet been reviewed by the PAA Internals group or DDR. As such, the contents are not necessarily accurate and care should be taken when dealing with customers who have encountered this error. If you are going to use the information held in this note then please take whatever steps are needed to in order to confirm that the information is accurate. Until the article has been set to EXTERNAL, we do not guarantee the contents. Thanks. PAA Internals Group (Note - this section will be deleted as the note moves to publication) Note: For additional ORA-600 related information please read Note 146580.1 PURPOSE: This article represents a partially published OERI note. It has been published because the ORA-600 error has been reported in at least one confirmed bug. Therefore, the SUGGESTIONS section of this article may help in terms of identifying the cause of the error. This specific ORA-600 error may be considered for full publication at a later date. If/when fully published, additional information will be available here on the nature of this error. PURPOSE: This article discusses the internal error "ORA-600 [qerfxFetch_01]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [qerfxFetch_01] VERSIONS: versions 9.2 DESCRIPTION: During database operations, user interrupts need to be handled correctly. ORA-600 [qerfxFetch_01] is raised when an interrupt has been trapped but has not been handled correctly. FUNCTIONALITY: Fixed table row source. IMPACT: NON CORRUPTIVE - No underlying data corruption. SUGGESTIONS: If the Known Issues section below does not help in terms of identifying a solution, please submit the trace files and alert.log to Oracle Support Services for further analysis. Known Issues: Bug# 2306106 See Note 2306106.8 OERI:[qerfxFetch_01] possible - affects OEM Fixed: 9.2.0.2, 10.1.0.2 INTERNAL ONLY SECTION - NOT FOR PUBLICATION OR DISTRIBUTION TO CUSTOMERS ======================================================================== Ensure that this note comes out on top in Metalink when searched ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 qerfxFetch_01 Note 2: ------- Doc ID : Note:2306106.8 Content Type: TEXT/X-HTML Subject: Support Description of Bug 2306106 Creation Date: 13-AUG-2003 Type: PATCH Last Revision Date: 14-AUG-2003 Status: PUBLISHED Click here for details of sections in this note. Bug 2306106 OERI:[qerfxFetch_01] possible - affects OEM This note gives a brief overview of bug 2306106. Affects: Product (Component) Oracle Server (RDBMS) Range of versions believed to be affected Versions >= 9.2 but < 10G Versions confirmed as being affected 9.2.0.1 Platforms affected Generic (all / most platforms affected) Fixed: This issue is fixed in 9.2.0.2 (Server Patch Set) 10G Production Base Release Symptoms: Error may occur Internal Error may occur (ORA-600) ORA-600 [qerfxFetch_01] Related To: (None Specified) Description ORA-600 [qerfxFetch_01] possible - affects OEM Note 3: ------- Bug 2306106 is fixed in the 9.2.0.2 patchset. This bug is not published and thus cannot be viewed externally in MetaLink. All it says on this bug is 'ORA-600 [qerfxFetch_01] possible - affects OEM'. 19.42 Undo corruption: ====================== Note 1: ------- Doc ID : Note:2431450.8 Content Type: TEXT/X-HTML Subject: Support Description of Bug 2431450 Creation Date: 08-AUG-2003 Type: PATCH Last Revision Date: 05-JAN-2004 Status: PUBLISHED Click here for details of sections in this note. Bug 2431450 SMU Undo corruption possible on instance crash This note gives a brief overview of bug 2431450. Affects: Product (Component) (Rdbms) Range of versions believed to be affected Versions >= 9 but < 10G Versions confirmed as being affected 9.0.1.4 9.2.0.3 Platforms affected Generic (all / most platforms affected) Fixed: This issue is fixed in 9.0.1.5 iAS Patch Set 9.2.0.4 (Server Patch Set) 10g Production Base Release Symptoms: Corruption (Physical) Internal Error may occur (ORA-600) ORA-600 [kteuPropTime-2] / ORA-600 [4191] Related To: System Managed Undo Description SMU (System Managed Undo) Undo corruption possible on instance crash. This can result in subsequent ORA-600 errors due to the undo corruption. Note 2: ------- Doc ID : Note:233864.1 Content Type: TEXT/X-HTML Subject: ORA-600 [kteuproptime-2] Creation Date: 28-MAR-2003 Type: REFERENCE Last Revision Date: 07-APR-2005 Status: PUBLISHED Note: For additional ORA-600 related information please read Note 146580.1 PURPOSE: This article discusses the internal error "ORA-600 [kteuproptime-2]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [kteuproptime-2] VERSIONS: versions 9.0 to 9.2 DESCRIPTION: Oracle has encountered an error propagating Extent Commit Times in the Undo Segment Header / Extent Map Blocks, for System Managed Undo Segments The extent being referenced is not valid. FUNCTIONALITY: UNDO EXTENTS IMPACT: INSTANCE FAILURE POSSIBLE PHYSICAL CORRUPTION SUGGESTIONS: If instance is down and fails to restart due to this error then set the following parameter, which will gather additional information to assist support in identifing the cause: # Dump Undo Segment Headers during transaction recovery event="10015 trace name context forever, level 10" Restart the instance and submit the trace files and alert.log to Oracle Support Services for further analysis. Do not set any other undo/rollback_segment parameters without direction from Support. Known Issues: Bug# 2431450 See Note 2431450.8 SMU Undo corruption possible on instance crash Fixed: 9.2.0.4, 10.1.0.2 Note 3: ------- Hi, apply patchset 9.2.0.2, bug 2431450 is fixed in 9.2.0.2 that made SMU (System Managed Undo) Undo corruption possible on instance crash. It's a very rare scenario : This will only cause a problem if there was an instance crash after a transaction committed but before it propogated the extent commit times to all its extents AND there was a shrink of extents before the transaction could be recovered. But still, this bug was not published (not for any particular reason except it was found internal). Greetings, Note 4: ------- From: Oracle, Ken Robinson 21-Feb-03 17:44 Subject: Re : ORA-600 kteuPropTime-2 Forgot to mention the second bug for this....bug 2689239. Regards, Ken Robinson Oracle Server EE Analyst ORA-600 [4191] possible on shrink of system managed undo segment. Note 5: ------- BUGBUSTER - System-managed undo segment corruption Affects Versions: 9.2.0.1.0, 9.2.0.2.0, 9.2.0.3.0 Fixed in: Patch 2431450, 9.2.0.4.0 BUG# (if recognised) 2431450 This info. correct on: 31-AUG-2003 Symptoms Oracle instance crashes and details of the ORA-00600 error are written to the alert.log ORA-00600: internal error code, arguments: [kteuPropTime-2], [], [], [] Followed by Fatal internal error happened while SMON was doing active transaction recovery. Then SMON: terminating instance due to error 600 Instance terminated by SMON, pid = 22972 This occurs as Oracle encounters an error when propagating Extent Commit Times in the Undo Segment Header Extent Map Blocks. It could be because SMON is over-enthusiastic in shrinking extents in SMU segments. As a result, extent commit times do not get written to all the extents and SMON causes the instance to crash, leaving one or more of the undo segments corrupt. When opening the database following the crash, Oracle tries to perform crash recovery and encounters problems recovering committed transactions stored in the corrupt undo segments. This leads to more ORA-00600 errors and a further instance crash. The net result is that the database cannot be opened: "Error 600 happened during db open, shutting down database" Workaround Until the corrupt undo segment can be identified and offlined then unfortunately the database will not open. Identify the corrupt undo segment by setting the following parameters in the init.ora file: _smu_debug_mode=1 event="10015 trace name context forever, level 10" (set event 10511) event="10511 trace name context forever, level 2" _smu_debug_mode simply collects diagnostic information for support purposes. Event 10015 is the undo segment recovery tracing event. Use this to identify corrupted rollback/undo segments when a database cannot be started. With these parameters set, an attempt to open the database will still cause a crash, but Oracle will write vital information about the corrupt rollback/undo segments to a trace file in user_dump_dest. This is an extract from such a trace file, revealing that undo segment number 6 (_SYSSMU6$) is corrupt. Notice that the information stored in the segment header about the number of extents was inconsistent with the extent map. Recovering rollback segment _SYSSMU6$ UNDO SEG (BEFORE RECOVERY): usn = 6 Extent Control Header ----------------------------------------------------------------- Extent Header:: spare1: 0 spare2: 0 #extents: 7 #blocks: 1934 last map 0x00805f89 #maps: 1 offset: 4080 Highwater:: 0x0080005b ext#: 0 blk#: 1 ext size: 7 #blocks in seg. hdr's freelists: 0 #blocks below: 0 mapblk 0x00000000 offset: 0 Unlocked Map Header:: next 0x00805f89 #extents: 5 obj#: 0 flag: 0x40000000 Extent Map ----------------------------------------------------------------- 0x0080005a length: 7 0x00800061 length: 8 0x0081ac89 length: 1024 0x00805589 length: 256 0x00805a89 length: 256 Retention Table ----------------------------------------------------------- Extent Number:0 Commit Time: 1060617115 Extent Number:1 Commit Time: 1060611728 Extent Number:2 Commit Time: 1060611728 Extent Number:3 Commit Time: 1060611728 Extent Number:4 Commit Time: 1060611728 Comment out parameters undo_management and undo_tablespace and set the undocumented _corrupted_rollback_segments parameter to tell Oracle to ignore any corruptions and force the database open: _corrupted_rollback_segments=(_SYSSMU6$) This time, Oracle will start and open OK, which will allow you to check the status of the undo segments by querying DBA_ROLLBACK_SEGS. select segment_id, segment_name, tablespace_name, status from dba_rollback_segs where owner='PUBLIC'; SEGMENT_ID SEGMENT_NAME TABLESPACE_NAME STATUS ---------- ------------ --------------- ---------------- 1 _SYSSMU1$ UNDOTS OFFLINE 2 _SYSSMU2$ UNDOTS OFFLINE 3 _SYSSMU3$ UNDOTS OFFLINE 4 _SYSSMU4$ UNDOTS OFFLINE 5 _SYSSMU5$ UNDOTS OFFLINE 6 _SYSSMU6$ UNDOTS NEEDS RECOVERY 7 _SYSSMU7$ UNDOTS OFFLINE 8 _SYSSMU8$ UNDOTS OFFLINE 9 _SYSSMU9$ UNDOTS OFFLINE 10 _SYSSMU10$ UNDOTS OFFLINE SMON will complain every 5 minutes by writing entries to the alert.log as long as there are undo segments in need of recovery SMON: about to recover undo segment 6 SMON: mark undo segment 6 as needs recovery At this point, you must either download and apply patch 2431450 or create private rollback segments. Note 6: ------- Repair UNDO log corruption Don Burleson In rare cases (usually DBA error) the Oracle UNDO tablespace can become corrupted. This manifests with this error: ORA-00376: file xx cannot be read at this time In cases of UNDO log corruption, you must: • Change the undo_management parameter from “AUTO” to “MANUAL” • Create a new UNDO tablespace • Drop the old UNDO tablespace Dropping the corrupt UNDO tablespace can be tricky and you may get the message: ORA-00376: file string cannot be read at this time To drop a corrupt UNDO tablespace: 1 – Identify the bad segment: select segment_name, status from dba_rollback_segs where tablespace_name='undotbs_corrupt' and status = ‘NEEDS RECOVERY’; SEGMENT_NAME STATUS ------------------------------ ---------------- _SYSSMU22$ NEEDS RECOVERY 2. Bounce the instance with the hidden parameter “_offline_rollback_segments”, specifying the bad segment name: _OFFLINE_ROLLBACK_SEGMENTS=_SYSSMU22$ 3. Bounce database, nuke the corrupt segment and tablespace: SQL> drop rollback segment "_SYSSMU22$"; Rollback segment dropped. SQL > drop tablespace undotbs including contents and datafiles; Tablespace dropped. Note 7: ------- Sometimes there can be trouble with an undo segment. Actually there might be something with a normal object: PUT the following in the init.ora- event = "10015 trace name context forever, level 10" Setting this event will generate a trace file that will reveal the necessary information about the transaction Oracle is trying to rollback and most importantly, what object Oracle is trying to apply the undo to. USE the following query to find out what object Oracle is trying to perform recovery on. select owner, object_name, object_type, status from dba_objects where object_id = ; THIS object must be dropped so the undo can be released. An export or relying on a backup may be necessary to restore the object after the corrupted rollback segment goes away. 19.43 ORA-1653 ============== Note 1: ------- Doc ID : Note:151994.1 Content Type: TEXT/PLAIN Subject: Overview Of ORA-01653: Unable To Extend Table %s.%s By %s In Tablespace %s Creation Date: 12-JUL-2001 Type: TROUBLESHOOTING Last Revision Date: 15-JUN-2004 Status: PUBLISHED PURPOSE ------- This bulletin is an overview of ORA-1653 error message for tablespace dictionary managed. SCOPE& APPLICATION ------------------ It is for users requiring further information on ORA-01653 error message. When looking to resolve the error by using any of the solutions suggested, please consult the DBA for assistance. Error: ORA-01653 Text: unable to extend table %s.%s by %s in tablespace %s ------------------------------------------------------------------------------- Cause: Failed to allocate an extent for table segment in tablespace. Action: Use ALTER TABLESPACE ADD DATAFILE statement to add one or more files to the tablespace indicated. Explanation: ------------ This error does not necessarily indicate whether or not you have enough space in the tablespace, it merely indicates that Oracle could not find a large enough area of free contiguous space in which to fit the next extent. Diagnostic Steps: ----------------- 1. In order to see the free space available for a particular tablespace, you must use the view DBA_FREE_SPACE. Within this view, each record represents one fragment of space. How the view DBA_FREE_SPACE can be used to determine the space available in the database is described in: [NOTE:121259.1] Using DBA_FREE_SPACE 2. The DBA_TABLES view describes the size of next extent (NEXT_EXTENT) and the percentage increase (PCT_INCREASE) for all tables in the database. The "next_extent" size is the size of extent that is trying to be allocated (and for which you have the error). When the extent is allocated : next_extent = next_extent * (1 + (pct_increase/100)) Algorythm to allocate extent for segment is described in the Concept Guide Chapter : Data Blocks, Extents, and Segments - How Extents Are Allocated 3. Look to see if any users have the tablespace in question as their temporary tablespace. This can be checked by looking at DBA_USERS (TEMPORARY_TABLESPACE). Possible solutions: ------------------- - Manually Coalesce Adjacent Free Extents ALTER TABLESPACE COALESCE; The extents must be adjacent to each other for this to work. - Add a Datafile: ALTER TABLESPACE ADD DATAFILE '' SIZE ; - Resize the Datafile: ALTER DATABASE DATAFILE '' RESIZE ; - Enable autoextend: ALTER DATABASE DATAFILE '' AUTOEXTEND ON MAXSIZE UNLIMITED; - Defragment the Tablespace: - Lower "next_extent" and/or "pct_increase" size: ALTER STORAGE ( next pctincrease ); - If the tablespace is being used as a temporary tablespace, temporary segments may be still holding the space. References: ----------- [NOTE:1025288.6] How to Diagnose and Resolve ORA-01650, ORA-01652, ORA-01653, ORA-01654, ORA-01688 : Unable to Extend < OBJECT > by %S in Tablespace [NOTE:1020090.6] Script to Report on Space in Tablespaces [NOTE:1020182.6] Script to Detect Tablespace Fragmentation [NOTE:1012431.6] Overview of Database Fragmentation [NOTE:121259.1] Using DBA_FREE_SPACE [NOTE:61997.1] SMON - Temporary Segment Cleanup and Free Space Coalescing Note 2: ------- Doc ID : Note:1025288.6 Content Type: TEXT/PLAIN Subject: How to Diagnose and Resolve ORA-01650,ORA-01652,ORA-01653,ORA-01654,ORA-01688 : Unable to Extend < OBJECT > by %S in Tablespace %S Creation Date: 02-JAN-1997 Type: TROUBLESHOOTING Last Revision Date: 10-JUN-2004 Status: PUBLISHED PURPOSE ------- This document can be used to diagnose and resolve space management errors - ORA-1650, ORA-1652, ORA-1653, ORA-1654 and ORA-1688. SCOPE & APPLICATION ------------------- You are working with the database and have encountered one of the following errors: ORA-01650: unable to extend rollback segment %s by %s in tablespace %s Cause: Failed to allocate extent for the rollback segment in tablespace. Action: Use the ALTER TABLESPACE ADD DATAFILE statement to add one or more files to the specified tablespace. ORA-01652: unable to extend temp segment by %s in tablespace %s Cause: Failed to allocate an extent for temp segment in tablespace. Action: Use ALTER TABLESPACE ADD DATAFILE statement to add one or more files to the tablespace indicated or create the object in other tablespace. ORA-01653: unable to extend table %s.%s by %s in tablespace %s Cause: Failed to allocate extent for table segment in tablespace. Action: Use the ALTER TABLESPACE ADD DATAFILE statement to add one or more files to the specified tablespace. ORA-01654: unable to extend index %s.%s by %s in tablespace %s Cause: Failed to allocate extent for index segment in tablespace. Action: Use the ALTER TABLESPACE ADD DATAFILE statement to add one or more files to the specified tablespace. ORA-01688: unable to extend table %s.%s partition %s by %s in tablespace %s Cause: Failed to allocate an extent for table segment in tablespace. Action: Use ALTER TABLESPACE ADD DATAFILE statement to add one or more files to the tablespace indicated. How to Solve the Following Errors About UNABLE TO EXTEND -------------------------------------------------------- An "unable to extend" error is raised when there is insufficient contiguous space available to extend the object. A. In order to address the UNABLE TO EXTEND issue, you need to get the following information: 1. The largest contiguous space available for the tablespace SELECT max(bytes) FROM dba_free_space WHERE tablespace_name = ''; The above query returns the largest available contiguous chunk of space. Please note that if the tablespace you are concerned with is of type TEMPORARY, then please refer to [NOTE:188610.1] . If this query is done immediately after the failure, it will show that the largest contiguous space in the tablespace is smaller than the next extent the object was trying to allocate. 2. => "next_extent" for the object => "pct_increase" for the object => The name of the tablespace in which the object resides Use the "next_extent" size with "pct_increase" in the following formula to determine the size of extent that is trying to be allocated. extent size = next_extent * (1 + (pct_increase/100) next_extent = 512000 pct_increase = 50 => extent size = 512000 * (1 + (50/100)) = 512000 * 1.5 = 768000 ORA-01650 Rollback Segment ========================== SELECT next_extent, pct_increase, tablespace_name FROM dba_rollback_segs WHERE segment_name = ''; Note: pct_increase is only needed for early versions of Oracle, by default in later versions pct_increase for a rollback segment is 0. ORA-01652 Temporary Segment =========================== SELECT next_extent, pct_increase, tablespace_name FROM dba_tablespaces WHERE tablespace_name = ''; Temporary segments take the default storage clause of the tablespace in which they are created. If this error is caused by a query, then try and ensure that the query is tuned to perform its sorts as efficiently as possible. To find the owner of a sort, please refer to [NOTE:1069041.6] ORA-01653 Table Segment ======================= SELECT next_extent, pct_increase , tablespace_name FROM dba_tables WHERE table_name = '' AND owner = ''; ORA-01654 Index Segment ======================= SELECT next_extent, pct_increase, tablespace_name FROM dba_indexes WHERE index_name = '' AND owner = ''; ORA-01688 Table Partition ========================= SELECT next_extent, pct_increase, tablespace_name FROM dba_tab_partitions WHERE partition_name='' AND table_owner = ''; B. Possible Solutions There are several options for solving errors due to failure to extend: a. Manually Coalesce Adjacent Free Extents --------------------------------------- ALTER TABLESPACE COALESCE; The extents must be adjacent to each other for this to work. b. Add a Datafile -------------- ALTER TABLESPACE ADD DATAFILE '' SIZE ; c. Lower "next_extent" and/or "pct_increase" size ---------------------------------------------- For non-temporary and non-partitioned segment problem: ALTER STORAGE ( next pctincrease ); For non-temporary and partitioned segment problem: ALTER TABLE MODIFY PARTITION STORAGE ( next pctincrease ); For a temporary segment problem: ALTER TABLESPACE DEFAULT STORAGE (initial next pctincrease ); d. Resize the Datafile ------------------- ALTER DATABASE DATAFILE '' RESIZE ; e. Defragment the Tablespace ------------------------- If you would like more information on fragmentation, the following documents are available from Oracle WorldWide Support . (this is not a comprehensive list) [NOTE:1020182.6] Script to Detect Tablespace Fragmentation [NOTE:1012431.6] Overview of Database Fragmentation [NOTE:30910.1] Recreating Database Objects Related Documents: ================== [NOTE:15284.1] Understanding and Resolving ORA-01547 Overview Of ORA-01653 Unable To Extend Table %s.%s By %s In Tablespace %s: Overview Of ORA-01654 Unable To Extend Index %s.%s By %s In Tablespace %s: [NOTE:188610.1] DBA_FREE_SPACE Does not Show Information about Temporary Tablespaces [NOTE:1069041.6] How to Find Creator of a SORT or TEMPORARY SEGMENT or Users Performing Sorts for Oracle8 and 9 Search Words: ============= ORA-1650 ORA-1652 ORA-1653 ORA-1654 ORA-1688 ORA-01650 ORA-01652 ORA-01653 ORA-01654 ORA-01688 1650 1652 1653 1654 1688 19.44: Other ORA- errors on 9i: =============================== Doc ID : Note:201342.1 Content Type: TEXT/X-HTML Subject: Top Internal Errors - Oracle Server Release 9.2.0 Creation Date: 27-JUN-2002 Type: BULLETIN Last Revision Date: 24-MAY-2004 Status: PUBLISHED Top Internal Errors - Oracle Server Release 9.2.0 Additional information or documentation on ORA-600 errors not listed here may be available from the ORA-600 Lookup tool : > > Oracle9i Release 2 (9.2) Support Status and Alerts ORA-600 [KSLAWE:!PWQ] Possible bugs: Fixed in: > BACKGROUND PROCESS GOT OERI:KSLAWE:!PWQ AND INSTANCE CRASHES 9.2.0.6, 10G References: > ALERT: ORA-600[KSLAWE:!PWQ] RAISED IN V92040 OR V92050 ON SUN 64BIT ORACLE ORA-600 [ksmals] Possible bugs: Fixed in: > ORA-7445 & HEAP CORRUPTION WHEN RUNNING APPS PROGRAM THAT DOES HEAVY INSERTS 9.2.0.4 References: > ORA-600 [ksmals] ORA-600 [4000] Possible bugs: Fixed in: > STARTUP after an ORA-701 fails with OERI[4000] 9.2.0.5, 10G > OERI:4506 / OERI:4000 possible against transported tablespace 8.1.7.4, 9.0.1.4, 9.2.0.1 References: > ORA-600 [4000] "trying to get dba of undo segment header block from usn" ORA-600 [4454] Possible bugs: Fixed in: > OERI:4411/OERI:4454 on long running job 8.1.7.3, 9.0.1.3, 9.2.0.1 References: > ORA-600 [4454] ORA-600 [kcbgcur_9] Possible bugs: Fixed in: > OERI:kcbgcur_9 on direct load into AUTO space managed segment 9.2.0.4, 10G > Direct path load may fail with OERI:kcbgcur_9 / OERI:ktfduedel2 9.2.0.4, 10G > OERI:KCBGCUR_9 possible from SMON dropping a rollback segment in locally managed tablespace 9.0.1.4, 9.2.0.1 > OERI:KCBGCUR_9 possible during TEMP space operations 9.0.1.3, 9.2.0.1 > OERI:KCBGCUR_9 possible from ONLINE REBUILD INDEX with concurrent DML 8.1.7.3, 9.0.1.3, 9.2.0.1 > OERI:kcbgcur_9 from CLOB TO CHAR or BLOB TO RAW conversion 9.2.0.2, 10G References: > ORA-600 [kcbgcur_9] "Block class pinning violation" ORA-600 [qerrmOFBu1], [1003] Possible bugs: Fixed in: > SQL*PLUS CRASH IN TTC LOGGING INTO ORACLE 7.3.4 DATABASE References: > ORA-600 [qerrmOFBu1] - "Error during remote row fetch operation > ALERT: Connections from Oracle 9.2 to Oracle7 are Not Supported ORA-600 [ktsgsp5] or ORA-600 [kdddgb2] Possible bugs: Fixed in: > ORA-600 [KDDDGB2] [435816] [2753588] & PROBABLE INDEX CORRUPTION 9.2.0.2 References: > ORA-600 [kdddgb2] > ORA-600 [ktsgsp5] > ALERT: Corruption / Internal Errors possible after Upgrading to 9.2.0.1 19.45: ADJUST SCN: ================== Note 1 Adjust SCN: ------------------ Doc ID: Note:30681.1 Subject: EVENT: ADJUST_SCN - Quick Reference Type: REFERENCE Status: PUBLISHED Content Type: TEXT/PLAIN Creation Date: 20-OCT-1997 Last Revision Date: 04-AUG-2000 Language: USAENG ADJUST_SCN Event ~~~~~~~~~~~~~~~~ *** WARNING *** This event should only ever be used under the guidance of an experienced Oracle analyst. If an SCN is ahead of the current database SCN, this indicates some form of database corruption. The database should be rebuilt after bumping the SCN. **************** The ADJUST_SCN event is useful in some recovery situations where the current SCN needs to be incremented by a large value to ensure it is ahead of the highest SCN in the database. This is typically required if either: a. An ORA-600 [2662] error is signalled against database blocks or b. ORA-1555 errors keep occuring after forcing the database open or ORA-604 / ORA-1555 errors occur during database open. (Note: If startup reports ORA-704 & ORA-1555 errors together then the ADJUST_SCN event cannot be used to bump the SCN as the error is occuring during bootstrap. Repeated startup/shutdown attempts may help if the SCN mismatch is small) or c. If a database has been forced open used _ALLOW_RESETLOGS_CORRUPTION (See ) The ADJUST_SCN event acts as described below. **NOTE: You can check that the ADJUST_SCN event has fired as it should write a message to the alert log in the form "Debugging event used to advance scn to %s". If this message is NOT present in the alert log the event has probably not fired. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ If the database will NOT open: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Take a backup. You can use event 10015 to trigger an ADJUST_SCN on database open: startup mount; alter session set events '10015 trace name adjust_scn level 1'; (NB: You can only use IMMEDIATE here on an OPEN database. If the database is only mounted use the 10015 trigger to adjust SCN, otherwise you get ORA 600 [2251], [65535], [4294967295] ) alter database open; If you get an ORA 600:2256 shutdown, use a higher level and reopen. Do *NOT* set this event in init.ora or the instance will crash as soon as SMON or PMON try to do any clean up. Always use it with the "alter session" command. ~~~~~~~~~~~~~~~~~~~~~~~~~~ If the database *IS* OPEN: ~~~~~~~~~~~~~~~~~~~~~~~~~~ You can increase the SCN thus: alter session set events 'IMMEDIATE trace name ADJUST_SCN level 1'; LEVEL: Level 1 is usually sufficient - it raises the SCN to 1 billion (1024*1024*1024) Level 2 raises it to 2 billion etc... If you try to raise the SCN to a level LESS THAN or EQUAL to its current setting you will get - See below. Ie: The event steps the SCN to known levels. You cannot use the same level twice. Calculating a Level from 600 errors: ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ To get a LEVEL for ADJUST_SCN: a) Determine the TARGET scn: ora-600 [2662] See Use TARGET >= blocks SCN ora-600 [2256] See Use TARGET >= Current SCN b) Multiply the TARGET wrap number by 4. This will give you the level to use in the adjust_scn to get the correct wrap number. c) Next, add the following value to the level to get the desired base value as well : Add to Level Base ~~~~~~~~~~~~ ~~~~~~~~~~~~ 0 0 1 1073741824 2 2147483648 3 3221225472 Note 2: Adjust SCN ------------------ Subject: OERR: 600 2662 Block SCN is ahead of Current SCN Creation Date: 21-OCT-1997 ORA-600 [2662] [a] [b] [c] [d] [e] Versions: 7.0.16 - 8.0.5 Source: kcrf.h =========================================================================== Meaning: There are 3 forms of this error. 4/5 argument forms - The SCN found on a block (dependant SCN) was ahead of the current SCN. See below for this 1 Argument (before 7.2.3): Oracle is in the process of writing a block to a log file. If the calculated block checksum is less than or equal to 1 (0 and 1 are reserved) ORA-600 [2662] is returned. This is a problem generating an offline immediate log marker (kcrfwg). *NOT DOCUMENTED HERE* --------------------------------------------------------------------------- Argument Description: Until version 7.2.3 this internal error can be logged for two separate reasons, which we will refer to as type I and type II. The two types can be distinguished by the number of arguments: Type I has four or five arguments after the [2662]. Type II has one argument after the [2662]. From 7.2.3 onwards type II no longer exists. Type I ~~~~~~ a. Current SCN WRAP b. Current SCN BASE c. dependant SCN WRAP d. dependant SCN BASE e. Where present this is the DBA where the dependant SCN came from. From kcrf.h: If the SCN comes from the recent or current SCN then a dba of zero is saved. If it comes from undo$ because the undo segment is not available then the undo segment number is saved, which looks like a block from file 0. If the SCN is for a media recovery redo (i.e. block number == 0 in change vector), then the dba is for block 0 of the relevant datafile. If it is from another database for distribute xact then dba is DBAINF(). If it comes from a TX lock then the dba is really usn<<16+slot. Type II ~~~~~~~ a. checksum -> log block checksum - zero if none (thread # in old format) --------------------------------------------------------------------------- Diagnosis: ~~~~~~~~~~ In addition to different basic types from above, there are different situations and coherences where ORA-600 [2662] type 'I' can be raised. For diagnosis we can split up startup-issues and no-startup-issues. Usually the startup-issues are more critical. Getting started: ~~~~~~~~~~~~~~~~ (1) is the error raised during normal database operations (i.e. when the database is up) or during startup of the database? (2) what is the SCN difference [d]-[b] ( subtract argument 'b' from arg 'd')? (3) is there a fifth argument [e] ? If so convert the dba to file# block# Is it a data dictionary object? (file#=1) If so find out object name with the help of reference dictionary from second database (4) What is the current SQL statement? (see trace) Which table is refered to? Does the table match the object you found in step before? Be careful at this point: there may be no relationship between DBA in [e] and real source of problem (blockdump). Deeper analysis: ~~~~~~~~~~~~~~~~ - investigate trace file this will be a user trace file normally but could be an smon trace too - search for: 'buffer' ("buffer dba" in Oracle7 dumps, "buffer tsn" in Oracle8 dumps) this will bring you to a blockdump which usually represents the 'real' source of OERI:2662 WARNING: There may be more than one buffer pinned to the process so ensure you check out all pinned buffers. -> does the blockdump match the dba from e.? -> what kind of blockdump is it? (a) rollbacksegment header (b) datablock (c) other SEE BELOW for EXAMPLES which demonstrate the sort of output you may see in trace files and the things to check. Check list and possible causes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - If Parallel Server check both nodes are using the same lock manager instance & point at the same control files. - If not Parallel Server check that 2 instances haven't mounted the same database (Is there a second PMON process around ?? - shut down any other instances to be sure) Possible causes: - doing an open resetlogs with _ALLOW_RESETLOGS_CORRUPTION enabled - a hardware problem, like a faulty controller, resulting in a failed write to the control file or the redo logs - restoring parts of the database from backup and not doing the appropriate recovery - restoring a control file and not doing a RECOVER DATABASE USING BACKUP CONTROLFILE - having _DISABLE_LOGGING set during crash recovery - problems with the DLM in a parallel server environment - a bug Solutions: - if the SCNs in the error are very close: Attempting a startup several times will bump up the dscn every time we open the database even if open fails. The database will open when dscn=scn. - ** You can bump the SCN on open using See [NOTE:30681.1] Be aware that you should really rebuild the database if you use this option. - Once this has occurred you would normally want to rebuild the database via exp/rebuild/imp as there is no guarantee that some other blocks are not ahead of time. Articles: ~~~~~~~~~ Solutions: [NOTE:30681.1] Details of the ADJUST_SCN Event [NOTE:1070079.6] alter system checkpoint Possible Causes: [NOTE:1021243.6] CHECK INIT.ORA SETTING _DISABLE_LOGGING [NOTE:74903.1] How to Force the Database Open (_ALLOW_RESETLOGS_CORRUPTION) [NOTE:41399.1] Forcing the database open with `_ALLOW_RESETLOGS_CORRUPTION` [NOTE:851959.9] OERI:2662 DURING CREATE SNAPSHOT AT MASTER SITE Known Bugs: ~~~~~~~~~~~ Fixed In. Bug No. Description ---------+------------+---------------------------------------------------- 7.0.14 BUG:153638 7.1.5 BUG:229873 7.1.3 Bug:195115 Miscalculation of SCN on startup for distributed TX ? 7.1.6.2.7 Bug:297197 Port specific Solaris OPS problem 7.3 Bug:336196 Port specific IBM SP AIX problem -> dlm issue 7.3.4.5 Bug:851959 OERI:2662 possible from distributed OPS select --------------------------------------------------------------------------- --------------------------------------------------------------------------- Examples: ~~~~~~~~ Below are some examples of this type of error and the information you will see in the trace files. ~~~~~~~~~~ CASE (a) ~~~~~~~~~~ blockdump should look like this: *** buffer dba: 0x05000002 inc: 0x00000001 seq: 0x0001a9c6 ver: 1 type: 1=KTU UNDO HEADER Extent Control Header ----------------------------------------------------------------- Extent Control:: inc#: 716918 tsn: 4 object#: 0 *** -> interpret: dba: 0x05000002 -> 83886082 (0x05000002) = 5,2 XXX tsn: 4 -> this is rollback segment 4 tsn: 4 -> this rollback segment is in tablespace 4 ORA-00600: Interner Fehlercode, Argumente: [2662], [0], [71183], [0], [71195], [83886082], [], [] -> [e] > 0 and represents dba from block which is in trace -> [d]-[b] = 71195 - 71183 = 12 -> convert [b] to hex: 71195 = 0x1161B so this value can be found in blockdump: *** TRN TBL:: index state cflags wrap# uel scn dba ------------------------------------------------------------------ ... 0x4e 9 0x00 0x00d6 0xffff 0x0000.0001161b 0x00000000 ... *** -> possible cause so in this case the CURRENT SCN is LOWER than the SCN on this transaction ie: The current SCN looks like it has decreased !! This could happen if the database is opened with the _allow_resetlogs_corruption parameter -> If some recovery steps have just been performed review these steps as the mismatch may be due to open resetlogs with _allow_resetlogs_corruption enabled or similar. See for information on this parameter. ------------------------------------------------------------------ ~~~~~~~~~~ CASE (b) ~~~~~~~~~~ blockdump looks like this: *** buffer dba: 0x0100012f inc: 0x00000815 seq: 0x00000d48 ver: 1 type: 6=trans data Block header dump: dba: 0x0100012f Object id on Block? Y seg/obj: 0xe csc: 0x00.5fed6 itc: 2 flg: O typ: 1 - DATA fsl: 0 fnx: 0x0 Itl Xid Uba Flag Lck Scn/Fsc 0x01 0x0000.00b.0000036c 0x0100261c.0138.04 --U- 1 fsc 0x0000.0005fed7 0x02 0x0000.00a.0000037b 0x0100261d.0138.01 --U- 1 fsc 0x0000.0005fed4 data_block_dump =============== ... *** interpret: dba: 0x0100012f -> 8,10 ==> 16777519 (0x0100012f) = 1,303 (0x1 0x12f) *** SVRMGR> SELECT SEGMENT_NAME, SEGMENT_TYPE FROM DBA_EXTENTS 2> WHERE FILE_ID = 1 AND 303 BETWEEN BLOCK_ID AND 3> BLOCK_ID + BLOCKS - 1; SEGMENT_NAME SEGMENT_TYPE ---------------------------------------------------------- ----------------- UNDO$ TABLE 1 row selected. *** -> current sql-statement (trace): *** update undo$ set name=:2,file#=:3,block#=:4,status$=:5,user#=:6, undosqn=:7,xactsqn=:8,scnbas=:9,scnwrp=:10,inst#=:11 where us#=:1 ksedmp: internal or fatal error ORA-00600: internal error code, arguments: [2662], [0], [392916], [0], [392919], [0], [], [] *** -> e. = 0 info not available -> d-b = 392919 - 392916 = 3 -> dba from blockdump matches the object from current sql statement -> convert b. to hex: = 0x5FED7 so this value can be found in blockdump -> see ITL slot 0x01! --------------------------------------------------------------------------- --------------------------------------------------------------------------- --------------------------------------------------------------------------- Some more internals: ~~~~~~~~~~~~~~~~~~~~ I will try to give another example in oder to answer question if current SCN is decreased or dependant SCN increase. hypothesis: current SCN decreased Evidence: reproduced ORA-600 [2662] by aborting tx and using _allow_resetlog_corruption while open resetlogs. check database SCN before! Prerequisits: _allow_resetlogs_corruption = true in init.ora shutdown/startup db *** BEGIN TESTCASE SVRMGR> drop table tx; Statement processed. SVRMGR> create table tx (scn# number); Statement processed. SVRMGR> insert into tx values( userenv('COMMITSCN') ); 1 row processed. SVRMGR> select * from tx; SCN# ---------- 392942 1 row selected. ************ another session ************** SQL> connect scott/tiger Connected. SQL> update emp set sal=sal+1; 13 rows processed. SQL> -- no commit here ******************************************* SVRMGR> insert into tx values( userenv('COMMITSCN') ); 1 row processed. SVRMGR> select * from tx; SCN# ---------- 392942 392943 2 rows selected. -- so current SCN will be 392943 SVRMGR> shutdown abort ORACLE instance shut down. -- this breaks tx SVRMGR> startup mount pfile=e:\jv734\initj734.ora ORACLE instance started. Total System Global Area 11018952 bytes Fixed Size 35760 bytes Variable Size 7698200 bytes Database Buffers 3276800 bytes Redo Buffers 8192 bytes Database mounted. SVRMGR> recover database until cancel; ORA-00279: Change 392925 generated at 10/26/99 17:13:03 needed for thread 1 ORA-00289: Suggestion : e:\jv734\arch\arch_2.arc ORA-00280: Change 392925 for thread 1 is in sequence #2 Specify log: {=suggested | filename | AUTO | CANCEL} cancel Media recovery cancelled. SVRMGR> alter database open resetlogs; alter database open resetlogs * ORA-00600: internal error code, arguments: [2662], [0], [392928], [0], [392931], [0], [], [] *** END TESTCASE because we know current SCN before (392943) we see, that current SCN has decreased after solving the problem with: shutdown abort/startup -> works SVRMGR> drop table tx; Statement processed. SVRMGR> create table tx (scn# number); Statement processed. SVRMGR> insert into tx values( userenv('COMMITSCN') ); 1 row processed. SVRMGR> select * from tx; SCN# ---------- 392943 1 row selected. so we have exactly reached the current SCN from before 'shutdown abort' So current SCN was bumpt up from 392928 to 392942. Note 3: Adjust SCN ------------------ Doc ID : Note:28929.1 Content Type: TEXT/X-HTML Subject: ORA-600 [2662] "Block SCN is ahead of Current SCN" Creation Date: 21-OCT-1997 Type: REFERENCE Last Revision Date: 15-OCT-2004 Status: PUBLISHED This note contains information that was not reviewed by DDR. As such, the contents are not necessarily accurate and care should be taken when dealing with customers who have encountered this error. Thanks. PAA Internals Group Note: For additional ORA-600 related information please read Note 146580.1 PURPOSE: This article discusses the internal error "ORA-600 [2662]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [2662] [a] [b] [c] [d] [e] VERSIONS: versions 6.0 to 10.1 DESCRIPTION: A data block SCN is ahead of the current SCN. The ORA-600 [2662] occurs when an SCN is compared to the dependent SCN stored in a UGA variable. If the SCN is less than the dependent SCN then we signal the ORA-600 [2662] internal error. ARGUMENTS: Arg [a] Current SCN WRAP Arg [b] Current SCN BASE Arg [c] dependent SCN WRAP Arg [d] dependent SCN BASE Arg [e] Where present this is the DBA where the dependent SCN came from. FUNCTIONALITY: File and IO buffer management for redo logs IMPACT: INSTANCE FAILURE POSSIBLE PHYSICAL CORRUPTION SUGGESTIONS: There are different situations where ORA-600 [2662] can be raised. It can be raised on startup or duing database operation. If not using Parallel Server, check that 2 instances have not mounted the same database. Check for SMON traces and have the alert.log and trace files ready to send to support. Check the SCN difference [argument d]-[argument b]. If the SCNs in the error are very close, then try to shutdown and startup the instance several times. In some situations, the SCN increment during startup may permit the database to open. Keep track of the number of times you attempted a startup. If the Known Issues section below does not help in terms of identifying a solution, please submit the trace files and alert.log to Oracle Support Services for further analysis. Known Issues: Bug# 2899477 See Note 2899477.8 Minimise risk of a false OERI[2662] Fixed: 9.2.0.5, 10.1.0.2 Bug# 2764106 See Note 2764106.8 False OERI[2662] possible on SELECT which can crash the instance Fixed: 9.2.0.5, 10.1.0.2 Bug# 2054025 See Note 2054025.8 OERI:2662 possible on new TEMPORARY index block Fixed: 9.0.1.3, 9.2.0.1 Bug# 851959 See Note 851959.8 OERI:2662 possible from distributed OPS select Fixed: 7.3.4.5 Bug# 647927 P See Note 647927.8 Digital Unix ONLY: OERI:2662 could occur under heavy load Fixed: 8.0.4.2, 8.0.5.0 INTERNAL ONLY SECTION - NOT FOR PUBLICATION OR DISTRIBUTION TO CUSTOMERS ======================================================================== There were 2 forms of this error until 7.2.3: Type I: 4/5 argument forms - The SCN found on a block (dependent SCN) is ahead of the current SCN. See below for this Type II: 1 Argument (before 7.2.3 only): Oracle is in the process of writing a block to a log file. If the calculated block checksum is less than or equal to 1 (0 and 1 are reserved) ORA-600 [2662] is returned. This is a problem generating an offline immediate log marker (kcrfwg). *NOT DOCUMENTED HERE* Type I ~~~~~~ a. Current SCN WRAP b. Current SCN BASE c. dependent SCN WRAP d. dependent SCN BASE e. Where present this is the DBA where the dependent SCN came from. From kcrf.h: If the SCN comes from the recent or current SCN then a dba of zero is saved. If it comes from undo$ because the undo segment is not available then the undo segment number is saved, which looks like a block from file 0. If the SCN is for a media recovery redo (i.e. block number == 0 in change vector), then the dba is for block 0 of the relevant datafile. If it is from another database for a distributed transaction then dba is DBAINF(). If it comes from a TX lock then the dba is really usn<<16+slot. Type II ~~~~~~~ a. checksum -> log block checksum - zero if none (thread # in old format) --------------------------------------------------------------------------- Diagnosis: ~~~~~~~~~~ In addition to different basic types from above, there are different situations where ORA-600 [2662] type I can be raised. Getting started: ~~~~~~~~~~~~~~~~ (1) is the error raised during normal database operations (i.e. when the database is up) or during startup of the database? (2) what is the SCN difference [d]-[b] ( subtract argument 'b' from arg 'd')? (3) is there a fifth argument [e] ? If so convert the dba to file# block# Is it a data dictionary object? (file#=1) If so find out object name with the help of reference dictionary from second database (4) What is the current SQL statement? (see trace) Which table is refered to? Does the table match the object you found in previous step? Be careful at this point: there may be no relationship between DBA in [e] and the real source of problem (blockdump). Deeper analysis: ~~~~~~~~~~~~~~~~ (1) investigate trace file: this will be a user trace file normally but could be an smon trace too (2) search for: 'buffer' ("buffer dba" in Oracle7 dumps, "buffer tsn" in Oracle8/Oracle9 dumps) this will bring you to a blockdump which usually represents the 'real' source of OERI:2662 WARNING: There may be more than one buffer pinned to the process so ensure you check out all pinned buffers. -> does the blockdump match the dba from e.? -> what kind of blockdump is it? (a) rollback segment header (b) datablock (c) other Check list and possible causes ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ If Parallel Server check both nodes are using the same lock manager instance & point at the same control files. Possible causes: (1) doing an open resetlogs with _ALLOW_RESETLOGS_CORRUPTION enabled (2) a hardware problem, like a faulty controller, resulting in a failed write to the control file or the redo logs (3) restoring parts of the database from backup and not doing the appropriate recovery (4) restoring a control file and not doing a RECOVER DATABASE USING BACKUP CONTROLFILE (5) having _DISABLE_LOGGING set during crash recovery (6) problems with the DLM in a parallel server environment (7) a bug Solutions: (1) if the SCNs in the error are very close, attempting a startup several times will bump up the dscn every time we open the database even if open fails. The database will open when dscn=scn. (2)You can bump the SCN either on open or while the database is open using (see Note 30681.1 ). Be aware that you should rebuild the database if you use this option. Once this has occurred you would normally want to rebuild the database via exp/rebuild/imp as there is no guarantee that some other blocks are not ahead of time. Articles: ~~~~~~~~~ Solutions: Note 30681.1 Details of the ADJUST_SCN Event Note 1070079.6 Alter System Checkpoint Possible Causes: Note 1021243.6 CHECK INIT.ORA SETTING _DISABLE_LOGGING Note 41399.1 Forcing the database open with `_ALLOW_RESETLOGS_CORRUPTION` Note 851959.9 OERI:2662 DURING CREATE SNAPSHOT AT MASTER SITE Known Bugs: ~~~~~~~~~~~ Fixed In. Bug No. Description ---------+------------+---------------------------------------------------- 7.1.5 Bug 229873 7.1.3 Bug 195115 Miscalculation of SCN on startup for distributed TX ? 7.1.6.2.7 Bug 297197 Port specific Solaris OPS problem 7.3 Bug 336196 Port specific IBM SP AIX problem -> dlm issue 7.3.4.5 Bug 851959 OERI:2662 possible from distributed OPS select Not fixed Bug 2216823 OERI:2662 reported when reusing tempfile with restored DB 8.1.7.4 Bug 2177050 OERI:729 space leak possible (with tags "define var info"/"oactoid info") can corrupt UGA and cause OERI:2662 --------------------------------------------------------------------------- Ensure that this note comes out on top in Metalink when searched ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 2662 in file header does not match given name You are certain that the file is good and that it belongs to that database. Solution: Check the file's properties in Windows Explorer and verify that it is not a "Hidden" file. Explanation: If you have set the "Show All Files' option under Explorer, View, Options, you are able to see 'hidden' files that other users and/or applications cannot. If any or all datafiles are marked as 'hidden' files, Oracle does not see them when it tries to recreate the controlfile. You must change the properties of the file by right-clicking on the file in Windows Explorer and then deselecting the check box marked "Hidden" under the General tab. You should then be able to create the controlfile. References: Note 1084048.6 ORA-01503, ORA-01161: on Create Controlfile. Note 2: ======= This message may result, if the db_name in the init.ora does not match with the set "db_name" given while creating the controlfile. Also, remove any old controlfiles present in the specified directory. Thanks, Note 3: ======= We ran into a similar problem when trying to create a new instance with datafiles from another database. The error comes in the create control file statement. Oracle uses REUSE as the default option when you do the alter database backup controlfile to trace. If you delete REUSE then the new database name you will change all the header information in all the database datafiles and you will be able to start up the instance. Hope this helps. Note 4: ======= Try this command "CREATE CONTROLFILE SET DATABASE..." instead of "CREATE CONTROLFILE REUSE DATABASE..." I think it would be better. 19.50. ORA-01031 ================ Note 1: ------- The 'OSDBA' and 'OSOPER' groups are chosen at installation time and usually both default to the group 'dba'. These groups are compiled into the 'oracle' executable and so are the same for all databases running from a given ORACLE_HOME directory. The actual groups being used for OSDBA and OSOPER can be checked thus: cd $ORACLE_HOME/rdbms/lib cat config.[cs] The line '#define SS_DBA_GRP "group"' should name the chosen OSDBA group. The line '#define SS_OPER_GRP "group"' should name the chosen OSOPER group. Note 2: ------- Bookmark Fixed font Go to End Doc ID: Note:69642.1 Content Type: TEXT/PLAIN Subject: UNIX: Checklist for Resolving Connect AS SYSDBA Issues Creation Date: 20-APR-1999 Type: TROUBLESHOOTING Last Revision Date: 31-DEC-2004 Status: PUBLISHED Introduction: ~~~~~~~~~~~~~ This bulletin lists the documented causes of getting ---> prompted for a password when trying to CONNECT as SYSDBA ---> errors such as ORA-01031, ORA-01034, ORA-06401, ORA-03113,ORA-09925, ORA-09817, ORA-12705, ORA-12547 a) SQLNET.ORA Checks: --------------------- 1. The "sqlnet.ora" can be found in the following locations (listed by search order): $TNS_ADMIN/sqlnet.ora $HOME/sqlnet.ora $ORACLE_HOME/network/admin/sqlnet.ora Depending upon your operating system, it may also be located in: /var/opt/oracle/sqlnet.ora /etc/sqlnet.ora A corrupted "sqlnet.ora" file, or one with security options set, will cause a 'connect internal' request to prompt for a password. To determine if this is the problem, locate the "sqlnet.ora" that is being used. The one being used will be the first one found according to the search order listed above. Next, move the file so that it will not be found by this search: % mv sqlnet.ora sqlnet.ora_save Try to connect internal again. If it still fails, search for other "sqlnet.ora" files according to the search order listed above and repeat using the move command until you are sure there are no other "sqlnet.ora" files being used. If this does not resolve the issue, use the move command to put all the "sqlnet.ora" files back where they were before you made the change: % mv sqlnet.ora_save sqlnet.ora If moving the "sqlnet.ora" resolves the issue, then verify the contents of the file: a) SQLNET.AUTHENTICATION_SERVICES If you are not using database links, comment this line out or try setting it to: SQLNET.AUTHENTICATION_SERVICES = (BEQ,NONE) b) SQLNET.CRYPTO_SEED This should not be set in a "sqlnet.ora" file on UNIX. If it is, comment the line out. (This setting is added to the "sqlnet.ora" if it is built by one of Oracle's network cofiguration products shipped with client products) c) AUTOMATIC_IPC If this is set to "ON" it can force a "TWO_TASK" connection. Try setting this to "OFF": AUTOMATIC_IPC = OFF 2. Set the permissions correctly in the "TNS_ADMIN" files. The environment variable TNS_ADMIN defines the directory where the "sqlnet.ora", "tnsnames.ora", and "listener.ora" files reside. These files must contain the correct permissions, which are set when "root.sh" runs during installation. As root, run "root.sh" or edit the permissions on the "sqlnet.ora", "tnsnames.ora", and "listener.ora" files by hand as follows: $ cd $TNS_ADMIN $ chmod 644 sqlnet.ora tnsnames.ora listener.ora $ ls -l sqlnet.ora tnsnames.ora listener.ora -rw-r--r-- 1 oracle dba 1628 Jul 12 15:25 listener.ora -rw-r--r-- 1 oracle dba 586 Jun 1 12:07 sqlnet.ora -rw-r--r-- 1 oracle dba 82274 Jul 12 15:23 tnsnames.ora b) Software and Operating System Issues: ---------------------------------------- 1. Be sure $ORACLE_HOME is set to the correct directory and does not have any typing mistakes: % cd $ORACLE_HOME % pwd If this returns a location other than your "ORACLE_HOME" or is invalid, you will need to reset the value of this environment variable: sh or ksh: ---------- $ ORACLE_HOME= $ export ORACLE_HOME Example: $ ORACLE_HOME=/u01/app/oracle/product/7.3.3 $ export ORACLE_HOME csh: ---- % setenv ORACLE_HOME Example: % setenv ORACLE_HOME /u01/app/oracle/product/7.3.3 If your "ORACLE_HOME" contains a link or the instance was started with the "ORACLE_HOME" set to another value, the instance may try to start using the memory location that another instance is using. An example of this might be: You have "ORACLE_HOME" set to "/u01/app/oracle/product/7.3.3" and start the instance. Then you do something like: % ln -s /u01/app/oracle/product/7.3.3 /u01/app/oracle/7.3.3 % setenv ORACLE_HOME /u01/app/oracle/7.3.3 % svrmgrl SVRMGR> connect internal If this prompts for a password then most likely the combination of your "ORACLE_HOME" and "ORACLE_SID" hash to the same shared memory address of another running instance. Otherwise you may be able to connect internal but you will receive an ORA-01034 "Oracle not available" error. In most cases using a link as part of your "ORACLE_HOME" is fine as long as you are consistent. Oracle recommends that links not be used as part of the "ORACLE_HOME", but their use is supported. 2. Check that $ORACLE_SID is set to the correct SID, (including capitalization), and does not have any typos: % echo $ORACLE_SID Refer to Note:1048876.6 for more information. 3. Ensure $TWO_TASK is not set. To check if "TWO_TASK" is set, do the following: sh, ksh or on HP/UX only csh: ----------------------------- env |grep -i two - or - echo $TWO_TASK csh: ---- setenv |grep -i two If any lines are returned such as: TWO_TASK= - or - TWO_TASK=PROD You will need to unset the environment variable "TWO_TASK": sh or ksh: ---------- unset TWO_TASK csh: ---- unsetenv TWO_TASK Example : $ TWO_TASK=V817 $ export TWO_TASK $ sqlplus /nolog SQL*Plus: Release 8.1.7.0.0 - Production on Fri Dec 31 10:12:25 2004 (c) Copyright 2000 Oracle Corporation. All rights reserved. SQL> conn / as sysdba ERROR: ORA-01031: insufficient privileges $ unset TWO_TASK $ sqlplus /nolog SQL> conn / as sysdba Connected. If you are running Oracle release 8.0.4, and upon starting "svrmgrl" you receive an ORA-06401 "NETCMN: invalid driver designator" error, you should also unset two_task. The login connect string may be getting its value from the TWO_TASK environment variable if this is set for the user. 4. Check the permissions on the Oracle executable: % cd $ORACLE_HOME/bin % ls -l oracle ('ls -n oracle' should work as well) The permissions should be rwsr-s--x, or 6751. If the permissions are incorrect, do the following as the "oracle" software owner: % chmod 6751 oracle If you receive an ORA-03113 "end-of-file on communication" error followed by a prompt for a password, then you may also need to check the ownership and permissions on the dump directories. These directories must belong to Oracle, group dba, (or the appropriates names for your installation). This error may occur while creating a database. Permissions should be: 755 (drwxr-xr-x) Also, the alert.log must not be greater than 2 Gigabytes in size. When you start up "nomount" an Oracle pseudo process will try to write the "alert.log" file in "udump". When Oracle cannot do this (either because of permissions or because of the "alert.log" being greater than 2 Gigabytes in size), it will issue the ORA-03113 error. 5. "osdba" group checks: a. Make sure the operating system user issuing the CONNECT INTERNAL belongs to the "osdba" group as defined in the "$ORACLE_HOME/rdbms/lib/config.s" or "$ORACLE_HOME/rdbms/lib/config.c". Typically this is set to "dba". To verify the operating system groups the user belongs to, do the following: % id uid=1030(oracle) gid=1030(dba) The "gid" here is "dba" so the "config.s" or "config.c" may contain an entry such as: /* 0x0008 15 */ .ascii "dba\0" If these do not match, you either need to add the operating system user to the group as it is seen in the "config" file, or modify the "config" file and relink the "oracle" binary. Refer to entry [NOTE:50507.1] section 3 for more details. b. Be sure you are not logged in as the "root" user and that the environment variables "USER", "USERNAME", and "LOGNAME" are not set to "root". The "root" user is a special case and cannot connect to Oracle as the "internal" user unless the effective group is changed to the "osdba" group, which is typically "dba". To do this, either modify the "/etc/password" file (not recommended) or use the "newgrp" command: # newgrp dba "newgrp" always opens a new shell, so you cannot issue "newgrp" from within a shell script. Keep this in mind if you plan on executing scripts as the "root" user. c. Verify that the "osdba" group is only listed once in the "/etc/group" file: % grep dba /etc/group dba::1010: dba::1100: If more than one line starting with the "osdba" group is returned, you need to remove the ones that are not correct. It is not possible to have more than one group use a group name. d. Check that the oracle user uid and gid are matching with /etc/passwd and /etc/group : $ id uid=500(oracle) gid=235(dba) $ grep oracle /etc/passwd oracle:x:500:235:oracle:/home/oracle:/bin/bash ^^^ $ grep dba /etc/group dba:x:253:oracle ^^^ The mismatch also causes an ORA-1031 error. 6. Verify that the file system is not mounted no set uid: % mount /u07 on /dev/md/dsk/d7 nosuid/read/write If the filesytem is mounted "nosuid", as seen in this example, you will need to unmount the filesystem and mount it without the "nosuid" option. Consult your operating system documentation or your operating system vendor for instruction on modifying mount options. 7. Please read the following warning before you attempt to use the information in this step: ****************************************************************** * * * WARNING: If you remove segments that belong to a running * * instance you will crash the instance, and this may * * cause database corruption. * * * * Please call Oracle Support Services for assistance * * if you have any doubts about removing shared memory * * segments. * * * ****************************************************************** If an instance crashed or was killed off using "kill" there may be shared memory segments hanging around that belong to the down instance. If there are no other instances running on the machine you can issue: % ipcs -b T ID KEY MODE OWNER GROUP SEGSZ Shared Memory: m 0 0x50000ffe --rw-r--r-- root root 68 m 1601 0x0eedcdb8 --rw-r----- oracle dba 4530176 In this case the "ID" of "1601" is owned by "oracle" and if there are no other instances running in most cases this can safely be removed: % ipcrm -m 1601 If your SGA is split into multiple segments you will have to remove all segments associated with the instance. If there are other instances running, and you are not sure which memory segments belong to the failed instance, you can do the following: a. Shut down all the instances on the machine and remove whatever shared memory still exists that is owned by the software owner. b. Reboot the machine. c. If your Oracle software is release 7.3.3 or newer, you can connect into each instance that is up and identify the shared memory owned by that instance: % svrmgrl SVRMGR> connect internal SVRMGR> oradebug ipc In Oracle8: ----------- Area #0 `Fixed Size', containing Subareas 0-0 Total size 000000000000b8c0, Minimum Subarea size 00000000 Subarea Shmid Size Stable Addr 0 7205 000000000000c000 80000000 In Oracle7: ----------- -------------- Shared memory -------------- Seg Id Address Size 2016 80000000 4308992 Total: # of segments = 1, size = 4308992 Note the "Shmid" for Oracle8 and "Seg Id" for Oracle7 for each running instance. By process of elimination find the segments that do not belong to an instance and remove them. 8. If you are prompted for a password and then receive error ORA-09925 "unable to create audit trail file" or error ORA-09817 "write to audit file failed", along with "SVR4 Error: 28: No space left on device", do the following: Check your "pfile". It is typically in the "$ORACLE_HOME/dbs" directory and will be named "init.ora, where "" is the value of "ORACLE_SID" in your environment. If the "init.ora" file has the "ifile" parameter set, you will also have to check the included file as well. You are looking for the parameter "audit_file_dest". If "audit_file_dest" is set, change to that directory; otherwise change to the "$ORACLE_HOME/rdbms/audit" directory, as this is the default location for audit files. If the directory does not exist, create it. Ensure that you have enough space to create the audit file. The audit file is generally 600 bytes in size. If it does exist, verify you can write to the directory: % touch afile If it could not create the called "afile", you need to change the permissions on your audit directory: % chmod 751 9. If connect internal prompts you for a password and then you receive an ORA-12705 "invalid or unknown NLS parameter value specified" error, you need to verify the settings for "ORA_NLS", "ORA_NLS32", "ORA_NLS33" or "NLS_LANG". You will need to consult your Installation and Configuration Guide for the proper settings for these environment variables. 10. If you have installed Oracle software and are trying to connect with Server Manager to create or start the database, and receive a TNS-12571 "packet writer failure" error, please refer to Note:1064635.6 11. If in SVRMGRL (Server Manager line mode), you are running the "startup.sql" script and receive the following error: ld:so.1: oracle_home/bin/svrmgrl fatal relocation error symbol not found kgffiop RDBMS v7.3.2 is installed. RDBMS v8.0.4 is a separate "oracle_home", and you are attempting to have it coexist. This is due to the wrong version of the client shared library "libclntsh.so.1" being used at runtime. Verify environment variable settings. You need to ensure that "ORACLE_HOME" and "LD_LIBRARY_PATH" are set correctly. For C-shell, type: % setenv LD_LIBRARY_PATH $ORACLE_HOME/lib % setenv ORACLE_HOME /u01/app/oracle/product/8.0.4 For Bourne or Korn shell, type: $ LD_LIBRARY_PATH=$ORACLE_HOME/lib $ export LD_LIBRARY_PATH $ ORACLE_HOME=/u01/app/oracle/product/8.0.4 $ export ORACLE_HOME 12. Ensure that the disk the instance resides on has not reached 100% capacity. % df -k If it has reached 100% capacity, this may be the cause of 'connect internal' prompting for a password. Additional disk space will need to be made available before 'connect internal' will work. For additional information refer to Note:97849.1 13. Delete process.dat and regid.dat files in $ORACLE_HOME/otrace/admin directory. Oracle Trace is enabled by default on 7.3.2 and 7.3.3 (depends on platform) This can caused high disk space usage by these files and cause a number of apparently mysterious side effects. See Note:45482.1 for more details. 14. When you get ora-1031 "Insufficient privileges" on connect internal after you supply a valid password and you have multiple instances running from the same ORACLE_HOME, be sure that if an instance has REMOTE_LOGIN_PASSWORDFILE set to exclusive that the file $ORACLE_HOME/dbs/orapw does exist, otherwise it defaults to the use of the file orapw that consequently causes access problems for any other database that has the parameter set to shared. Set the parameter REMOTE_LOGIN_PASSWORDFILE to shared for all instances that share the common password file and create an exclusive orapw password files for any instances that have this set to exclusive. 15. Check permissions on /etc/passwd file (Unix only). If Oracle cannot open the password file, connect internal fails with ORA-1031, since Oracle is not able to verify if the user trying to connect is indeed in the dba group. Example: -------- # chmod 711 /etc/passwd # ls -ltr passwd -rwx--x--x 1 root sys 901 Sep 21 14:26 passwd $ sqlplus '/ as sysdba' SQL*Plus: Release 9.2.0.1.0 - Production on Sat Sep 21 16:21:18 2002 Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved. ERROR: ORA-01031: insufficient privileges Trussing sqlplus will show also the problem: 25338: munmap(0xFF210000, 8192) = 0 25338: lwp_mutex_wakeup(0xFF3E0778) = 0 25338: lwp_mutex_lock(0xFF3E0778) = 0 25338: time() = 1032582594 25338: open("/etc/passwd", O_RDONLY) Err#13 EACCES 25338: getrlimit(RLIMIT_NOFILE, 0xFFBE8B28) = 0 c) Operating System Specific checks: ------------------------------------ 1. On OpenVMS, check that the privileges have been granted at the Operating System level: $ SET DEFAULT SYS$SYSTEM: $ RUN AUTHORIZE If the list returned by AUTHORIZE does not contain ORA__DBA, or ORA_DBA, then you do not have the correct OS privileges to issue a connect internal. If ORA__DBA was added AFTER ORA_DBA, then ORA_DBA needs to be removed and granted again to be updated. Please refer to Note:1010852.6 for more details. 2. On Windows NT, check if DBA_AUTHORIZATION is set to BYPASS in the registry. 3. On Windows NT, if you are able to connect internally but then startup fails for some reason, successive connect internal attempts might prompt for a password. You may also receive errors such as: ORA-12705: invalid or unknown NLS parameter value specified ORA-01012: not logged on LCC-00161: Oracle error (possible syntax error) ORA-01031: insufficient privileges Refer to entry Note:1027964.6 for suggestions on how to resolve this problem 4. If you are using Multi-Threaded Server (MTS), make sure you are using a dedicated server connection. A dedicated server connection is required to start up or shutdown the database. Unless the database alias in the "TNSNAMES.ORA" file includes a parameter to make a dedicated server connection, it will make a shared connection to a dispatcher. See Note:1058680.6 for more details. 5. On Solaris, if the file "/etc/.name_service_door" has incorrect permissions, Oracle cannot read the file. You will receive a message that "The Oracle user cannot access "/etc/.name_service_door" (permission denied). This file is a flavor of IPC specific to Solaris which Oracle software is using This can also cause connect internal problems. See entry Note:1066589.6 6. You are on Digital Unix, running SVRMGRL (Server Manager line mode), and you receive an ORA-12547 "TNS:lost contact" error and a password prompt. This problem occurs when using Parallel Server and the True Cluster software together. If Parallel Server is not linked in, svrmgrl works as expected. Oracle V8.0.5 requires an Operating System patch which previous versions of Oracle did not require. The above patch allows svrmgrl to communicate with the TCR software. You can determine if the patch is applied by running: % nm /usr/ccs/lib/libssn.a | grep adjust If this returns nothing, then you need to: 1. Obtain the patch for TCR 1.5 from Digital. This patch is for the MC SCN and adds the symbol "adjustSequenceNumber" to the library /usr/ccs/lib/libssn.a. 2. Apply the patch. 3. Relink Oracle Another possibility is that you need to raise the value of kernel parameter per-proc-stack-size when increased from its default value of 2097152 to 83886080 resolved this problem. 7. You are on version 6.2 of the Silicon Graphics UNIX (IRIX) operating system and you have recently installed RDBMS release 8.0.3. If you are logged on as "oracle/dba" and an attempt to log in to Server Manager using "connect/internal" prompts you for a password, you should refer to entry Note:1040607.6 8. On AIX 4.3.3 after applying ML5 or higher you can not longer connect as internal or if on 9.X '/ as sysdba' does not work as well. This is a known AIX bug and it occurs on all RS6000 ports including SP2. There is two workarounds and one solution. They are as follows: 1) Use mkpasswd command to remove the index. This is valid until a new user is added to "/etc/passwd" or modified: # mkpasswd -v -d 2) Touch the "/etc/passwd" file. If the "/etc/passwd" file is newer than the index it will not use the password file index: # touch /etc/passwd 3) Obtain APAR IY22458 from IBM. Any questions about this APAR should be directed to IBM. d) Additional Information: -------------------------- 1. In the "Oracle7 Administrator's Reference for UNIX", there is a note that states: If REMOTE_OS_AUTHENT is set to true, users who are members of the dba group on the remote machine are able to connect as INTERNAL without a password. However, if you are connecting remotely, that is connecting via anything except the bequeath adapter, you will be prompted for a password regardless of the value of "REMOTE_OS_AUTHENT". Refer to bug 644988 References: ~~~~~~~~~~~ [NOTE:1048876.6] UNIX: Connect internal prompts for password after install [NOTE:1064635.6] ORA-12571: PACKET WRITER FAILURE WHEN STARTING SVRMGR [NOTE:1010852.6] OPENVMS: ORA-01031: WHEN ISSUING "CONNECT INTERNAL" IN SQL*DBA OR SERVER MANAGER [NOTE:1027964.6] LCC-00161 AND ORA-01031 ON STARTUP [NOTE:1058680.6] ORA-00106 or ORA-01031 ERROR when trying to STARTUP or SHUTDOWN DATABASE [NOTE:1066589.6] UNIX: Connect Internal asks for password when TWO_TASK is set [NOTE:1040607.6] SGI: ORA-01012 ORA-01031: WHEN USING SRVMGR AFTER 8.0.3 INSTALL [NOTE:97849.1] Connect internal Requires Password [NOTE:50507.1] SYSDBA and SYSOPER Privileges in Oracle8 and Oracle7 [NOTE:18089.1] UNIX: Connect INTERNAL / AS SYSBDA Privilege on Oracle 7/8 [BUG:644988] REMOTE_OS_AUTHENT=TRUE: NOT ALLOWING USERS TO CONNECT INTERNAL WITHOUT PASSWORD Search Words: ~~~~~~~~~~~~~ svrmgrm sqldba sqlplus sqlnet remote_login_passwordfile Note 3: ------- ORA-01031: insufficient privileges Cause: An attempt was made to change the current username or password without the appropriate privilege. This error also occurs if attempting to install a database without the necessary operating system privileges. Action: Ask the database administrator to perform the operation or grant the required privileges. Note 4: ------- ORA-01031: insufficient privileges In most cases, the user receiving this error lacks a privilege to create an object (such as a table, view, procedure and the like). Grant the required privilege like so: grant create table to user_lacking_privilege; Startup If someone receives this error while trying to startup the instance, the logged on user must belong to the ora_dba group on Windows or dba group on Unix. Note 5: ------- I am not sure it is the same, but I got this error today in windows when sql_authentication in sqlnet.ora was NONE. Changing it to NTS solved the problem. 19.51 ORA-00600: internal error code, arguments: [17059]: ========================================================= Note 1: ------- Doc ID : Note:138554.1 Content Type: TEXT/PLAIN Subject: ORA-600 [17059] Creation Date: 02-APR-2001 Type: REFERENCE Last Revision Date: 09-DEC-2004 Status: PUBLISHED Note: For additional ORA-600 related information please read [NOTE:146580.1] PURPOSE: This article discusses the internal error "ORA-600 [17059]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [17059] [a] VERSIONS: versions 7.1 to 10.1 DESCRIPTION: While building a table to hold the list of child cursor dependencies relating to a given parent cursor, we exceed the maximum possible size of the table. ARGUMENTS: Arg [a] Object containing the table FUNCTIONALITY: Kernel Generic Library cache manager IMPACT: PROCESS FAILURE NON CORRUPTIVE - No underlying data corruption. SUGGESTIONS: One symptom of this error is that the session will appear to hang for a period of time prior to this error being reported. If the Known Issues section below does not help in terms of identifying a solution, please submit the trace files and alert.log to Oracle Support Services for further analysis. Issuing this SQL as SYS (SYSDBA) may help show any problem objects in the dictionary: select do.obj#, po.obj# , p_timestamp, po.stime , decode(sign(po.stime-p_timestamp),0,'SAME','*DIFFER*') X from sys.obj$ do, sys.dependency$ d, sys.obj$ po where P_OBJ#=po.obj#(+) and D_OBJ#=do.obj# and do.status=1 /*dependent is valid*/ and po.status=1 /*parent is valid*/ and po.stime!=p_timestamp /*parent timestamp not match*/ order by 2,1 ; Normally the above select would return no rows. If any rows are returned the listed dependent objects may need recompiling. Known Issues: Bug# 3555003 See [NOTE:3555003.8] View compilation hangs / OERI:17059 after DBMS_APPLY_ADM.SET_DML_HANDLER Fixed: 9.2.0.6 Bug# 2707304 See [NOTE:2707304.8] OERI:17059 / OERI:kqlupd2 / PLS-907 after adding partitions to Partitioned IOT Fixed: 9.2.0.3, 10.1.0.2 Bug# 2636685 See [NOTE:2636685.8] Hang / OERI:[17059] after adding a list value to a partition Fixed: 9.2.0.3, 10.1.0.2 Bug# 2626347 See [NOTE:2626347.8] OERI:17059 accessing view after ADD / SPLIT PARTITION Fixed: 9.2.0.3, 10.1.0.2 Bug# 2306331 See [NOTE:2306331.8] Hang / OERI[17059] on view after SET_KEY or SET_DML_INVOKATION on base table Fixed: 9.2.0.2 Bug# 1115424 See [NOTE:1115424.8] Cursor authorization and dependency lists too long - can impact shared pool / OERI:17059 Fixed: 8.0.6.2, 8.1.6.2, 8.1.7.0 Bug# 631335 See [NOTE:631335.8] OERI:17059 from extensive re-user of a cursor Fixed: 8.0.4.2, 8.0.5.0, 8.1.5.0 Bug# 558160 See [NOTE:558160.8] OERI:17059 from granting privileges multiple times Fixed: 8.0.3.2, 8.0.4.0, 8.1.5.0 Note 2: ------- Doc ID : Note:234457.1 Content Type: TEXT/X-HTML Subject: ORA-600 [17059] Error When Compiling A Package Creation Date: 19-FEB-2003 Type: PROBLEM Last Revision Date: 24-AUG-2004 Status: PUBLISHED fact: fact: Oracle Server - Enterprise Edition fact: Partitioned Tables / Indexes symptom: ORA-600 [17059] Error When Compiling A Package symptom: When Compiling a Package symptom: The Package Accesses a Partitioned Table symptom: ORA-00600: internal error code, arguments: [%s], [%s], [%s], [%s], [%s], [%s], [%s] symptom: internal error code, arguments: [17059], [352251864] symptom: Calling Location kglgob symptom: Calling Location kgldpo symptom: Calling Location kgldon symptom: Calling Location pkldon symptom: Calling Location pkloud symptom: Calling Location - phnnrl_name_resolve_by_loading cause: This is due to > fixed in 10i, and occurs when accessing a partitioned table via a dblink within the package, where DDL (such as adding/dropping partitions) is performed on the table. fix: This is fixed in 9.0.1.4, 9.2.0.2 & 10i. One-off patches are available for 8.1.7.4. A workaround is to flush the shared pool. Note 3: ------- Doc ID : Note:239796.1 Content Type: TEXT/PLAIN Subject: ORA-600 [17059] when querying dba_tablespaces, dba_indexes, dba_ind_partitions etc Creation Date: 28-MAY-2003 Type: PROBLEM Last Revision Date: 13-AUG-2004 Status: PUBLISHED Problem: ~~~~~~~~ The information in this article applies to: Internal Error ORA-600 [17059] when querying Data dictionary views like dba_tablespaces, dba_indexes, dba_ind_partitions etc Symptom(s) ~~~~~~~~~~ While querying Data dictionary views like dba_tablespaces, dba_indexes, dba_ind_partitions etc, getting internal error ORA-600 [17059] Change(s) ~~~~~~~~~~ You probably altered some objects or executed some cat*.sql scripts. Cause ~~~~~~~ Some SYS objects are INVALID. Fix ~~~~ Connect SYS run $ORACLE_HOME/rdbms/admin/utlrp.sql and make sure all the objects are valid. 19.52: ORA-00600: internal error code, arguments: [17003] ========================================================= Note 1: ------- The information in this article applies to: Oracle Forms - Version: 9.0.2.7 to 9.0.2.12 Oracle Server - Enterprise Edition - Version: 9.2 This problem can occur on any platform. Errors ORA 600 "internal error code, arguments: [%s],[%s],[%s], [%s], [%s], Symptoms The following error occurs when compiling a form or library ( fmb / pll ) against RDBMS 9.2 PL/SQL ERROR 0 at line 0, column 0 ORA-00600: internal error code, arguments: [17003], [0x11360BC], [275], [1], [], [], [], [] The error reproduces everytime. Triggers / local program units in the form / library contain calls to stored database procedures and / or functions. The error does not occur when compiling against RDBMS 9.0.1 or lower. Cause This is a known bug / issue. The compilation error occurs when the form contains a call to a stored database function / procedure which has two DATE IN variables receiving DEFAULT values such as SYSDATE. Reference: Abstract: INTERNAL ERROR [1401] WHEN COMPILE FUNCTION WITH 2 DEFAULT DATE VARIABLES ON 9.2 Fix The bug is fixed in Oracle Forms 10g (9.0.4). There is no backport fix available for Forms 9i (9.0.2) To work-around, modify the offending calls to the stored database procedure/ functions so that DEFAULT parameter values are not passed directly . For example, pass the DEFAULT value SYSDATE indirectly to the stored database procedure/ function by first assigning it to a local variable in the form. Note 2: ------- Doc ID : Note:138537.1 Content Type: TEXT/PLAIN Subject: ORA-600 [17003] Creation Date: 02-APR-2001 Type: REFERENCE Last Revision Date: 15-OCT-2004 Status: PUBLISHED Note: For additional ORA-600 related information please read [NOTE:146580.1] PURPOSE: This article discusses the internal error "ORA-600 [17003]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [17003] [a] [b] [c] VERSIONS: versions 7.0 to 10.1 DESCRIPTION: The error indicates that we have tried to lock a library cache object by using the dependency number to identify the target object and have found that no such dependency exists. Under this situation we will raise an ORA-600 [17003] if the dependency number that we are using exceeds the number of entries in the dependency table or the dependency entry is not marked as invalidated. ARGUMENTS: Arg [a] Library Cache Object Handle Arg [b] Dependency number Arg [c] 1 or 2 (indicates where the error was raised internally) FUNCTIONALITY: Kernel Generic Library cache manager IMPACT: PROCESS MEMORY FAILURE NO UNDERLYING DATA CORRUPTION. SUGGESTIONS: A common condition where this error is seen is problematic upgrades. If a patchset has recently been applied, please confirm that there were no errors associated with this upgrade. Specifically, there are some XDB related bugs which can lead to this error being reported. Known Issues: Bug# 2611590 See [NOTE:2611590.8] OERI:[17003] running XDBRELOD.SQL Fixed: 9.2.0.3, 10.1.0.2 Bug# 3073414 XDB may not work after applying a 9.2 patch set Fixed: 9.2.0.5 19.53: ORA-00600: internal error code, arguments: [qmxiUnpPacked2], [121], [], [], [], [], [], [] ================================================================================================= Note 1. ------- Doc ID: Note:222876.1 Content Type: TEXT/PLAIN Subject: ORA-600 [qmxiUnpPacked2] Creation Date: 09-DEC-2002 Type: REFERENCE Last Revision Date: 15-OCT-2004 Status: PUBLISHED Note: For additional ORA-600 related information please read [NOTE:146580.1] PURPOSE: This article discusses the internal error "ORA-600 [qmxiUnpPacked2]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [qmxiUnpPacked2] [a] VERSIONS: versions 9.2 to 10.1 DESCRIPTION: When unpickling an XOB or an array of XOBs an unexpected datatype was found. Generally due to XMLType data that has not been successfully upgraded from a previous version. ARGUMENTS: Arg [a] Type of XOB FUNCTIONALITY: Qernel xMl support Xob to/from Image IMPACT: PROCESS FAILURE NON CORRUPTIVE - No underlying data corruption. SUGGESTIONS: Please review the following article on Metalink : [NOTE:235423.1] How to resolve ORA-600 [qmxiUnpPacked2] during upgrade If you still encounter the error having tried the suggestions in the above article, or the article isn't applicible to your environment then ensure that the upgrade to current version was completed succesfully without error. If the Known Issues section below does not help in terms of identifying a solution, please submit the trace files and alert.log to Oracle Support Services for further analysis. Known Issues: Bug# 2607128 See [NOTE:2607128.8] OERI:[qmxiUnpPacked2] if CATPATCH.SQL/XDBPATCH.SQL fails Fixed: 9.2.0.3 Bug# 2734234 CONSOLIDATION BUG FOR ORA-600 [QMXIUNPPACKED2] DURING CATPATCH.SQL 9.2.0.2 Note 2. ------- Doc ID: Note:235423.1 Content Type: TEXT/X-HTML Subject: How to resolve ORA-600 [qmxiUnpPacked2] during upgrade Creation Date: 14-APR-2003 Type: HOWTO Last Revision Date: 18-MAR-2005 Status: PUBLISHED The information in this article applies to: Oracle 9.2.0.2 Multiple Platforms, 64-bit Symptom(s) ~~~~~~~~~~ ORA-600 [qmxiUnpPacked2] [] Cause ~~~~~ If the error is seen after applying 9.2.0.2 on a 9.2.0.1 database or if using DBCA in 9.2.0.2 to create a new database (which is using the 9.2.0.1 seed database) then it is very likely that either shared_pool_size or java_pool_size was too small when catpatch.sql was executed. Error is generally seen as ORA-600: internal error code, arguments: [qmxiUnpPacked2], [121] There are 3 options to proceed from here:- Fix ~~~~ Option 1 ======== If your shared_pool_size and java_pool_size are less than 150Mb the do the following :- 1/ Set your shared_pool_size and java_pool_size to 150Mb each. In some case you may need to use larger pool sizes. 2/ Get the xdbpatch.sql script from Note 237305.1 3/ Copy xdbpatch.sql to $ORACLE_HOME/rdbms/admin/xdbpatch.sql having taken a backup of the original file first 4/ Restart the instance with: startup migrate; 5/ spool catpatch @?/rdbms/admin/catpatch.sql Option 2 ======== If you already have shared_pool_size and java_pool_size set at greater than 150Mb then the problem may be caused by the shared memory allocated during the JVM upgrade is not released properly. In which case do the following :- 1/ Set your shared_pool_size and java_pool_size to 150Mb each. In some case you may need to use larger pool sizes. 2/ Get the xdbpatch.sql script from Note 237305.1 3/ Edit the xdbpatch.sql script and add the following as the first line in the script:- alter system flush shared_pool; 3/ Copy xdbpatch.sql to $ORACLE_HOME/rdbms/admin/xdbpatch.sql having taken a backup of the original file first 3/ Restart the instance with: startup migrate; 4/ spool catpatch @?/rdbms/admin/catpatch.sql Option 3 ======== If XDB is NOT in use and there are NO registered XML Schemas an alternative is to drop, and maybe re-install XDB :- 1/ To drop the XDB subsystem connect as sys and run @?/rdbms/admin/catnoqm.sql 2/ You can then run catpatch.sql to perform the upgrade startup migrate; @?/rdbms/admin/catpatch.sql 3/ Once complete you may chose to re-install the XDB subsystem, if so connect as sys and run catqm.sql @?/rdbms/admin/catqm.sql If the error is seen during normal database operation, ensure that upgrade to current version was completed succesfully without error. Once this is confirmed attempt to reproduce the error, if successful forward ALERT.LOG, trace files and full error stack to Oracle Support Services for further analysis. References ~~~~~~~~~~~ Bug 2734234 CONSOLIDATION BUG FOR ORA-600 [QMXIUNPPACKED2] DURING CATPATCH.SQL 9.2.0.2 Note 237305.1 Modified xdbpatch.sql 19.54 ORA-00600: internal error code, arguments: [kcbget_37], [1], [], [], [], [], [], [] ========================================================================================= ORA-00600: internal error code, arguments: [kcbso1_1], [], [], [], [], [], [], [] ORA-00600: internal error code, arguments: [kcbget_37], [1], [], [], [], [], [], [] Doc ID: Note:2652771.8 Subject: Support Description of Bug 2652771 Type: PATCH Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 13-AUG-2003 Last Revision Date: 14-AUG-2003 Click here for details of sections in this note. Bug 2652771 AIX: OERI[1100] / OERI[KCBGET_37] SGA corruption This note gives a brief overview of bug 2652771. Affects: Product (Component) Oracle Server (RDBMS) Range of versions believed to be affected Versions < 10G Versions confirmed as being affected 8.1.7.4 9.2.0.2 Platforms affected Aix 64bit 5L Aix 64bit 433 Fixed: This issue is fixed in 9.2.0.3 (Server Patch Set) Symptoms: Memory Corruption Internal Error may occur (ORA-600) ORA-600 [1100] / ORA-600 [kcbget_37] Known Issues: Bug# 2652771 P See [NOTE:2652771.8] AIX: OERI[1100] / OERI[KCBGET_37] SGA corruption Fixed: 9.2.0.3 19.55 ORA-00600: internal error code, arguments: [kcbzwb_4], [], [], [], [], [], [], [] ======================================================================================= Doc ID: Note:4036717.8 Subject: Bug 4036717 - Truncate table in exception handler can causes OERI:kcbzwb_4 Type: PATCH Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 25-FEB-2005 Last Revision Date: 09-MAR-2005 Click here for details of sections in this note. Bug 4036717 Truncate table in exception handler can causes OERI:kcbzwb_4 This note gives a brief overview of bug 4036717. Affects: Product (Component) PL/SQL (Plsql) Range of versions believed to be affected Versions < 10.2 Versions confirmed as being affected 10.1.0.3 Platforms affected Generic (all / most platforms affected) Fixed: This issue is fixed in 9.2.0.7 (Server Patch Set) 10.1.0.4 (Server Patch Set) 10g Release 2 (future version) Symptoms: Related To: Internal Error May Occur (ORA-600) ORA-600 [kcbzwb_4] PL/SQL Truncate Description Truncate table in exception handler can cause OERI:kcbzwb_4 with the fix for bug 3768052 installed. Workaround: Turn off or deinstall the fix for bug 3768052. Note that the procedure containing the affected transactional commands will have to be recompiled after backing out the bug fix. Doc ID: Note:4036717.8 Subject: Bug 4036717 - Truncate table in exception handler can causes OERI:kcbzwb_4 Type: PATCH Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 25-FEB-2005 Last Revision Date: 09-MAR-2005 Click here for details of sections in this note. Bug 4036717 Truncate table in exception handler can causes OERI:kcbzwb_4 This note gives a brief overview of bug 4036717. Affects: Product (Component) PL/SQL (Plsql) Range of versions believed to be affected Versions < 10.2 Versions confirmed as being affected 10.1.0.3 Platforms affected Generic (all / most platforms affected) Fixed: This issue is fixed in 9.2.0.7 (Server Patch Set) 10.1.0.4 (Server Patch Set) 10g Release 2 (future version) Symptoms: Related To: Internal Error May Occur (ORA-600) ORA-600 [kcbzwb_4] PL/SQL Truncate Description Truncate table in exception handler can cause OERI:kcbzwb_4 with the fix for bug 3768052 installed. Workaround: Turn off or deinstall the fix for bug 3768052. Note that the procedure containing the affected transactional commands will have to be recompiled after backing out the bug fix. 19.56 ORA-00600: internal error code, arguments: [kcbgtcr_6], [], [], [], [], [], [], [] ======================================================================================== Doc ID: Note:248874.1 Subject: ORA-600 [kcbgtcr_6] Type: REFERENCE Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 18-SEP-2003 Last Revision Date: 25-MAR-2004 This note contains information that has not yet been reviewed by DDR. As such, the contents are not necessarily accurate and care should be taken when dealing with customers who have encountered this error. Thanks. PAA Internals Group Note: For additional ORA-600 related information please read Note 146580.1 PURPOSE: This article discusses the internal error "ORA-600 [kcbgtcr_6]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [kcbgtcr_6] [a] VERSIONS: versions 8.0 to 10.1 DESCRIPTION: Two buffers have been found in the buffer cache that are both current and for the same DBA (Data Block Address). We should not have two 'current' buffers for the same DBA in the cache, if this is the case then this error is raised. ARGUMENTS: Arg [a] Buffer class Note that for Oracle release 9.2 and earlier there are no additional arguments reported with this error. FUNCTIONALITY: Kernel Cache Buffer management IMPACT: PROCESS FAILURE POSSIBLE INSTANCE FAILURE NON CORRUPTIVE - No underlying data corruption. SUGGESTIONS: Retry the operation. Does the error still occur after an instance bounce? If using 64bit AIX then ensure that minimum version in use is 9.2.0.3 or patch for Bug 2652771 has been applied. If the Known Issues section below does not help in terms of identifying a solution, please submit the trace files and alert.log to Oracle Support Services for further analysis. Known Issues: Bug 2652771 Shared data structures corrupted around latch code on 64bit AIX ports. Fixed 9.2.0.3 backports available for older versions (8.1.7) from Metalink. ORA-600 [kcbgtcr_6] Versions: 8.0.5 - 10.1 Source: kcb.c Meaning: We have two 'CURRENT' buffers for the same DBA. Argument Description: None --------------------------------------------------------------------------- Explanation: We have identified two 'CURRENT' buffers for the same DBA in the cache, this is incorrect, and this error will be raised. --------------------------------------------------------------------------- Diagnosis: Check the trace file, this will show the buffers i.e :- BH (0x70000003ffe9800) file#: 39 rdba: 0x09c131e6 (39/78310) class 1 ba: 0x70000003fcf0000 set: 6 dbwrid: 0 obj: 11450 objn: 11450 hash: [70000000efa9b00,70000004d53a870] lru: [70000000efa9b68,700000006fb8d68] ckptq: [NULL] fileq: [NULL] st: XCURRENT md: NULL rsop: 0x0 tch: 1 LRBA: [0x0.0.0] HSCN: [0xffff.ffffffff] HSUB: [255] RRBA: [0x0.0.0] BH (0x70000000efa9b00) file#: 39 rdba: 0x09c131e6 (39/78310) class 1 ba: 0x70000000e4f6000 set: 6 dbwrid: 0 obj: 11450 objn: 11450 hash: [70000004d53a870,70000003ffe9800] lru: [700000012fbaf68,70000003ffe9868] ckptq: [NULL] fileq: [NULL] st: XCURRENT md: NULL rsop: 0x0 tch: 2 LRBA: [0x0.0.0] HSCN: [0xffff.ffffffff] HSUB: [255] RRBA: [0x0.0.0] Here it is clear that we have two current buffers for the dba. Most likely cause for this is 64bit AIX Bug 2652771. If this isn't the case check the error reproduces consistently after bouncing the instance? Via SQLplus? What level of concurrency to reproduce? Is a testcase available? Check OS memory for errors. --------------------------------------------------------------------------- Known Bugs: Bug 2652771 Shared data structures corrupted around latch code on 64bit AIX ports. - Fixed 9.2.0.3, backports available for older versions. 19.57 ORA-00600: internal error code, arguments: [1100], [0x7000002FDF83F40], [0x7000002FDF83F40], [], [], [], [], [] ===================================================================================================================== Doc ID: Note:138123.1 Subject: ORA-600 [1100] Type: REFERENCE Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 28-MAR-2001 Last Revision Date: 08-FEB-2005 Note: For additional ORA-600 related information please read Note 146580.1 PURPOSE: This article discusses the internal error "ORA-600 [1100]", what it means and possible actions. The information here is only applicable to the versions listed and is provided only for guidance. ERROR: ORA-600 [1100] [a] [b] [c] [d] [e] VERSIONS: versions 6.0 to 9.2 DESCRIPTION: This error relates to the management of standard double-linked (forward and backward) lists. Generally, if the list is damaged an attempt to repair the links is performed. Additional information will accompany this internal error. A dump of the link and often a core dump will coincide with this error. This is a problem with a linked list structure in memory. FUNCTIONALITY: GENERIC LINKED LISTS IMPACT: PROCESS FAILURE POSSIBLE INSTANCE FAILURE IF DETECTED BY PMON PROCESS No underlying data corruption. SUGGESTIONS: Known Issues: Bug# 3724548 See Note 3724548.8 OERI[kglhdunp2_2] / OERI[1100] under high load Fixed: 9.2.0.6, 10.1.0.4, 10.2 Bug# 3691672 + See Note 3691672.8 OERI[17067]/ OERI[26599] / dump (kgllkdl) from JavaVM / OERI:1100 from PMON Fixed: 10.1.0.4, 10.2 Bug# 2652771 P See Note 2652771.8 AIX: OERI[1100] / OERI[KCBGET_37] SGA corruption Fixed: 9.2.0.3 Bug# 1951929 See Note 1951929.8 ORA-7445 in KQRGCU/kqrpfr/kqrpre possible Fixed: 8.1.7.3, 9.0.1.2, 9.2.0.1 Bug# 959593 See Note 959593.8 CTRL-C During a truncate crashes the instance Fixed: 8.1.6.3, 8.1.7.0 INTERNAL ONLY SECTION - NOT FOR PUBLICATION OR DISTRIBUTION TO CUSTOMERS No internal information at the present time. Ensure that this note comes out on top in Metalink when searched ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 ora-600 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 1100 Note 2: ------- Doc ID: Note:3724548.8 Subject: Bug 3724548 - OERI[kglhdunp2_2] / OERI[1100] under high load Type: PATCH Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 24-SEP-2004 Last Revision Date: 13-JAN-2005 Click here for details of sections in this note. Bug 3724548 OERI[kglhdunp2_2] / OERI[1100] under high load This note gives a brief overview of bug 3724548. Affects: Product (Component) Oracle Server (Rdbms) Range of versions believed to be affected Versions < 10.2 Versions confirmed as being affected 9.2.0.4 9.2.0.5 Platforms affected Generic (all / most platforms affected) Fixed: This issue is fixed in 9.2.0.6 (Server Patch Set) 10.1.0.4 (Server Patch Set) 10g Release 2 (future version) Symptoms: Related To: Memory Corruption Internal Error May Occur (ORA-600) ORA-600 [kglhdunp2_2] ORA-600 [1100] (None Specified) Description When an instance is under high load it is possible for sessions to get ORA-600[KGLHDUNP2_2] and ORA-600 [1100] errors. This can also show as a corrupt linked list in the SGA. The full bug text (if published) can be seen at (This link will not work for UNPUBLISHED bugs) You can search for any interim patches for this bug here (This link will Error if no interim patches exist) 19.58 Compilation problems DBI DBD: =================================== We upgraded Oracle from 8.1.6 to 9.2.0.5 and I tried to rebuild the DBD::Oracle module but it threw errors like: . gcc: unrecognized option `-q64' ld: 0711-736 ERROR: Input file /lib/crt0_64.o: XCOFF64 object files are not allowed in 32-bit mode. collect2: ld returned 8 exit status make: 1254-004 The error code from the last command is 1. Stop. After some digging I found out that this is because the machine is AIX 5.2 running under 32-bit and it is looking at the oracle's lib directory which has 64 bit libraries. So after running "perl Makefile.PL", I edited the Makefile 1. changing the references to Oracle's ../lib to ../lib32, 2. changing change crt0_64.o to crt0_r.o. 3. Remove the -q32 and/or -q64 options from the list of libraries to link with. Now when I ran "make" it went smoothly, so did make test and make install. I ran my own simple perl testfile which connects to the Oracle and gets some info and it works fine. Now I have an application which can be customised to call perl scripts and when I call this test script from that application it fails with: install_driver(Oracle) failed: Can't load '/usr/local/perl/lib/site_perl/5.8.5/a ix/auto/DBD/Oracle/Oracle.so' for module DBD::Oracle: 0509-022 Cannot load mod ule /usr/local/perl/lib/site_perl/5.8.5/aix/auto/DBD/Oracle/Oracle.so. 0509-150 Dependent module /u00/oracle/product/9.2.0/lib/libclntsh.a(sh r.o) could not be loaded. 0509-103 The module has an invalid magic number. 0509-022 Cannot load module /u00/oracle/product/9.2.0/lib/libclntsh.a. 0509-150 Dependent module /u00/oracle/product/9.2.0/lib/libclntsh.a co uld not be loaded. at /usr/local/perl/lib/5.8.5/aix/DynaLoader.pm line 230. at (eval 3) line 3 Compilation failed in require at (eval 3) line 3. Perhaps a required shared library or dll isn't installed where expected at /opt/dscmdevc/src/udps/test_oracle_dbd.pl line 45 whats happening here is that the application sets its own LIBPATH to include oracle's lib(instead of lib32) in the beginning and that makes perl look at the wrong place for the file - libclntsh.a .Unfortunately it will take too long for the application developers to change this in their application and I am looking for a quick solution. The test script is something like: use Env; use strict; use lib qw( /opt/harvest/common/perl/lib ) ; #use lib qw( $ORACLE_HOME/lib32 ) ; use DBI; my $connect_string="dbi:Oracle:"; my $datasource="d1ach2"; $ENV{'LIBPATH'} = "${ORACLE_HOME}/lib32:$ENV{'LIBPATH'}" ; . . my $dbh = DBI->connect($connect_string, $dbuser, $dbpwd) or die "Can't connect to $datasource: $DBI::errstr"; . . Adding 'use lib' or using'$ENV{LIBPATH}' to change the LIBPATH is not working because I need to make this work in this perl script and the "use DBI" is run (or whatever the term is) in the compile-phase before the LIBPATH is set in the run-phase. I have a work around for it: write a wrapper ksh script which exports the LIBPATH and then calls the perl script which works fine but I was wondering if there is a way to set the libpath or do something else inside the current perl script so that it knows where to look for the right library files inspite of the wrong LIBPATH? Or did I miss something when I changed the Makefile and did not install everything right? Is there anyway I check this? (the make install didnot throw any errors) Any help or thoughts on this would be much appreciated. Thanks! Rachana. note 12: -------- P550:/ # find . -name "libclnt*" -print ./apps/oracle/product/9.2/lib/libclntst9.a ./apps/oracle/product/9.2/lib/libclntsh.a ./apps/oracle/product/9.2/lib32/libclntst9.a ./apps/oracle/product/9.2/lib32/libclntsh.a ./apps/oracle/oui/bin/aix/libclntsh.so.9.0 P550:/ # 19.59 Listener problem: IBM/AIX RISC System/6000 Error: 13: Permission denied ----------------------------------------------------------------------------- When starting listener start listener TNS-12546: TNS:permission denied TNS-12560: TNS:protocol adapter error TNS-00516: Permission denied IBM/AIX RISC System/6000 Error: 13: Permission denied Note 1: 'TNS-12531: TNS:cannot allocate memory' may be misleading, it seems to be a permission problem (see also IBM/AIX RISC System/6000 Error: 13: Permission denied). A possible reason is: Oracle (more specific the listener) is unable to read /etc/hosts, because of permission problems. So host resolution is not possible. .. .. The problem really was in permissions of /etc/hosts on the node2. It was -rw-r----- (640). Now it is -rw-rw-r-- (664) and everything goes ok. Thank you! BUGS WITH REGARDS TO PRO*COBOL ON 9i: 10.59 Listener problem: IBM/AIX RISC System/6000 Error: 79: Connection refused ------------------------------------------------------------------------------ d0planon@zb121l01:/data/oracle/d0planon/admin/home/$ lsnrctl LSNRCTL for IBM/AIX RISC System/6000: Version 10.2.0.3.0 - Production on 12-OCT-2007 08:29:14 Copyright (c) 1991, 2006, Oracle. All rights reserved. Welcome to LSNRCTL, type "help" for information. LSNRCTL> status Connecting to (ADDRESS=(PROTOCOL=tcp)(HOST=)(PORT=1521)) TNS-12541: TNS:no listener TNS-12560: TNS:protocol adapter error TNS-00511: No listener IBM/AIX RISC System/6000 Error: 79: Connection refused Answer 1: Check if the oracle user can read /etc/hosts Answer 2: Maybe there are multiple instances of the listener, so if you try the following LSNRCTL> status You might have a correct response. 19.61: 64BIT PRO*COBOL IS NOT THERE EVNN AFTER UPGRDING TO 9.2.0.3 ON AIX-5L BOX -------------------------------------------------------------------------------- Bookmark Fixed font Go to End Monitor Bug Bug No. 2859282 Filed 19-MAR-2003 Updated 01-NOV-2003 Product Precompilers Product Version 9.2.0.3 Platform AIX5L Based Systems (64-bit) Platform Version 5.* Database Version 9.2.0.3 Affects Platforms Port-Specific Severity Severe Loss of Service Status Closed, Duplicate Bug Base Bug 2440385 Fixed in Product Version No Data Problem statement: 64BIT PRO*COBOL IS NOT THERE EVNN AFTER UPGRDING TO 9.2.0.3 ON AIX-5L BOX *** 03/19/03 10:13 am *** 2889686.996 . ========================= PROBLEM: . 1. Clear description of the problem encountered: . cst. has upgraded from 9.2.0.2 to 9.2.0.3 on a AIX 5L 64-Bit Box and is not seeing the 64-bit Procob executable. Actually the same problem existed when upgraded from 9.2.0.1 to 9.2.0.2, but the one-off patch has been provided in the Bug#2440385 to resolve the issue. As per the Bug, problem has been fixed in 9.2.0.3. But My Cst. is facing the same problem on 9.2.0.3 also. . This is what the Cst. says ============================ This is the original bug # 2440385. The fix provides 64 bit versions of Pro*Cobol.There are two versions of the patch for the bug: one is for the 9.2.0.1 RDBMS and the other is for 9.2.0.2. So the last time I hit this issue, I applied the 9.2.0.2 RDBMS patch to the 9.2.0.1 install. The 9.2.0.2 patch also experienced the relinking problem on rtsora just like the 9.2.0.1 install did. I ignored the error to complete the patch application. Then I used the patch for the 2440385 bug to get 64 bit procob/rtsora executables (the patch actually provides executables rather than performing a successful relinking) to get the Pro*Cobol 1.8.77 precompiler to work with the MicroFocus Server Express 2.0.11 (64 bit) without encountering "bad magic number" error. . The current install that I am performing I've downloaded the Oracle 9.2.0.3 Pro*Cobol capability fix either so the rtsora relinking fails as well. Thus I don't have a working Pro*Cobol precompiler to allow me to generate our Cobol programs against the database. . 2. Pertinent configuration information (MTS/OPS/distributed/etc) . 3. Indication of the frequency and predictability of the problem . 4. Sequence of events leading to the problem . 5. Technical impact on the customer. Include persistent after effects. . ========================= DIAGNOSTIC ANALYSIS: . One-off patch should be provided on top of 9.2.0.3 as provided on top of 9.2.0.2/9.2.0.1 . ========================= WORKAROUND: . . ========================= RELATED BUGS: . 2440385 . ========================= REPRODUCIBILITY: . 1. State if the problem is reproducible; indicate where and predictability . 2. List the versions in which the problem has reproduced . 9.2.0.3 . 3. List any versions in which the problem has not reproduced Further notes on PRO*COBOL: =========================== Note 1: ======= 9201,9202,9203,9204,9205 32 bit cobol: procob32 or procob18_32. 64 bit cobol: procob or procob18 PATCHES: 1. Patch 2663624: (Cobol patch for 9202 AIX 5L) ----------------------------------------------- PSE FOR BUG2440385 ON 9.2.0.2 FOR AIX5L PORT 212 Patchset Exception: 2663624 / Base Bug 2440385 #------------------------------------------------------------------------- # # DATE: November 26, 2002 # ----------------------- # Platform Patch for : AIX Based Systems (Oracle 64bit) for 5L # Product Version # : 9.2.0.2 # Product Patched : RDBMS # # Bugs Fixed by this patch: # ------------------------- # 2440385 : PLEASE PROVIDE THE PATCH FOR SUPPORTING 64BIT PRO*COBOL # # Patch Installation Instructions: # -------------------------------- # To apply the patch, unzip the PSE container file; # # % unzip p2440385_9202_AIX64-5L.zip # # Set your current directory to the directory where the patch # is located: # # % cd 2663624 # # Ensure that the directory containing the opatch script appears in # your $PATH; then enter the following command: # # % opatch apply 2. Patch 2440385: ----------------- Results for Platform : AIX5L Based Systems (64-bit) Patch Description Release Updated Size 2440385 Pro*COBOL: PATCH FOR SUPPORTING 64BIT PRO*COBOL 9.2.0.3 27-APR-2003 34M 2440385 Pro*COBOL: PATCH FOR SUPPORTING 64BIT PRO*COBOL 9.2.0.2 26-NOV-2002 17M 2440385 Pro*COBOL: PATCH FOR SUPPORTING 64BIT PRO*COBOL 9.2.0.1 01-OCT-2002 17M 3. Patch 3501955 9205: ---------------------- Also includes 2440385. Provide the patch for supporting 64-bit Pro*COBOL. Note 2: ======= Problem precompiling Cobol program under Oracle 9i...... Hi, we recently upgraded to 9i. However, we still have 32 bit Cobol, so we're using the procob18_32 precompiler to compile our programs. Some of my compiles have worked successfully. However, I'm receiving the follow error in one of my compiles: 1834 183400 01 IB0-STATUS PIC 9. 7SA 350 1834 ...................................^ PCC-S-0018: Expected "PICTURE clause", but found "9" at line 1834 in file What's strange is that if I compile the program against the same DB using procob instead of procob18_32, it compiles cleanly. I noticed in my compile that failed using procob18_32, it had the following message: System default option values taken from: /u01/app/oracle/product/9.2.0.4/precomp /admin/pcccob.cfg Yet, when I used procob, it had this message: System default option values taken from: /u01/app/oracle/product/9.2.0.4/precomp /admin/pcbcfg.cfg .. .. Hi, I started using procob32 instead of procob18_32, and that resolved my problem. Thanks for any help you may have already started to provide. Note 3: ======= Doc ID: Note:257934.1 Content Type: TEXT/X-HTML Subject: Pro*COBOL Application Fails in Runtime When Using Customized old Make Files With Signal 11 (MF Errror 114) Creation Date: 20-NOV-2003 Type: PROBLEM Last Revision Date: 04-APR-2005 Status: MODERATED The information in this article applies to: Precompilers - Version: 9.2.0.4 This problem can occur on any platform. Symptoms After upgrading from Oracle server and Pro*COBOL 9.2.0.3.0 to 9.2.0.4.0 application are failing with cobol runtime error 114 when using 32-bit builds. Platform is AIX 4.3.3 which does not support 64-bit builds with Micro Focus Server Express 2.0.11. Execution error : file 'sample1' error code: 114, pc=0, call=1, seg=0 114 Attempt to access item beyond bounds of memory (Signal 11) Changes Upgraded from 9.2.0.3.0 to 9.2.0.4.0. Cause The customized old make files for building 32-bit applications invoked the 64-bit precompilers procob or procob18 instead of procob32 or procob18_32. Fix Use the Oracle Supplied make templates or change the customized old make files for 32-bit application builds $ORACLE_HOME/precomp/demo/procob2/demo_procob_32.mk, $ORACLE_HOME/precomp/demo/procob/demo_procob_32.mk and $ORACLE_HOME/precomp/demo/procob/demo_procob18_32.mk invoke the wrong precompiler. To fix the problem add the following to $ORACLE_HOME/precomp/demo/procob2/demo_procob_32.mk: PROCOB=procob32 Using $ORACLE_HOME/precomp/demo/procob/demo_procob_32.mk: PROCOB_32=procob32 Using $ORACLE_HOME/precomp/demo/procob/demo_procob18_32.mk PROCOB18_32=procob18_32 The change can be added to the bottom of the make file. References Bug 3220095 - Procobol App Fails114 Attempt To Access Item Beyond Bounds Of Memory (Signal 11) Note 4: ======= Displayed below are the messages of the selected thread. Thread Status: Closed From: Jean-Daniel DUMAS 23-Nov-04 16:39 Subject: PROCOB18_32 Problem at execution ORA-00933 PROCOB18_32 Problem at execution ORA-00933 We try to migrate from Oracle 8.1.7.4 to Oracle 9.2.0.5. We've got problems with a lot of procobol programs using host table variables in PL SQL blocks like: EXEC SQL EXECUTE BEGIN FOR nIndice IN 1..:WI-NB-APPELS-TFO009S LOOP UPDATE tmp_edition_erreur SET mon_nb_dec = :WTI-S2-MON-NB-DEC (nIndice) WHERE mon_cod = :WTC-S2-MON-COD (nIndice) AND run_id = :WC-O-RUN-ID; END LOOP; END; END-EXEC At execution, we've got "ORA-00933 SQL command not properly ended". The problem seems to appear only if the host table variable is used inside a SELECT,UPDATE or DELETE command. For the INSERT VALUES command, it seems that we've got no problem. A workaround consists to assign host table variables into oracle table variables and replace inside SQL command host table variables by oracle table variables. But, as we've got a lot a program like this, we don't enjoy to do this. Have somebody another idea ? jddumas@eram.fr From: Oracle, Amit Joshi 05-Jan-05 06:26 Subject: Re : PROCOB18_32 Problem at execution ORA-00933 Hi Please refer to bug 3802067 on Metalink. From the details provided , it seems you are hitting the same. Best Regards Amit Joshi Note 5: ======= Re: Server Express 64bit and Oracle 9i problem (114) on AIX 5.2 Hi Wayne (and Panos) Apologies if you're aware of some of this already, but I just wanted to clarify the steps involved in creating and executing a Pro*COBOL application with Micro Focus Server Express on UNIX. When installing Pro*COBOL on UNIX (as part of the main Oracle installation), you need to have your COBOL environment setup, in order for the installer to relink a COBOL RTS containing the Oracle support libraries (rtsora/rtsora32/rtsora64). The 64-bit edition of Oracle 9i on AIX 5.x creates rtsora -- the 64-bit version of the run-time -- and rtsora32 -- the 32-bit version of the run-time. It's imperative that you use the correct edition of Server Express, i.e. 32-bit or 64-bit -- note well, that these are separate products on this platform -- for the mode in which you wish to use Oracle. In addition, you need to ensure that LIBPATH is set to point to the correct Oracle 'lib' directory -- $ORACLE_HOME/lib32 for 32-bit, or $ORACLE_HOME/lib for 64-bit If you wish to recreate those executables, say if you've updated your COBOL environment since installing Oracle, then from looking at the makefiles -- ins_precomp.mk and env_precomp.mk -- then the effective commands to use to re-link the run-time correctly are as follows (logged in under your Oracle user ID) : either mode: export PATH=$COBDIR/bin:$ORACLE_HOME/bin:$PATH 32-bit : export LIBPATH=$COBDIR/lib:$ORACLE_HOME/lib32:$LIBPATH cd $ORACLE_HOME/precomp/lib make LIBDIR=lib32 -f ins_precomp.mk EXE=rtsora32 rtsora32 64-bit: export LIBPATH=$COBDIR/lib:$ORACLE_HOME/lib:$LIBPATH cd $ORACLE_HOME/precomp/lib make -f ins_precomp.mk rtsora Regarding precompiling your application, Oracle provide two versions of Pro*COBOL. Again, you need to use the correct one depending on whether you're creating a 32-bit or 64-bit application, as the precompiler will generate different code. If invoking Pro*COBOL directly, you need to use : 32-bit : procob32 / procob18_32 , e.g. procob32 myapp.pco cob -it myapp.cob rtsora32 myapp.int or 64-bit : procob / procob18 , e.g. procob myapp.pco cob -it myapp.cob rtsora myapp.int If you're using Server Express 2.2 SP1 or later, you can also compile using the Cobsql preprocessor, which will invoke the correct version of Pro*COBOL under the covers, allowing for a single precompile-compile step, e.g. cob -ik myapp.pco -C "p(cobsql) csqlt==oracle8 endp" This method also aids debugging, as you will see the original source code while animating, rather than the output from the precompiler. See the Server Express Database Access manual. Prior to SX 2.2 SP1, Cobsql only supported the creation of 32-bit applications. I hope this helps -- if you're still having problems, please let me know. Regards, SimonT. Re: Re: Server Express 64bit and Oracle 9i problem (114) on AIX 5.2 Hi Simon (and anyone else) Thanks for that. We still seem to be getting a very unusual error with our c ompiles in or makes. A bit of background: we are "upgrading" from Oracle8i, SAS6, Solaris, MF COB OL 4.5 to AIX 5L, Oracle9i, SAS8 and MF Server Express COBOL. When we attempt to compile our COBOL it works fine. However if the COBOL has embedded Oracle SQL our procomp makes try to access ADA. We do not use ADA. I thought this must have been included by accident; but can find no flag or install option for it. So can you give us any clues as to why we are suffer ing an ADA plague :-)) Wayne Re: Server Express 64bit and Oracle 9i problem (114) on AIX 5.2 Hi Wayne. On the surface, it appears as if you're not picking up the correct Pro*COBOL binary. If you invoke 'procob' from the command line, you should see something along the lines of : Pro*COBOL: Release 9.2.0.4.0 - Production on Mon Apr 19 13:38:07 2004 followed by a list of Pro*COBOL options. Do you see this, or do you see a different banner (say, Pro*ADA, or Pro*Fortran)? Assuming you see something other than a Pro*COBOL banner, then if you invoke 'whence procob', does it show procob as being picked up from your Oracle bin directory (/home/oracle/9.2.0/bin/procob in my case) ? If you're either not seeing the correct Pro*COBOL banner, or it's not located in the correct directory, I'd suggest rebuilding the procob and procob32 binaries. Logged in under your Oracle user ID, with the Oracle environment set up : cd $ORACLE_HOME/precomp/lib make -f ins_precomp.mk procob32 procob and then try your compilation process again. Regards, SimonT. Re: Re: Server Express 64bit and Oracle 9i problem (114) on AIX 5.2 Hi Simon Firstly, thanks for all your help, it was greatly appreciated. We have the solution to our problem: The problem is resolved by modifying the line in the job from: make -f $SRC_DIR/procob.mk COBS="$SRC_DIR/PFEM025A.cob SYSDATE.cob CNTLGET. cob" EXE=$SRC_DIR/PFEM025A to make -f $SRC_DIR/procob.mk build COBS="$SRC_DIR/PFEM025A.cob SYSDATE.cob CN TLGET.cob" EXE=$SRC_DIR/PFEM025A It appears this (build keyword) is not a requirement for the job to run on S olaris but is for AIX. All is working fine. Cheers Wayne Note 6: ======= Doc ID: Note:2440385.8 Content Type: TEXT/X-HTML Subject: Support Description of Bug 2440385 Creation Date: 08-AUG-2003 Type: PATCH Last Revision Date: 15-AUG-2003 Status: PUBLISHED Click here for details of sections in this note. Bug 2440385 AIX: Support for 64 bit ProCobol This note gives a brief overview of bug 2440385. Affects: Product (Component) Precompilers (Pro*COBOL) Range of versions believed to be affected Versions >= 7 but < 10G Versions confirmed as being affected 9.2.0.3 Platforms affected Aix 64bit 5L Fixed: This issue is fixed in 9.2.0.4 (Server Patch Set) Symptoms: (None Specified) Related To: Pro* Precompiler Description Add support for 64 bit ProCobol The full bug text (if published) can be seen at Bug 2440385 This link will not work for UNPUBLISHED bugs. Note 7: ======= Displayed below are the messages of the selected thread. Thread Status: Closed From: Cathy Agada 18-Sep-03 21:40 Subject: How do I relink rtsora for 64 bit processing How do I relink rtsora for 64 bit processing I have the following error while relinking "rtsora" on AIX 5L/64bit platform on oracle 9.2.0.3 (I believe my patch is up-to-date). Our Micro Focus compiler version is 2.0.11 $>make -f ins_precomp.mk relink EXENAME=rtsora /bin/make -f ins_precomp.mk LIBDIR=lib32 EXE=/app/oracle/product/9.2.0/precomp/lib/rtsora rtsora32 Linking /app/oracle/product/9.2.0/precomp/lib/rtsora cob64: bad magic number: /app/oracle/product/9.2.0/precomp/lib32/cobsqlintf.o make: 1254-004 The error code from the last command is 1. Stop. make: 1254-004 The error code from the last command is 2. My environment variable is as follows: COBDIR=/usr/lpp/cobol LD_LIBRARY_PATH=$ORACLE_HOME/lib:/app/oracle/product/9.2.0/network/lib SHLIB_PATH=$ORACLE_HOME/lib64:/app/oracle/product/9.2.0/lib32 I added 'define=bit64' on precomp config file. Any ideas on what could be wrong. Thanks. From: Oracle, Amit Chitnis 19-Sep-03 05:26 Subject: Re : How do I relink rtsora for 64 bit processing Cathy, Support for 64 bit Pro*Cobol 9.2.0.3 on AIX 5.1 was provided through one off patch for bug 2440385 You will need to download and apply the patch for bug 2440385. ==OR== You can dowload and apply the latest 9.2.0.4 patchset where the bug is fixed. Thanks, Amit Chitnis. Note 8: ======= Doc ID: Note:215279.1 Content Type: TEXT/X-HTML Subject: Building Pro*COBOL Programs Fails With "cob64: bad magic number:" Creation Date: 08-APR-2003 Type: PROBLEM Last Revision Date: 15-APR-2003 Status: PUBLISHED fact: Pro*COBOL 9.2.0.2 fact: Pro*COBOL 9.2.0.1 fact: AIX-Based Systems (64-bit) symptom: Building Pro*COBOL programs fails symptom: cob64: bad magic number: %s symptom: /oracle/product/9.2.0/precomp/lib32/cobsqlintf.o cause: Bug 2440385 AIX: Support for 64 bit ProCobol fix: This is fixed in Pro*COBOL 9.2.0.3 One-Off patch for Pro*COBOL 9.2.0.2 has been provided in Metalink Patch Number 2440385 Reference: How to Download a Patch from Oracle Note 9: ======= If you wish to recreate those executables, say if you've updated your COBOL environment since installing Oracle, then from looking at the makefiles -- ins_precomp.mk and env_precomp.mk -- then the effective commands to use to re-link the run-time correctly are as follows (logged in under your Oracle user ID) : either mode: export PATH=$COBDIR/bin:$ORACLE_HOME/bin:$PATH 32-bit : export LIBPATH=$COBDIR/lib:$ORACLE_HOME/lib32:$LIBPATH cd $ORACLE_HOME/precomp/lib make LIBDIR=lib32 -f ins_precomp.mk EXE=rtsora32 rtsora32 64-bit: export LIBPATH=$COBDIR/lib:$ORACLE_HOME/lib:$LIBPATH cd $ORACLE_HOME/precomp/lib make -f ins_precomp.mk rtsora Note 10: ======== On 9.2.0.5, try to get the pro cobol patch for 9203. Then just copy the procobol files to the cobol directory. 19.62: ORA-12170: ================= Connection Timeout. Doc ID: Note:274303.1 Content Type: TEXT/X-HTML Subject: Description of parameter SQLNET.INBOUND_CONNECT_TIMEOUT Creation Date: 26-MAY-2004 Type: BULLETIN Last Revision Date: 10-FEB-2005 Status: MODERATED *** This article is being delivered in Draft form and may contain errors. Please use the MetaLink "Feedback" button to advise Oracle of any issues related to this article. *** PURPOSE ------- To specify the time, in seconds, for a client to connect with the database server and provide the necessary authentication information. Description of parameter SQLNET.INBOUND_CONNECT_TIMEOUT ------------------------------------------------------- This parameter has been introduced in 9i version. This has to be configured in sqlnet.ora file. Use the SQLNET.INBOUND_CONNECT_TIMEOUT parameter to specify the time, in seconds, for a client to connect with the database server and provide the necessary authentication information. If the client fails to establish a connection and complete authentication in the time specified, then the database server terminates the connection. In addition, the database server logs the IP address of the client and an ORA-12170: TNS:Connect timeout occurred error message to the sqlnet.log file. The client receives either an ORA-12547: TNS:lost contact or an ORA-12637: Packet receive failed error message. Without this parameter, a client connection to the database server can stay open indefinitely without authentication. Connections without authentication can introduce possible denial-of-service attacks, whereby malicious clients attempt to flood database servers with connect requests that consume resources. To protect both the database server and the listener, Oracle Corporation recommends setting this parameter in combination with the INBOUND_CONNECT_TIMEOUT_listener_name parameter in the listener.ora file. When specifying values for these parameters, consider the following recommendations: *Set both parameters to an initial low value. *Set the value of the INBOUND_CONNECT_TIMEOUT_listener_name parameter to a lower value than the SQLNET.INBOUND_CONNECT_TIMEOUT parameter. For example, you can set INBOUND_CONNECT_TIMEOUT_listener_name to 2 seconds and INBOUND_CONNECT_TIMEOUT parameter to 3 seconds. If clients are unable to complete connections within the specified time due to system or network delays that are normal for the particular environment, then increment the time as needed. By default is set to None Example SQLNET.INBOUND_CONNECT_TIMEOUT=3 RELATED DOCUMENTS ----------------- Oracle9i Net Services Reference Guide, Release 2 (9.2), Part Number A96581-02 SQLNET.EXPIRE_TIME: ------------------- Purpose: Determines time interval to send a probe to verify the session is alive See Also: Oracle Advanced Security Administrator's Guide Default: None Minimum Value: 0 minutes Recommended Value: 10 minutes Example: sqlnet.expire_time=10 sqlnet.expire_time Enables dead connection detection, that is, after the specifed time (in minutes) the server checks if the client is still connected. If not, the server process exits. This parameter must be set on the server PROBLEM: Long query (20 minutes) returns ORA-01013 after about a minute. SOLUTION: The SQLNET.ORA parameter SQLNET.EXPIRE_TIME was set to a one(1). The parameter was changed to... SQLNET.EXPIRE_TIME=2147483647 This allowed the query to complete. This is documented in the Oracle Troubleshooting manual on page 324. The manual part number is A54757.01. Keywords: SQLNET.EXPIRE_TIME,SQLNET.ORA,ORA-01013 sqlnet.expire_time should be set on the server. The server sends keep alive traffic over connections that have already been established. You won't need to change your firewall. sqlnet.expire_time is actually intended to test connections in order to allow oracle to clean up resources from connection that abnormally terminated. The architecture to do that means that the server will send a probe packet to the client. That probe packet is viewed by the most firewalls as traffic on the line. That will in short reset the idle timers on the firewall. If you happen to have the disconnects from idle timers then it may help. It was not intended for that feature but it is a byproduct of the design. 19.63: Tracing SQLNET: ====================== Note 1: ------- Doc ID: Note:219968.1 Subject: SQL*Net, Net8, Oracle Net Services - Tracing and Logging at a Glance Type: BULLETIN Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 20-NOV-2002 Last Revision Date: 26-AUG-2003 TITLE ----- SQL*Net, Net8, Oracle Net Services - Tracing and Logging at a Glance. PURPOSE ------- The purpose of Oracle Net tracing and logging is to provide detailed information to track and diagnose Oracle Net problems such as connectivity issues, abnormal disconnection and connection delay. Tracing provides varying degrees of information that describe connection-specific internal operations during Oracle Net usage. Logging reports summary, status and error messages. Oracle Net Services is the replacement name for the Oracle Networking product formerly known as SQL*Net (Oracle7 [v2.x]) and Net8 (Oracle8/8i [v8.0/8.1]). For consistency, the term Oracle Net is used thoughout this article and refers to all Oracle Net product versions. SCOPE & APPLICATION ------------------- The aim of this document is to overview SQL*Net, Net8, Oracle Net Services tracing and logging facilities. The intended audience includes novice Oracle users and DBAs alike. Although only basic information on how to enable and disable tracing and logging features is described, the document also serves as a quick reference. The document provides the reader with the minimum information necessary to generate trace and log files with a view to forwarding them to Oracle Support Services (OSS) for further diagnosis. The article does not intend to describe trace/log file contents or explain how to interpret them. LOG & TRACE PARAMETER OVERVIEW ------------------------------ The following is an overview of Oracle Net trace and log parameters. TRACE_LEVEL_[CLIENT|SERVER|LISTENER] = [0-16|USER|ADMIN|SUPPORT|OFF] TRACE_FILE_[CLIENT|SERVER|LISTENER] = TRACE_DIRECTORY_[CLIENT|SERVER|LISTENER] = TRACE_UNIQUE_[CLIENT|SERVER|LISTENER] = [ON|TRUE|OFF|FALSE] TRACE_TIMESTAMP_[CLIENT|SERVER|LISTENER] = [ON|TRUE|OFF|FALSE] #Oracle8i+ TRACE_FILELEN_[CLIENT|SERVER|LISTENER] = #Oracle8i+ TRACE_FILENO_[CLIENT|SERVER|LISTENER] = #Oracle8i+ LOG_FILE_[CLIENT|SERVER|LISTENER] = LOG_DIRECTORY_[CLIENT|SERVER|LISTENER] = LOGGING_LISTENER = [ON|OFF] TNSPING.TRACE_LEVEL = [0-16|USER|ADMIN|SUPPORT|OFF] TNSPING.TRACE_DIRECTORY = NAMES.TRACE_LEVEL = [0-16|USER|ADMIN|SUPPORT|OFF] NAMES.TRACE_FILE = NAMES.TRACE_DIRECTORY = NAMES.TRACE_UNIQUE = [ON|OFF] NAMES.LOG_FILE = NAMES.LOG_DIRECTORY = NAMES.LOG_UNIQUE = [ON|OFF] NAMESCTL.TRACE_LEVEL = [0-16|USER|ADMIN|SUPPORT|OFF] NAMESCTL.TRACE_FILE = NAMESCTL.TRACE_DIRECTORY = NAMESCTL.TRACE_UNIQUE = [ON|OFF] Note: With the exception of parameters suffixed with LISTENER, all other parameter suffixes and prefixes [CLIENT|NAMES|NAMESCTL|SERVER|TNSPING] are fixed and cannot be changed. For parameters suffixed with LISTENER, the suffix name should be the actual Listener name. For example, if the Listener name is PROD_LSNR, an example trace parameter name would be TRACE_LEVEL_PROD_LSNR=OFF. CONFIGURATION FILES ------------------- Files required to enable Oracle Net tracing and logging features include: Oracle Net Listener LISTENER.ORA LISTENER.TRC Oracle Net - Client SQLNET.ORA on client SQLNET.TRC Oracle Net - Server SQLNET.ORA on server SQLNET.TRC TNSPING Utility SQLNET.ORA on client/Server TNSPING.TRC Oracle Name Server NAMES.ORA NAMES.TRC Oracle NAMESCTL SQLNET.ORA on server Oracle Connection Manager CMAN.ORA CONSIDERATIONS WHEN USING LOGGING/TRACING ----------------------------------------- 1. Verify which Oracle Net configuration files are in use. By default, Oracle Net configuration files are sought and resolved from the following locations: TNS_ADMIN environment variable (incl. Windows Registry Key) /etc or /var/opt/oracle (Unix) $ORACLE_HOME/network/admin (Unix) %ORACLE_HOME%/Network/Admin or %ORACLE_HOME%/Net80/Admin (Windows) Note: User-specific Oracle Net parameters may also reside in $HOME/sqlnet.ora file. An Oracle Net server installation is also a client. 2. Oracle Net tracing and logging can consume vast quantities of disk space. Monitor for sufficient disk space when tracing is enabled. On some Unix operating systems, /tmp is used for swap space. Although generally writable by all users, this is not an ideal location for trace/log file generation. 3. Oracle Net tracing should only be enabled for the duration of the issue at hand. Oracle Net tracing should always be disabled after problem resolution. 4. Large trace/log files place an overhead on the processes that generate them. In the absence of issues, the disabling of tracing and/or logging will improve Oracle Net overall efficiency. Alternatively, regularly truncating log files will also improve efficiency. 5. Ensure that the target trace/log directory is writable by the connecting user, Oracle software owner and/or user that starts the Net Listener. LOG & TRACE PARAMETERS ---------------------- This section provides a detailed description of each trace and log parameter. TRACE LEVELS TRACE_LEVEL_[CLIENT|SERVER|LISTENER] = [0-16|USER|ADMIN|SUPPORT|OFF] Determines the degree to which Oracle Net tracing is provided. Configuration file is SQLNET.ORA, LISTENER.ORA. Level 0 is disabled - level 16 is the most verbose tracing level. Listener tracing requires the Net Listener to be reloaded or restarted after adding trace parameters to LISTENER.ORA. Oracle Net (client/server) tracing takes immediate effect after tracing parameters are added to SQLNET.ORA. By default, the trace level is OFF. OFF (equivalent to 0) disabled - provides no tracing. USER (equivalent to 4) traces to identify user-induced error conditions. ADMIN (equivalent to 6) traces to identify installation-specific problems. SUPPORT (equivalent to 16) trace information required by OSS for troubleshooting. TRACE FILE NAME TRACE_FILE_[CLIENT|SERVER|LISTENER] = Determines the trace file name. Any valid operating system file name. Configuration file is SQLNET.ORA, LISTENER.ORA. Trace file is automatically appended with '.TRC'. Default trace file name is SQLNET.TRC, LISTENER.TRC. TRACE DIRECTORY TRACE_DIRECTORY_[CLIENT|SERVER|LISTENER] = Determines the directory in which trace files are written. Any valid operating system directory name. Configuration file is SQLNET.ORA, LISTENER.ORA. Directory should be writable by the connecting user and/or Oracle software owner. Default trace directory is $ORACLE_HOME/network/trace. UNIQUE TRACE FILES TRACE_UNIQUE_[CLIENT|SERVER|LISTENER] = [ON|TRUE|OFF|FALSE] Allows generation of unique trace files per connection. Trace file names are automatically appended with '_.TRC'. Configuration file is SQLNET.ORA, LISTENER.ORA. Unique tracing is ideal for sporadic issues/errors that occur infrequently or randomly. Default value is OFF TRACE TIMING TRACE_TIMESTAMP_[CLIENT|SERVER|LISTENER] = [ON|TRUE|OFF|FALSE] A timestamp in the form of [DD-MON-YY 24HH:MI;SS] is recorded against each operation traced by the trace file. Configuration file is SQLNET.ORA, LISTENER.ORA Suitable for hanging or slow connection issues. Available from Oracle8i onwards. Default value is is OFF. MAXIMUM TRACE FILE LENGTH TRACE_FILELEN_[CLIENT|SERVER|LISTENER] = Determines the maximum trace file size in Kilobytes (Kb). Configuration file is SQLNET.ORA, LISTENER.ORA. Available from Oracle8i onwards. Default value is UNLIMITED. TRACE FILE CYCLING TRACE_FILENO_[CLIENT|SERVER|LISTENER] = Determines the maximum number of trace files through which to perform cyclic tracing. Configuration file is SQLNET.ORA, LISTENER.ORA. Suitable when disk space is limited or when tracing is required to be enabled for long periods. Available from Oracle8i onwards. Default value is 1 (file). LOG FILE NAME LOG_FILE_[CLIENT|SERVER|LISTENER] = Determines the log file name. May be any valid operating system file name. Configuration file is SQLNET.ORA, LISTENER.ORA. Log file is automatically appended with '.LOG'. Default log file name is SQLNET.LOG, LISTENER.LOG. LOG DIRECTORY LOG_DIRECTORY_[CLIENT|SERVER|LISTENER] = Determines the directory in which log files are written. Any valid operating system directory name. Configuration file is SQLNET.ORA, LISTENER.ORA. Directory should be writable by the connecting user or Oracle software owner. Default directory is $ORACLE_HOME/network/log. DISABLING LOGGING LOGGING_LISTENER = [ON|OFF] Disables Listener logging facility. Configuration file is LISTENER.ORA. Default value is ON. ORACLE NET TRACE/LOG EXAMPLES ----------------------------- CLIENT (SQLNET.ORA) trace_level_client = 16 trace_file_client = cli trace_directory_client = /u01/app/oracle/product/9.0.1/network/trace trace_unique_client = on trace_timestamp_client = on trace_filelen_client = 100 trace_fileno_client = 2 log_file_client = cli log_directory_client = /u01/app/oracle/product/9.0.1/network/log tnsping.trace_directory = /u01/app/oracle/product/9.0.1/network/trace tnsping.trace_level = admin SERVER (SQLNET.ORA) trace_level_server = 16 trace_file_server = svr trace_directory_server = /u01/app/oracle/product/9.0.1/network/trace trace_unique_server = on trace_timestamp_server = on trace_filelen_server = 100 trace_fileno_server = 2 log_file_server = svr log_directory_server = /u01/app/oracle/product/9.0.1/network/log namesctl.trace_level = 16 namesctl.trace_file = namesctl namesctl.trace_directory = /u01/app/oracle/product/9.0.1/network/trace namesctl.trace_unique = on LISTENER (LISTENER.ORA) trace_level_listener = 16 trace_file_listener = listener trace_directory_listener = /u01/app/oracle/product/9.0.1/network/trace trace_timestamp_listener = on trace_filelen_listener = 100 trace_fileno_listener = 2 logging_listener = off log_directory_listener = /u01/app/oracle/product/9.0.1/network/log log_file_listener=listener NAMESERVER TRACE (NAMES.ORA) names.trace_level = 16 names.trace_file = names names.trace_directory = /u01/app/oracle/product/9.0.1/network/trace names.trace_unique = off CONNECTION MANAGER TRACE (CMAN.ORA) tracing = yes RELATED DOCUMENTS ----------------- Note 16658.1 (7) Tracing SQL*Net/Net8 Note 111916.1 SQLNET.ORA Logging and Tracing Parameters Note 39774.1 Log & Trace Facilities on Net v2 Note 73988.1 How to Get Cyclic SQL*Net Trace Files when Disk Space is Limited Note 1011114.6 SQL*Net V2 Tracing Note 1030488.6 Net8 Tracing Note 2: ------- Doc ID: Note:39774.1 Subject: LOG & TRACE Facilities on NET v2. Type: FAQ Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 25-JUL-1996 Last Revision Date: 31-JAN-2002 LOG AND TRACE FACILITIES ON SQL*NET V2 ====================================== This article describes the log and trace facilities that can be used to examine application connections that use SQL*Net. This article is based on usage of SQL*NET v2.3. It explains how to invoke the trace facility and how to use the log and trace information to diagnose and resolve operating problems. Following topics are covered below: o What the log facility is o What the trace facility is o How to invoke the trace facility o Logging and tracing parameters o Sample log output o Sample trace output Note: Information in this section is generic to all operating system environments. You may require further information from the Oracle operating system-specific documentation for some details of your specific operating environment. ________________________________________ 1. What is the Log Facility? ============================ All errors encountered in SQL*Net are logged to a log file for evaluation by a network or database administrator. The log file provides additional information for an administrator when the error on the screen is inadequate to understand the failure. The log file, by way of the error stack, shows the state of the TNS software at various layers. The properties of the log file are: o Error information is appended to the log file when an error occurs. o Generally, a log file can only be replaced or erased by an administrator, although client log files can be deleted by the user whose application created them. (Note that in general it is bad practice to delete these files while the program using them is still actively logging.) o Logging of errors for the client, server, and listener cannot be disabled. This is an essential feature that ensures all errors are recorded. o The Navigator and Connection Manager components of the MultiProtocol Interchange may have logging turned on or off. If on, logging includes connection statistics. o The Names server may have logging turned on or off. If on, a Names server's operational events are written to a specified logfile. You set logging parameters using the Oracle Network Manager. ________________________________________ 2. What is the Trace Facility? ============================== The trace facility allows a network or database administrator to obtain more information on the internal operations of the components of a TNS network than is provided in a log file. Tracing an operation produces a detailed sequence of statements that describe the events as they are executed. All trace output is directed to trace output files which can be evaluated after the failure to identify the events that lead up to an error. The trace facility is typically invoked during the occurrence of an abnormal condition, when the log file does not provide a clear indication of the cause. Attention: The trace facility uses a large amount of disk space and may have a significant impact upon system performance. Therefore, you are cautioned to turn the trace facility ON only as part of a diagnostic procedure and to turn it OFF promptly when it is no longer necessary. Components that can be traced using the trace facility are: o Network listener o SQL*Net version 2 components - SQL*Net client - SQL*Net server o MultiProtocol Interchange components - the Connection Manager and pumps - the Navigator o Oracle Names - Names server - Names Control Utility The trace facility can be used to identify the following types of problems: - Difficulties in establishing connections - Abnormal termination of established connections - Fatal errors occurring during the operation of TNS network components ________________________________________ 3. What is the Difference between Logging and Tracing? ====================================================== While logging provides the state of the TNS components at the time of an error, tracing provides a description of all software events as they occur, and therefore provides additional information about events prior to an error. There are three levels of diagnostics, each providing more information than the previous level. The three levels are: 1. The reported error from Oracle7 or tools; this is the single error that is commonly returned to the user. 2. The log file containing the state of TNS at the time of the error. This can often uncover low level errors in interaction with the underlying protocols. 3. The trace file containing English statements describing what the TNS software has done from the time the trace session was initiated until the failure is recreated. When an error occurs, a simple error message is displayed and a log file is generated. Optionally, a trace file can be generated for more information. (Remember, however, that using the trace facility has an impact on your system performance.) In the following example, the user failed to use Oracle Network Manager to create a configuration file, and misspelled the word "PORT" as "POT" in the connect descriptor. It is not important that you understand in detail the contents of each of these results; this example is intended only to provide a comparison. Reported Error (On the screen in SQL*Forms): ERROR: ORA-12533: Unable to open message file (SQL-02113) Logged Error (In the log file, SQLNET.LOG): **************************************************************** Fatal OSN connect error 12533, connecting to: (DESCRIPTION=(CONNECT_DATA=(SID=trace)(CID=(PROGRAM=)(HOST=lala) (USER=ginger)))(ADDRESS_LIST=(ADDRESS=(PROTOCOL=ipc) (KEY=bad_port))(ADDRESS=(PROTOCOL=tcp)(HOST=lala)(POT=1521)))) VERSION INFORMATION: TNS for SunOS: Version 2.0.14.0.0 - Developer's Release Oracle Bequeath NT Protocol Adapter for SunOS: Version 2.0.14.0.0 - Developer's Release Unix Domain Socket IPC NT Protocol Adaptor for SunOS: Version 2.0.14.0.0 - Developer's Release TCP/IP NT Protocol Adapter for SunOS: Version 2.0.14.0.0 - Developer's Release Time: 07-MAY-93 17:38:50 Tracing to file: /home/ginger/trace_admin.trc Tns error struct: nr err code: 12206 TNS-12206: TNS:received a TNS error while doing navigation ns main err code: 12533 TNS-12533: TNS:illegal ADDRESS parameters ns secondary err code: 12560 nt main err code: 503 TNS-00503: Illegal ADDRESS parameters nt secondary err code: 0 nt OS err code: 0 Example of Trace of Error ------------------------- The trace file, SQLNET.TRC at the USER level, contains the following information: --- TRACE CONFIGURATION INFORMATION FOLLOWS --- New trace stream is "/private1/oracle/trace_user.trc" New trace level is 4 --- TRACE CONFIGURATION INFORMATION ENDS --- --- PARAMETER SOURCE INFORMATION FOLLOWS --- Attempted load of system pfile source /private1/oracle/network/admin/sqlnet.ora Parameter source was not loaded Error stack follows: NL-00405: cannot open parameter file Attempted load of local pfile source /home/ginger/.sqlnet.ora Parameter source loaded successfully -> PARAMETER TABLE LOAD RESULTS FOLLOW <- Some parameters may not have been loaded See dump for parameters which loaded OK -> PARAMETER TABLE HAS THE FOLLOWING CONTENTS <- TRACE_DIRECTORY_CLIENT = /private1/oracle trace_level_client = USER TRACE_FILE_CLIENT = trace_user --- PARAMETER SOURCE INFORMATION ENDS --- --- LOG CONFIGURATION INFORMATION FOLLOWS --- Attempted open of log stream "/tmp_mnt/home/ginger/sqlnet.log" Successful stream open --- LOG CONFIGURATION INFORMATION ENDS --- Unable to get data from navigation file tnsnav.ora local names file is /home/ginger/.tnsnames.ora system names file is /etc/tnsnames.ora -- failure, error stack follows -- NL-00427: bad list -- NOTE: FILE CONTAINS ERRORS, SOME NAMES MAY BE MISSING Calling address: (DESCRIPTION=(CONNECT_DATA=(SID=trace)(CID=(PROGRAM=)(HOST=lala)(USER=ginger))) (ADDRESS_LIST=(ADDRESS=(PROTOCOL=ipc)(KEY=bad_port))(ADDRESS=(PROTOCOL=tcp)(HOST Getting local community information Looking for local addresses setup by nrigla No addresses in the preferred address list TNSNAV.ORA is not present. No local communities entry. Getting local address information Address list being processed... No community information so all addresses are "local" Resolving address to use to call destination or next hop Processing address list... No community entries so iterate over address list This a local community access Got routable address information Making call with following address information: (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=ipc)(KEY=bad_port))) Calling with outgoing connect data (DESCRIPTION=(CONNECT_DATA=(SID=trace)(CID=(PROGRAM=)(HOST=lala)(USER=ginger))) (ADDRESS_LIST=(ADDRESS=(PROTOCOL=tcp)(HOST=lala)(POT=1521)))) (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=ipc)(KEY=bad_port))) KEY = bad_port connecting... opening transport... -- sd=8, op=1, resnt[0]=511, resnt[1]=2, resnt[2]=0 -- unable to open transport -- nsres: id=0, op=1, ns=12541, ns2=12560; nt[0]=511, nt[1]=2, nt[2]=0 connect attempt failed Call failed... Call made to destination Processing address list so continuing Getting local community information Looking for local addresses setup by nrigla No addresses in the preferred address list TNSNAV.ORA is not present. No local communities entry. Getting local address information Address list being processed... No community information so all addresses are "local" Resolving address to use to call destination or next hop Processing address list... No community entries so iterate over address list This a local community access Got routable address information Making call with following address information: (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=tcp)(HOST=lala)(POT=1521))) Calling with outgoing connect data (DESCRIPTION=(CONNECT_DATA=(SID=trace)(CID=(PROGRAM=)(HOST=lala)(USER=ginger))) (ADDRESS_LIST=(ADDRESS=(PROTOCOL=tcp)(HOST=lala)(POT=521)))) (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=tcp)(HOST=lala)(POT=1521))) -- failed to recognize: POT -- nsres: id=0, op=13, ns=12533, ns2=12560; nt[0]=503, nt[1]=0, nt[2]=0 Call failed... Exiting NRICALL with following termination result -1 -- error from nricall -- nr err code: 12206 -- ns main err code: 12533 -- ns (2) err code: 12560 -- nt main err code: 503 -- nt (2) err code: 0 -- nt OS err code: 0 -- Couldn't connect, returning 12533 In the trace file, note that unexpected events are preceded with an -- stamp. These events may represent serious errors, minor errors, or merely unexpected results from an internal operation. More serious and probably fatal errors are stamped with the -- prefix. In this example trace file, you can see that the root problem, the misspelling of "PORT," is indicated by the trace line: -- failed to recognize: POT Most tracing is very similar to this. If you have a basic understanding of the events the components will perform, you can identify the probable cause of an error in the text of the trace. ________________________________________ 4. Log File Names ================= Log files produced by different components have unique names. The default file names are: SQLNET.LOG Contains client and/or server information LISTENER.LOG Contains listener information INTCHG.LOG Contains Connection Manager and pump information NAVGATR.LOG Contains Navigator information NAMES.LOG Contains Names server information You can control the name of the log file. For each component, any valid string can be used to create a log file name. The parameters are of the form: LOG_FILE_component = string For example: LOG_FILE_LISTENER = TEST Some platforms have restrictions on the properties of a file name. See your Oracle operating system specific manuals for platform specific restrictions. _____________________________________ 5. Using Log Files ================== Follow these steps to track an error using a log file: 1. Browse the log file for the most recent error that matches the error number you have received from the application. This is almost always the last entry in the log file. Notice that an entry or error stack in the log file is usually many lines in length. In the example earlier in this chapter, the error number was 12207. 2. Starting at the bottom, look up to the first non-zero entry in the error report. This is usually the actual cause. In the example earlier in this chapter, the last non-zero entry is the "ns" error 12560. 3. Look up the first non-zero entry in later chapters of this book for its recommended cause and action. (For example, you would find the "ns" error 12560 under ORA-12560.) To understand the notation used in the error report, see the previous chapter, "Interpreting Error Messages." 4. If that error does not provide the desired information, move up the error stack to the second to last error and so on. 5. If the cause of the error is still not clear, turn on tracing and re-execute the statement that produced the error message. The use of the trace facility is described in detail later in this chapter. Be sure to turn tracing off after you have re-executed the command. ________________________________________ 6. Using the Trace Facility =========================== The steps used to invoke tracing are outlined here. Each step is fully described in subsequent sections. 1. Choose the component to be traced from the list: o Client o Server o Listener o Connection Manager and pump (cmanager) o Navigator (navigator) o Names server o Names Control Utility 2. Save existing trace file if you need to retain information on it. By default most trace files will overwrite an existing ones. TRACE_UNIQUE parameter needs to be included in appropriate config. files if unique trace files are required. This appends Process Id to each file. For Example: For Names server tracing, NAMES.TRACE_UNIQUE=ON needs to be set in NAMES. ORA file. For Names Control Utility, NAMESCTL.TRACING_UNIQUE=TRUE needs to be in SQLNET.ORA. TRACE_UNIQUE_CLIENT=ON in SQLNET.ORA for Client Tracing. 3. For any component, you can invoke the trace facility by editing the component configuration file that corresponds to the component traced. The component config. files are SQLNET.ORA, LISTENER.ORA, INTCHG.ORA, and NAMES. ORA. 4. Execute or start the component to be traced. If the trace component configuration files are modified while the component is running, the modified trace parameters will take effect the next time the component is invoked or restarted. Specifically for each component: CLIENT: Set the trace parameters in the client-side SQLNET.ORA and invoke a client application, such as SQL*Plus, a Pro*C application, or any application that uses the Oracle network products. SERVER: Set the trace parameters in the server-side SQLNET.ORA. The next process started by the listener will have tracing enabled. The trace parameters must be created or edited manually. LISTENER: Set the trace parameters in the LISTENER.ORA CONNECTION MANAGER: Set the trace parameters in INTCHG.ORA and start the Connection Manager from the Interchange Control Utility or command line. The pumps are started automatically with the Connection Manager, and their trace files are controlled by the trace parameters for the Connection Manager. NAVIGATOR:Again, set the trace parameters in INTCHG.ORA and start the Navigator NAMES SERVER: Trace parameters needs to be set in NAMES.ORA and start the Names server. NAMES CONTROL UTILITY: Set the trace parameters in SQLNET.ORA and start the Names Control Utility 5. Be sure to turn tracing off when you do not need it for a specific diagnostic purpose. ________________________________________ 7. Setting Trace Parameters =========================== The trace parameters are defined in the same configuration files as the log parameters. Table below shows the configuration files for different network components and the default names of the trace files they generate. -------------------------------------------------------- | Trace Parameters | Configuration | | | Corresponding to | File | Output Files | |-------------------|-----------------|------------------| | | | | | Client | SQLNET.ORA | SQLNET.TRC | | Server | | SQLNET.TRC | | TNSPING Utility | | TNSPING.TRC | | Names Control | | | | Utility | | NAMESCTL.TRC | |-------------------|-----------------|------------------| | Listener | LISTENER.ORA | LISTENER.TRC | |-------------------|-----------------|------------------| | Interchange | INTCHG.ORA | | | Connection | | | | Manager | | CMG.TRC | | Pumps | | PMP.TRC | | Navigator | | NAV.TRC | |-------------------|-----------------|------------------| | Names server | NAMES.ORA | NAMES.TRC | |___________________|_________________|__________________| The configuration files for each component are located on the computer running that component. The trace characteristics for two or more components of an Interchange are controlled by different parameters in the same configuration file. For example, there are separate sets of parameters for the Connection Manager and the Navigator that determine which components will be traced, and at what level. Similarly, if there are multiple listeners on a single computer, each listener is controlled by parameters that include the unique listener name in the LISTENER.ORA file. For each component, the configuration files contain the following information: o A valid trace level to be used (Default is OFF) o The trace file name (optional) o The trace file directory (optional) ________________________________________ 7a. Valid SQLNET.ORA Diagnostic Parameters ========================================== The SQLNET.ORA caters for: o Client Logging & Tracing o Server Logging & Tracing o TNSPING utility o NAMESCTL program ------------------------------------------------------------------------------ | | | | | PARAMETERS | VALUES | Example (DOS client, UNIX server) | | | | | |------------------------|----------------|------------------------------------| |Parameters for Client | |===================== | |------------------------------------------------------------------------------| | | | | | TRACE_LEVEL_CLIENT | OFF/USER/ADMIN | TRACE_LEVEL_CLIENT=USER | | | | | | TRACE_FILE_CLIENT | string | TRACE_FILE_CLIENT=CLIENT | | | | | | TRACE_DIRECTORY_CLIENT | valid directory| TRACE_DIRECTORY_CLIENT=c:\NET\ADMIN| | | | | | TRACE_UNIQUE_CLIENT | OFF/ON | TRACE_UNIQUE_CLIENT=ON | | | | | | LOG_FILE_CLIENT | string | LOG_FILE_CLIENT=CLIENT | | | | | | LOG_DIRECTORY_CLIENT | valid directory| LOG_DIRECTORY_CLIENT=c:\NET\ADMIN | |------------------------------------------------------------------------------| |Parameters for Server | |===================== | |------------------------------------------------------------------------------| | | | | | TRACE_LEVEL_SERVER | OFF/USER/ADMIN | TRACE_LEVEL_SERVER=ADMIN | | | | | | TRACE_FILE_SERVER | string | TRACE_FILE_SERVER=unixsrv_2345.trc | | | | | | TRACE_DIRECTORY_SERVER | valid directory| TRACE_DIRECTORY_SERVER=/tmp/trace | | | | | | LOG_FILE_SERVER | string | LOG_FILE_SERVER=unixsrv.log | | | | | | LOG_DIRECTORY_SERVER | valid directory| LOG_DIRECTORY_SERVER=/tmp/trace | |------------------------------------------------------------------------------| ---(SQLNET.ORA Cont.)--------------------------------------------------------- | | | | | PARAMETERS | VALUES | Example (DOS client, UNIX server) | | | | | |------------------------|----------------|------------------------------------| | |Parameters for TNSPING | |====================== | |------------------------------------------------------------------------------| | | | | | TNSPING.TRACE_LEVEL | OFF/USER/ADMIN | TNSPING.TRACE_LEVEL=user | | | | | | TNSPING.TRACE_DIRECTORY| directory |TNSPING.TRACE_DIRECTORY= | | | | /oracle7/network/trace | | | | | |------------------------------------------------------------------------------| |Parameters for Names Control Utility | |==================================== | |------------------------------------------------------------------------------| | | | | | NAMESCTL.TRACE_LEVEL | OFF/USER/ADMIN |NAMESCTL.TRACE_LEVEL=user | | | | | | NAMESCTL.TRACE_FILE | file |NAMESCTL.TRACE_FILE=nc_south.trc | | | | | | NAMESCTL.TRACE_DIRECTORY| directory |NAMESCTL.TRACE_DIRECTORY=/o7/net/trace| | | | | | NAMESCTL.TRACE_UNIQUE | TRUE/FALSE |NAMESCTL.TRACE_UNIQUE=TRUE or ON/OFF| | | | | ------------------------------------------------------------------------------ Note: You control log and trace parameters for the client through Oracle Network Manager. You control log and trace parameters for the server by manually adding the desired parameters to the SQLNET.ORA file. Parameters for Names Control Utility & TNSPING Utility need to be added manually to SQLNET.ORA file. You cannot create them using Oracle Network Manager. ________________________________________ 7b. Valid LISTENER.ORA Diagnostic Parameters ============================================ The following table shows the valid LISTENER.ORA parameters used in logging and tracing of the listener. ------------------------------------------------------------------------------ | | | | | PARAMETERS | VALUES | Example (DOS client, UNIX server) | | | | | |------------------------|----------------|------------------------------------| | | | | |TRACE_LEVEL_LISTENER | USER | TRACE_LEVEL_LISTENER=OFF | | | | | |TRACE_FILE_LISTENER | string | TRACE_FILE_LISTENER=LISTENER | | | | | |TRACE_DIRECTORY_LISTENER| valid directory| TRACE_DIRECTORY_LISTENER=$ORA_SQLNETV2 | | | | | |LOG_FILE_LISTENER | string | LOG_FILE_LISTENER=LISTENER | | | | | |LOG_DIRECTORY_LISTENER | valid directory| LOG_DIRECTORY_LISTENER=$ORA_ERRORS | | | | | ------------------------------------------------------------------------------ ________________________________________ 7c. Valid INTCHG.ORA Diagnostic Parameters ========================================== The following table shows the valid INTCHG.ORA parameters used in logging and tracing of the Interchange. ---------------------------------------------------------------------------------- | | | | | PARAMETERS | VALUES | Example (DOS client, UNIX server) | | | (default)| | |------------------------|--------------------|------------------------------------| | | | | |TRACE_LEVEL_CMANAGER | OFF|USER|ADMIN | TRACE_LEVEL_CMANAGER=USER | | | | | |TRACE_FILE_CMANAGER | string (CMG.TRC) | TRACE_FILE_CMANAGER=CMANAGER | | | | | |TRACE_DIRECTORY_CMANAGER| valid directory | TRACE_DIRECTORY_CMANAGER=C:\ADMIN | | | | | |LOG_FILE_CMANAGER | string (INTCHG.LOG)| LOG_FILE_CMANAGER=CMANAGER | | | | | |LOG_DIRECTORY_CMANAGER | valid directory | LOG_DIRECTORY_CMANAGER=C:\ADMIN | | | | | |LOGGING_CMANAGER | OFF/ON | LOGGING_CMANAGER=ON | | | | | |LOG_INTERVAL_CMANAGER | Any no of minutes | LOG_INTERVAL_CMANAGER=60 | | | (60 minutes)| | |TRACE_LEVEL_NAVIGATOR | OFF/USER/ADMIN | TRACE_LEVEL_NAVIGATOR=ADMIN | | | | | |TRACE_FILE_NAVIGATOR | string (NAV.TRC)| TRACE_FILE_NAVIGATOR=NAVIGATOR | | | | | |TRACE_DIRECTORY_NAVIGATOR| valid directory | TRACE_DIRECTORY_NAVIGATOR=C:\ADMIN | | | | | |LOG_FILE_NAVIGATOR |string (NAVGATR.LOG)| LOG_FILE_NAVIGATOR=NAVIGATOR | | | | | |LOG_DIRECTORY_NAVIGATOR | valid directory | LOG_DIRECTORY_NAVIGATOR=C:\ADMIN | | | | | |LOGGING_NAVIGATOR | OFF/ON | LOGGING_NAVIGATOR=OFF | | | | | |LOG_LEVEL_NAVIGATOR | ERRORS|ALL (ERRORS)| LOG_LEVEL_NAVIGATOR=ERRORS | | | | | ---------------------------------------------------------------------------------- Note: The pump component shares the trace parameters of the Connection Manager, but it generates a separate trace file with the unchangeable default name PMPpid.TRC. ________________________________________ 7d. Valid NAMES.ORA Diagnostic Parameters ========================================= The following table shows the valid NAMES.ORA parameters used in logging and tracing of the Names server. ------------------------------------------------------------------------------ | | | | | PARAMETERS | VALUES | Example (DOS client, UNIX server) | | | (default)| | |------------------------|----------------|------------------------------------| | | | | | NAMES.TRACE_LEVEL | OFF/USER/ADMIN | NAMES.TRACE_LEVEL=ADMIN | | | | | | NAMES.TRACE_FILE | file(names.trc)| NAMES.TRACE_FILE=nsrv3.trc | | | | | | NAMES.TRACE_DIRECTORY | directory | NAMES.TRACE_DIRECTORY=/o7/net/trace| | | | | | NAMES.TRACE_UNIQUE | TRUE/FALSE | NAMES.TRACE_UNIQUE=TRUE or ON/OFF | | | | | | NAMES.LOG_FILE | file(names.log)| NAMES.LOG_FILE=nsrv1.log | | | | | | NAMES.LOG_DIRECTORY | directory | NAMES.LOG_DIRECTORY= /o7/net/log | | | | | ------------------------------------------------------------------------------ ________________________________________ 8. Example of a Trace File =========================== In the following example, the SQLNET.ORA file includes the following line: TRACE_LEVEL_CLIENT = ADMIN The following trace file is the result of a connection attempt that failed because the hostname is invalid. The trace output is a combination of debugging aids for Oracle specialists and English information for network administrators. Several key events can be seen by analyzing this output from beginning to end: (A) The client describes the outgoing data in the connect descriptor used to contact the server. (B) An event is received (connection request). (C) A connection is established over the available transport (in this case TCP/IP). (D) The connection is refused by the application, which is the listener. (E) The trace file shows the problem, as follows: -- ***hostname lookup failure! *** (F) Error 12545 is reported back to the client. If you look up Error 12545 in Chapter 3 of this Manual, you will find the following description: ORA-12545 TNS:Name lookup failure Cause: A protocol specific ADDRESS parameter cannot be resolved. Action: Ensure the ADDRESS parameters have been entered correctly; the most likely incorrect value is the node name. ++++++ NOTE: TRACE FILE EXTRACT +++++++ --- TRACE CONFIGURATION INFORMATION FOLLOWS --- New trace stream is "/private1/oracle/trace_admin.trc" New trace level is 6 --- TRACE CONFIGURATION INFORMATION ENDS --- ++++++ NOTE: Loading Parameter files now. +++++++ --- PARAMETER SOURCE INFORMATION FOLLOWS --- Attempted load of system pfile source /private1/oracle/network/admin/sqlnet.ora Parameter source was not loaded Error stack follows: NL-00405: cannot open parameter file Attempted load of local pfile source /home/ginger/.sqlnet.ora Parameter source loaded successfully -> PARAMETER TABLE LOAD RESULTS FOLLOW <- Some parameters may not have been loaded See dump for parameters which loaded OK -> PARAMETER TABLE HAS THE FOLLOWING CONTENTS <- TRACE_DIRECTORY_CLIENT = /private1/oracle trace_level_client = ADMIN TRACE_FILE_CLIENT = trace_admin --- PARAMETER SOURCE INFORMATION ENDS --- ++++++ NOTE: Reading Parameter files. +++++++ --- LOG CONFIGURATION INFORMATION FOLLOWS --- Attempted open of log stream "/private1/oracle/sqlnet.log" Successful stream open --- LOG CONFIGURATION INFORMATION ENDS --- Unable to get data from navigation file tnsnav.ora local names file is /home/ginger/.tnsnames.ora system names file is /etc/tnsnames.ora initial retry timeout for all servers is 500 csecs max request retries per server is 2 default zone is [root] Using nncin2a() to build connect descriptor for (possibly remote) database. initial load of /home/ginger/.tnsnames.ora -- failure, error stack follows -- NL-00405: cannot open parameter file -- NOTE: FILE CONTAINS ERRORS, SOME NAMES MAY BE MISSING initial load of /etc/tnsnames.ora -- failure, error stack follows -- NL-00427: bad list -- NOTE: FILE CONTAINS ERRORS, SOME NAMES MAY BE MISSING Inserting IPC address into connect descriptor returned from nncin2a(). ++++++ NOTE: Looking for Routing Information. +++++++ Calling address: (DESCRIPTION=(CONNECT_DATA=(SID=trace)(CID=(PROGRAM=)(HOST=lala) (USER=ginger)))(ADDRESS_LIST=(ADDRESS=(PROTOCOL=ipc (KEY=bad_host))(ADDRESS=(PROTOCOL=tcp)(HOST=lavender) (PORT=1521)))) Getting local community information Looking for local addresses setup by nrigla No addresses in the preferred address list TNSNAV.ORA is not present. No local communities entry. Getting local address information Address list being processed... No community information so all addresses are "local" Resolving address to use to call destination or next hop Processing address list... No community entries so iterate over address list This a local community access Got routable address information ++++++ NOTE: Calling first address (IPC). +++++++ Making call with following address information: (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=ipc)(KEY=bad_host))) Calling with outgoing connect data (DESCRIPTION=(CONNECT_DATA=(SID=trace)(CID=(PROGRAM=)(HOST=lala) (USER=ginger)))(ADDRESS_LIST=(ADDRESS=(PROTOCOL=tcp) (HOST=lavender)(PORT=1521)))) (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=ipc)(KEY=bad_host))) KEY = bad_host connecting... opening transport... -- sd=8, op=1, resnt[0]=511, resnt[1]=2, resnt[2]=0 -- unable to open transport -- nsres: id=0, op=1, ns=12541, ns2=12560; nt[0]=511, nt[1]=2, nt[2]=0 connect attempt failed Call failed... Call made to destination Processing address list so continuing ++++++ NOTE: Looking for Routing Information. +++++++ Getting local community information Looking for local addresses setup by nrigla No addresses in the preferred address list TNSNAV.ORA is not present. No local communities entry. Getting local address information Address list being processed... No community information so all addresses are "local" Resolving address to use to call destination or next hop Processing address list... No community entries so iterate over address list This a local community access Got routable address information ++++++ NOTE: Calling second address (TCP/IP). +++++++ Making call with following address information: (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=tcp) (HOST=lavender)(PORT=1521))) Calling with outgoing connect data (DESCRIPTION=(CONNECT_DATA=(SID=trace)(CID=(PROGRAM=)(HOST=lala) (USER=ginger)))(ADDRESS_LIST=(ADDRESS=(PROTOCOL=tcp) (HOST=lavender) (PORT=1521)))) (DESCRIPTION=(EMPTY=0)(ADDRESS=(PROTOCOL=tcp) (HOST=lavender)(PORT=1521))) port resolved to 1521 looking up IP addr for host: lavender -- *** hostname lookup failure! *** -- nsres: id=0, op=13, ns=12545, ns2=12560; nt[0]=515, nt[1]=0, nt[2]=0 Call failed... Exiting NRICALL with following termination result -1 -- error from nricall -- nr err code: 12206 -- ns main err code: 12545 -- ns (2) err code: 12560 -- nt main err code: 515 -- nt (2) err code: 0 -- nt OS err code: 0 -- Couldn't connect, returning 12545 Most tracing is very similar to this. If you have a basic understanding of the events the components will perform, you can identify the probable cause of an error in the text of the trace. 19.64 ORA-01595: error freeing extent (2) of rollback segment (9)): =================================================================== Note 1: ORA-01595, 00000, "error freeing extent (%s) of rollback segment (%s))" Cause: Some error occurred while freeing inactive rollback segment extents. Action: Investigate the accompanying error. Note 2: Two factors are necessary for this to happen. A rollback segment has extended beyond OPTIMAL. There are two or more transactions sharing the rollback segment at the time of the shrink. What happens is that the first process gets to the end of an extent, notices the need to shrink and begins the recursive transaction to do so. But the next transaction blunders past the end of that extent before the recursive transaction has been committed. The preferred solution is to have sufficient rollback segments to eliminate the sharing of rollback segments between processes. Look in V$RESOURCE_LIMIT for the high-water-mark of transactions. That is the number of rollback segments you need. The alternative solution is to raise OPTIMAL to reduce the risk of the error. Note 3: This error is harmless. You can try (and probably should) set optimal to null and maxextents to unlimited (which might minimize the frequency of these errors). These errors happen sometimes when oracle is shrinking the rollback segments upto the optimal size. The undo data for shrinking is also kept in the rollback segments. So when it attempts to shrink the same rollback segment where its trying to write the undo, it throws this warning. Its not a failure per se .. since oracle will retry and succeed. 19.65: OUI-10022: oraInventory cannot be used because it is in an invalid state =============================================================================== Note 1: ------- If there are other products installed through the OUI, create a copy of = the oraInst.loc file (depending on the UNIX system, possibly in /etc or /var/opt/oracle). Modify the inventory_loc parameter to point to a different location for = the OUI to create the oraInventory directory. Run the installer using the -invPtrLoc parameter (eg: runInstaller -invPtrLoc /PATH/oraInst.loc). This will retain the existing oraInventory directory and create a new = one for use by the new product. 19.66: Failure to extend rollback segment because of 30036 condition ==================================================================== Not a serious problem. Do some undo tuning. 19.67: ORA-06502: PL/SQL: numeric or value error: character string buffer too small =================================================================================== Note 1: Hi, I am having a strange problem with an ORA-06502 error I am getting and don't understand why. I would expect this error to be quite easy to fix, it would suggest that a variable is not large enough to cope with a value being assigned to it. But I'm fairly sure that isn't the problem. Anyway I have a stored procedure similar to the following: PROCEDURE myproc(a_user IN VARCHAR2, p_1 OUT .%TYPE, p_2 OUT .%TYPE) IS BEGIN SELECT my_first_column, my_second_column INTO p_1, p_2 FROM my_table WHERE user_id = a_user; END; / The procedure is larger than this, but using error_position variables I have tracked it down to one SQL statement. But I don't understand why I'm getting the ORA-06502, because the variables I am selecting into are defined as the same types as the columns I'm selecting. The variable I am selecting into is in fact a VARCHAR2(4), but if I replace the sql statement with p_1 := 'AB'; it still fails. It succeeds if I do p_1 := 'A'; Has anyone seen this before or anything similar that they might be able to help me with please? Thanks, mtae. -- Answer 1: It is the code from which you are calling it that has the problem, e.g. DECLARE v1 varchar2(1); v2 varchar2(1); BEGIN my_proc ('USER',v1,v2); END; / -- Answer 2 try this: PROCEDURE myproc(a_user IN VARCHAR2, p_1 OUT varchar2, p_2 OUT varchar2) IS v_1 .%TYPE; v_2 .%TYPE; BEGIN SELECT my_first_column, my_second_column INTO v_1, v_2 FROM my_table WHERE user_id = a_user; p_1 := v_1; p_2 := v_2; END; / Comment from mtae Date: 07/28/2004 04:24AM PDT Author Comment It was the size of the variable that was being used as the actual parameter being passed in. Feeling very silly, but thanks, sometimes you can look at a problem too long. 19.68 ORA-00600: internal error code, arguments: [LibraryCacheNotEmptyOnClose], [], [], [], [], [], [], [] ========================================================================================================== thread: see this error every time I shutdown a 10gR3 grid control database on 10.2.0.3 RDBMS, even though all opmn and OMS processes are down. So far, I have not seen any problems, apart from the annoying shutdown warning. Note 365103.1 seems to indicate it can be ignored: Cause This is due to unpublished Bug 4483084 'ORA-600 [LIBRARYCACHENOTEMPTYONCLOSE]' This is a bug in that an ORA-600 error is reported when it is found that something is still going on during shutdown. It does not indicate any damage or a problem in the system. Solution At the time of writing, it is likely that the fix will be to report a more meaningful external error, although this has not been finalised. The error is harmless so it is unlikely that this will be backported to 10.2. The error can be safely ignored as it does not indicate a problem with the database. thread: ORA-00600: internal error code, arguments: [LibraryCacheNotEmptyOnClose], [],[], [], [], [], [], [] 14-DEC-06 05:15:35 GMT Hi, There is no patch available for the bug 4483084. You need to Ignore this error, as there is absolutely no impact to the database due to this error. Thanks, Ram thread: 19.69: ORA-12518 Tns: Listener could not hand off: -------------------------------------------------- >>>> thread 1: Q: ORA-12518 Tns: Listener could not hand off client conenction Posted: May 31, 2007 2:02 AM Reply Dear exeprts, Plz tell me how can I resolve ORA-12518 Tns: Listener could not hand off client conenction. ORA-12518: TNS:listener could not hand off client connection A: Your server is probably running out of memory and need to swap memory to disk. One cause can be an Oracle process consuming too much memory. A possible workaround is to set following parameter in the listener.ora and restart the listener: DIRECT_HANDOFF_TTC_LISTENER=OFF You might need to increase the value of large_pool_size. Regards. >>>> thread 2: Q: Hi All, I'm using oracle 10g in window XP system. Java programmers will be accessing the database. Frequently they will get "ORA-12518: TNS:listener could not hand off" error and through sqlplus also i'll get this error. But, after sometime it works fine. I checked tnsnames.ora and listner.ora files entry. they seems to be ok. i have used system name itself for HOST flag instead of IP address. But still i'm getting this error. Can anybody tell me what might be the problem? Thanks, A: From Oracle's error messages docco, we see -------- TNS-12518 TNS:listener could not hand off client connection Cause: The process of handing off a client connection to another process failed. Action: Turn on listener tracing and re-execute the operation. Verify that the listener and database instance are properly configured for direct handoff. If the problem persists, contact Oracle Support Services. -------- So what does the listener trace indicate? A: Did you by any chance upgrade with SP2? If so, you could be running into firewall problems - 1521 is open, the initial contact made, but the handoff to a random (blocked!) port fails... -- Regards, Frank van Bortel >>>> thread 3: Q: I install Oracle9i and Oracle8i on Win2000 Server. I used Listener of 9i. My database based on Oracle8i. I found this error:ORA-12518: TNS:listener could not hand off client connectionwhen I logged on database. If I restarted database and listener it run but a few minutes it failed. Can u help me??? A: Are you usting MTS? First start the listener and then the database ( both the databases). Now check the status of listener. if nothing works, try DIRECT_HANDOFF_TTC_ = OFF in listener.ora. >>>> thread 4 Q: This weekend I installed Oracle Enterprise 10g release 2 on Windows 2003 server. The server is a Xeon dual processor 2.5MHz each with 3GB RAM and 300GB harddisk on RAID 1. The installation was fine, I then installed our application on it, that went smoothly as well. I had 3 users logged in to test the installation and everything was ok. Today morning we had 100 users trying to login and some got access, but majority got the ORA error above and have no access. I checked the tnsnames.ora file and sqlnet.ora file, service on the database all looks ok. I also restarted the listener service on the server, but I still get this error message. I've also increased no of sessions to 1000. Has anyone ever come across a issue like this in Oracle 10g. Regards A: I think I've resolved the problem, majority of my users are away on easter break so when they return I will know whether this tweak has paid off or not. Basically my SGA settings were quite high, so 60% of RAM was being used by SGA and 40% by Windows. I basically reduced the total SGA to 800 MB and i've had no connection problems, ever since. >>>> thread 5 ORA-12518: TNS:listener could not hand off client connection Your server is probably running out of memory and need to swap memory to disk. One cause can be an Oracle process consuming too much memory. A possible workaround is to set following parameter in the listener.ora and restart the listener: DIRECT_HANDOFF_TTC_LISTENER=OFF Should you be working with Multi threaded server connections, you might need to increase the value of large_pool_size. 19.70: Private strand flush not complete: ----------------------------------------- -- thread: Q: I just upgraded to Oracle 10g release 2 and I keep getting this error in my alert log Thread 1 cannot allocate new log, sequence 509 Private strand flush not complete Current log# 2 seq# 508 mem# 0: /usr/local/o1_mf_2_2cx5wnw5_.log Current log# 2 seq# 508 mem# 1: /usr/local/o1_mf_2_2cx5wrjk_.log What causes the "private strand flush not complete" message? A: This is not a bug, it's the expected behavior in 10gr2. The "private strand flush not complete" is a "noise" error, and can be disregarded because it relates to internal cache redo file management. Oracle Metalink note 372557.1 says that a "strand" is a new 10gr2 term for redo latches. It notes that a strand is a new mechanism to assign redo latches to multiple processes, and it's related to the log_parallelism parameter. The note says that the number of strands depends on the cpu_count. When you switch redo logs you will see this alert log message since all private strands have to be flushed to the current redo log. -- thread: Q: HI, I'm using the Oracle 10g R2 in a server with Red Hat ES 4.0, and i received the following message in alert log "Private strand flush not complete", somebody knows this error? The part of log, where I found this error is: Fri Feb 10 10:30:52 2006 Thread 1 advanced to log sequence 5415 Current log# 8 seq# 5415 mem# 0: /db/oradata/bioprd/redo081.log Current log# 8 seq# 5415 mem# 1: /u02/oradata/bioprd/redo082.log Fri Feb 10 10:31:21 2006 Thread 1 cannot allocate new log, sequence 5416 Private strand flush not complete Current log# 8 seq# 5415 mem# 0: /db/oradata/bioprd/redo081.log Current log# 8 seq# 5415 mem# 1: /u02/oradata/bioprd/redo082.log Thread 1 advanced to log sequence 5416 Current log# 13 seq# 5416 mem# 0: /db/oradata/bioprd/redo131.log Current log# 13 seq# 5416 mem# 1: /u02/oradata/bioprd/redo132.log Thanks, A: Hi, Note:372557.1 has brief explanation of this message. Best Regards, -- thread: Q: Hi, I`m having such info in alert_logfile... maybee some ideas or info... Private strand flush not complete What could this posible mean ?? Thu Feb 9 22:03:44 2006 Thread 1 cannot allocate new log, sequence 387 Private strand flush not complete Current log# 2 seq# 386 mem# 0: /path/redo02.log Thread 1 advanced to log sequence 387 Current log# 3 seq# 387 mem# 0: /path/redo03.log Thanks A: see http://download-uk.oracle.com/docs/cd/B19306_01/server.102/b14237/waitevents003.htm#sthref4478 regards log file switch (private strand flush incomplete) User sessions trying to generate redo, wait on this event when LGWR waits for DBWR to complete flushing redo from IMU buffers into the log buffer; when DBWR is complete LGWR can then finish writing the current log, and then switch log files. Wait Time: 1 second Parameters: None Error message :Thread 1 cannot allocate new log ----------------------------------------------- Note 1: ------- Q: Hi Iam getting error message "Thread 1 cannot allocate new log", sequence40994 can any one help me out , how to overcome this problem. Give me a solution. regards A: Perhaps this will provide some guidance. Rick Sometimes, you can see in your alert.log file, the following corresponding messages: Thread 1 advanced to log sequence 248 Current log# 2 seq# 248 mem# 0: /prod1/oradata/logs/redologs02.log Thread 1 cannot allocate new log, sequence 249 Checkpoint not complete This message indicates that Oracle wants to reuse a redo log file, but the corresponding checkpoint associated is not terminated. In this case, Oracle must wait until the checkpoint is completely realized. This situation may be encountered particularly when the transactional activity is important. This situation may also be checked by tracing two statistics in the BSTAT/ESTAT report.txt file. The two statistics are: - Background checkpoint started. - Background checkpoint completed. These two statistics must not be different more than once. If this is not true, your database hangs on checkpoints. LGWR is unable to continue writing the next transactions until the checkpoints complete. Three reasons may explain this difference: - A frequency of checkpoints which is too high. - A checkpoints are starting but not completing - A DBWR which writes too slowly. The number of checkpoints completed and started as indicated by these statistics should be weighed against the duration of the bstat/estat report. Keep in mind the goal of only one log switch per hour, which ideally should equate to one checkpoint per hour as well. The way to resolve incomplete checkpoints is through tuning checkpoints and logs: 1) Give the checkpoint process more time to cycle through the logs - add more redo log groups - increase the size of the redo logs 2) Reduce the frequency of checkpoints - increase LOG_CHECKPOINT_INTERVAL - increase size of online redo logs 3) Improve the efficiency of checkpoints enabling the CKPT process with CHECKPOINT_PROCESS=TRUE 4) Set LOG_CHECKPOINT_TIMEOUT = 0. This disables the checkpointing based on time interval. 5) Another means of solving this error is for DBWR to quickly write the dirty buffers on disk. The parameter linked to this task is: DB_BLOCK_CHECKPOINT_BATCH. DB_BLOCK_CHECKPOINT_BATCH specifies the number of blocks which are dedicated inside the batch size for writing checkpoints. When you want to accelerate the checkpoints, it is necessary to increase this value. Note 2: ------- Q: Hi All, Lets generate a good discussion thread for this database performance issue. Sometimes this message is found in the alert log generated. Thread 1 advanced to log sequence xxx Current log# 2 seq# 248 mem# 0: /df/sdfds Thread 1 cannot allocate new log, sequence xxx Checkpoint not complete I would appreciate a discussion on the following 1. What are the basic reasons for this warning 2. What is the preventive measure to be taken / Methods to detect its occurance 3. What are the post occurance measures/solutions for this. Regards A: Increase size of your redo logs. A: Amongst other reasons, this happens when redo logs are not sized properly. A checkpoint could not be completed because a new log is trying to be allocated while it is still in use (or hasn't been archived yet). This can happen if you are running very long transactions that are producing large amounts of redo (which you did not anticipate) and the redo logs are too small to handle it. If you are not archiving, increasing the size of your logfiles should help (each log group should have at least 2 members on separate disks). Also, be aware of what type of hardware you are using. Typically, raid-5 is slower for writes than raid-1. If you are archiving and have increased the size of the redo logs, also try adding an additional arch process. I have read plenty of conflicting documentation on how to resolve this problem. One of the "solutions" is to increase the size of your logbuffer. I have not found this to be helpful (for my particular databases). In the future, make sure to monitor the ratio of redo log entries to requests (it should be around 5000 to 1). If it slips below this ratio, you may want to consider adding addtional members to your log groups and increasing their size. A: Configuring redo logs is an art and you may never archieve 100% of the time that there is no waiting for available log files. But in my opinion, the best bet for your situation is to add one (or more) redo log instead of increase the size of the redo logs. Because even if your redo logs are huge, but if your disk controller is slow, a large transaction (for example, data loading) may use up all three redo logs before the first redo log completes the archive and becomes available, thus Oracle will halt until the archive is completed. 19.71: tkcrrsarc: (WARN) Failed to find ARCH for message (message:0x10): ------------------------------------------------------------------------ -- thread 1: Q: tkcrrsarc: (WARN) Failed to find ARCH for message (message:0x1) tkcrrpa: (WARN) Failed initial attempt to send ARCH message (message:0x1) Most repords speak of a harmless message. Some reports refer to a bug affecting Oracle versions op to 10.2.0.2 19.72 ORA-600 12333 ------------------- thread 1: ORA-600[12333] is reported with three additional numeric values when arequest is being received from a network packet and the request code inthe packet is not recognized. The three additional values report theinvalid request values received. The error may have a number of different root causes. For example, anetwork error may have caused bad data to be received, or the clientapplication may have sent wrong data, or the data in the network buffermay have been overwritten. Since there are many potential causes of thiserror, it is essential to have a reproducible testcase to correctlydiagnose the underlying cause. If operating system network logs areavailable, it is advisable to check them for evidence of networkfailures which may indicate network transmission problems. thread 2: We just found out that it was related to Block option DML RETURNINGVALUE in Forms4.5 We set it to NO, and the problem was solved Thanks anyway thread 3: From: Oracle, Kalpana Malligere 05-Oct-99 22:09 Subject: Re : ORA-00600: internal error code, arguments: [12333], [0], [3], [81], [], [], [] Hello, An ORA-600 12333 occurs because there has been a client/server protocol violation. There can be many reasons for this: Network errors, network hardware problems, etc. Where do you see or when do you get this error? Do you have any idea what was going on at the time of this error? Which process received it, i.e., was it a background or user process? Were you running sql*loader? Does this error have any adverse impact on the application or database? We cannot generally progress unless there is reproducible test case or reproducible environment. There are many bugs logged for this error which are closed as 'could not reproduce'. In one such bug, the developer indicated that "The problem does not normally have any bad side effects." So suggest you try to isolate what is causing it as much as possible. The error can be due to underlying network problems as well. It is not indicative of a problem with the database itself. 19.73: SMON: Parallel transaction recovery tried: ------------------------------------------------- Note 1: ------- Q: I was inserting 2.000.000 records in a table and the connection has been killed. in my alert file I found the following message : "SMON: Parallel transaction recovery tried" here the content of the smon log file: Redo thread mounted by this instance: 1 Oracle process number: 6 Windows thread id: 2816, image: ORACLE.EXE *** 2006-06-29 21:33:05.484 *** SESSION ID:(5.1) 2006-06-29 21:33:05.453 *** 2006-06-29 21:33:05.484 SMON: Restarting fast_start parallel rollback *** 2006-06-30 02:50:54.695 SMON: Parallel transaction recovery tried A: Hi, This is an expected message when cleanup is occuring and you have fast_start_parallel_rollback set to cleanup rollback segments after a failed transaction Note 2: ------- You get this message if SMON failed to generate the slave servers necessary to perform a parallel rollback of a transaction. Check the value for the parameter, FAST_START_PARALLEL_ROLLBACK (default is LOW). LOW limits the number of rollback processes to 2 * CPU_COUNT. HIGH limits the number of rollback processes to 4 * CPU_COUNT. You may want to set the value of this parameter to FALSE. Received on Wed Mar 10 2004 - 23:58:40 CST Note 3: ------- Q: SMON: Parallel transaction recovery tried We found above message in alert_sid.log file. A: No need to worry about it. it is information message ... SMON start recovery in parrallel but failed and done in serial mode. Note 4: ------- The system monitor process (SMON) performs recovery, if necessary, at instance startup. SMON is also responsible for cleaning up temporary segments that are no longer in use and for coalescing contiguous free extents within dictionary managed tablespaces. If any terminated transactions were skipped during instance recovery because of file-read or offline errors, SMON recovers them when the tablespace or file is brought back online. SMON checks regularly to see whether it is needed. Other processes can call SMON if they detect a need for it. With Real Application Clusters, the SMON process of one instance can perform instance recovery for a failed CPU or instance. 19.74: KGX Atomic Operation: ============================ Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production With the Partitioning, OLAP and Data Mining options ORACLE_HOME = /dbms/tdbaaccp/ora10g/home System name: AIX Node name: pl003 Release: 3 Version: 5 Machine: 00CB560D4C00 Instance name: accptrid Redo thread mounted by this instance: 1 Oracle process number: 16 Unix process pid: 2547914, image: oracle@pl003 (TNS V1-V3) *** 2008-03-20 07:22:28.571 *** SERVICE NAME:(SYS$USERS) 2008-03-20 07:22:28.570 *** SESSION ID:(161.698) 2008-03-20 07:22:28.570 KGX cleanup... KGX Atomic Operation Log 700000036eb4350 Mutex 70000003f9adcf8(161, 0) idn 0 oper EXAM Cursor Parent uid 161 efd 5 whr 26 slp 0 oper=DEFAULT pt1=700000039ce1c30 pt2=700000039ce1e18 pt3=700000039ce2338 pt4=0 u41=0 stt=0 Note 1: ------- Q: Hi there, Oracle has started using mutexes and it is said that they are more efficient as compared to latches. Questions 1)What is mutex?I know mutex are mutual exclusions and they are the concept of multiple threads.What I want to know that how this concept is implemented in Oracledatabase? 2) How they are better than latches?both are used for low level locking so how one is better than the other? Any input is welcome. Thanks and regards Aman.... A: 1) Simply put mutexes are memory structures. They are used to serialize the access to shared structures. IMHO their most important characteristics are two. First, they can be taken in shared or exclusive mode. Second, getting a mutex can be done in wait or no-wait mode. 2) The main advantages over latches are that mutexes requires less memory and are faster to get and release. A: In Oracle, latches and mutexes are different things and managed using different modules. KSL* modules for latches and KGX* for mutexes. As Chris said, general mutex operatins require less CPU instructions than latch operations (as they aren't as sophisticated as latches and don't maintain get/miss counts as latches do). But the main scalability benefit comes from that there's a mutex structure in each child cursor handle and the mutex itself acts as cursor pin structure. So if you have a cursor open (or cached in session cursor cache) you don't need to get the library cache latch (which was previously needed for changing cursor pin status), but you can modify the cursor's mutex refcount directly (with help of pointers in open cursor state area in sessions UGA). Therefore you have much higher scalability when pinning/unpinning cursors (no library cache latching needed, virtually no false contention) and no separate pin structures need to be allocated/maintained. Few notes: 1) library cache latching is still needed for parsing etc, the mutexes address only the pinning issue in library cache 2) mutexes are currently used for library cache cursors (not other objects like PL/SQL stored procs, table defs etc) 3) As mutexes are a generic mechanism (not library cache specific) they're used in V$SQLSTATS underlying structures too 4) When mutexes are enabled, you won't see cursor pins from X$KGLPN anymore (as X$KGLPN is a fixed table based on the KGL pin array - which wouldn't be used for cursors anymore) 19.75: ktsmgtur(): TUR was not tuned for 361 secs: ================================================== [pl101][tdbaprod][/dbms/tdbaprod/prodrman/admin/dump/bdump] cat prodrman_mmnl_1011950.trc /dbms/tdbaprod/prodrman/admin/dump/bdump/prodrman_mmnl_1011950.trc Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production With the Partitioning, OLAP and Data Mining options ORACLE_HOME = /dbms/tdbaprod/ora10g/home System name: AIX Node name: pl101 Release: 3 Version: 5 Machine: 00CB85FF4C00 Instance name: prodrman Redo thread mounted by this instance: 1 Oracle process number: 12 Unix process pid: 1011950, image: oracle@pl101 (MMNL) *** 2008-03-25 06:58:08.841 *** SERVICE NAME:(SYS$BACKGROUND) 2008-03-25 06:58:08.811 *** SESSION ID:(105.1) 2008-03-25 06:58:08.811 ktsmgtur(): TUR was not tuned for 361 secs What does this mean? Note 1: ------- Tur works along with a pathchecker, and if a SAN connection is lost, TUR will complain. Note 2: ------- The MMNL process does also tracking query lengths and is involved in tuning undo_retention. Here is means that TUR was not tuned for 361 seconds. If these are occasional messages, it is not to worry about. Otherwise you should dive into undo tuning. 19.76: tkcrrpa: (WARN) Failed initial attempt to send ARCH message: =================================================================== > *** SERVICE NAME:() 2008-03-22 14:56:43.590 > *** SESSION ID:(221.1) 2008-03-22 14:56:43.590 > Maximum redo generation record size = 132096 bytes > Maximum redo generation change vector size = 98708 bytes > tkcrrsarc: (WARN) Failed to find ARCH for message (message:0x10) > tkcrrpa: (WARN) Failed initial attempt to send ARCH message (message:0x10) No good answer yet. 19.77: Weird errors 1: ====================== In a trace file of an Oracle 10.2.0.3 db on AIX 5.3 we can find: >>>> DATABASE CALLED PRODTRID: > OS pid = 3907726 > loadavg : 1.12 1.09 1.13 > swap info: free_mem = 49.16M rsv = 24.00M > alloc = 2078.75M avail = 6144.00M swap_free = 4065.25M > F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD > 240001 A tdbaprod 3907726 1 0 60 20 1cfff7400 90692 06:00:39 - 0:00 ora_m000_prodtrid > open: Permission denied > 3907726: ora_m000_prodtrid > 0x00000001000f81e0 sskgpwwait(??, ??, ??, ??, ??) + ?? > 0x00000001000f5c54 skgpwwait(??, ??, ??, ??, ??) + 0x94 > 0x000000010010ba00 ksliwat(??, ??, ??, ??, ??, ??, ??, ??) + 0x640 > 0x0000000100116744 kslwaitns_timed(??, ??, ??, ??, ??, ??, ??, ??) + 0x24 > 0x0000000100170374 kskthbwt(0x0, 0x7000000, 0x0, 0x0, 0x15ab3c, 0x28284288, 0xfffffff, 0x7000000) + 0x214 > 0x0000000100116884 kslwait(??, ??, ??, ??, ??, ??) + 0x84 > 0x00000001002c8fb0 ksvrdp() + 0x550 > 0x00000001041c8c34 opirip(??, ??, ??) + 0x554 > 0x0000000102ab4ba8 opidrv(??, ??, ??) + 0x448 > 0x000000010409df30 sou2o(??, ??, ??, ??) + 0x90 > 0x0000000100000870 opimai_real(??, ??) + 0x150 > 0x00000001000006d8 main(??, ??) + 0x98 > 0x0000000100000360 __start() + 0x90 > *** 2008-04-01 06:01:43.294 At other instances we find: >>>> DATABASE CALLED PRODRMAN 06:01:41 - Check for changes since lastscan in file: /dbms/tdbaprod/prodrman/admin/dump/bdump/prodrman_cjq0_1003754.trc Warning: Errors detected in file /dbms/tdbaprod/prodrman/admin/dump/bdump/prodrman_cjq0_1003754.trc > OS pid = 3997922 > loadavg : 1.00 1.09 1.17 > swap info: free_mem = 62.76M rsv = 24.00M > alloc = 2087.91M avail = 6144.00M swap_free = 4056.09M > F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD > 240001 A tdbaprod 3997922 1 4 62 20 1322c8400 91516 05:43:28 - 0:00 ora_j000_prodrman > open: Permission denied > 3997922: ora_j000_prodrman > 0x00000001000f81e0 sskgpwwait(??, ??, ??, ??, ??) + ?? > 0x00000001000f5c54 skgpwwait(??, ??, ??, ??, ??) + 0x94 > 0x000000010010ba00 ksliwat(??, ??, ??, ??, ??, ??, ??, ??) + 0x640 > 0x0000000100116744 kslwaitns_timed(??, ??, ??, ??, ??, ??, ??, ??) + 0x24 > 0x0000000100170374 kskthbwt(0x0, 0x0, 0x7000000, 0x7000000, 0x15ab10, 0x1, 0xfffffff, 0x7000000) + 0x214 > 0x0000000100116884 kslwait(??, ??, ??, ??, ??, ??) + 0x84 > 0x00000001021d4fcc kkjsexe() + 0x32c > 0x00000001021d5d58 kkjrdp() + 0x478 > 0x00000001041c8bd0 opirip(??, ??, ??) + 0x4f0 > 0x0000000102ab4ba8 opidrv(??, ??, ??) + 0x448 > 0x000000010409df30 sou2o(??, ??, ??, ??) + 0x90 > 0x0000000100000870 opimai_real(??, ??) + 0x150 > 0x00000001000006d8 main(??, ??) + 0x98 > 0x0000000100000360 __start() + 0x90 > *** 2008-04-01 05:46:23.170 05:46:20 - Check for changes since lastscan in file: /dbms/tdbaprod/prodrman/admin/dump/bdump/prodrman_cjq0_1003754.trc Warning: Errors detected in file /dbms/tdbaprod/prodrman/admin/dump/bdump/prodrman_cjq0_1003754.trc > /dbms/tdbaprod/prodrman/admin/dump/bdump/prodrman_cjq0_1003754.trc > Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production > With the Partitioning, OLAP and Data Mining options > ORACLE_HOME = /dbms/tdbaprod/ora10g/home > System name: AIX > Node name: pl101 > Release: 3 > Version: 5 > Machine: 00CB85FF4C00 > Instance name: prodrman > Redo thread mounted by this instance: 1 > Oracle process number: 10 > Unix process pid: 1003754, image: oracle@pl101 (CJQ0) > > *** 2008-04-01 05:46:17.709 > *** SERVICE NAME:(SYS$BACKGROUND) 2008-04-01 05:44:28.394 > *** SESSION ID:(107.1) 2008-04-01 05:44:28.394 > Waited for process J000 to initialize for 60 seconds > *** 2008-04-01 05:46:17.709 > Dumping diagnostic information for J000: >>>> DATABASE CALLED ACCPROSS 06:01:26 - Check for changes since lastscan in file: /dbms/tdbaaccp/accpross/admin/dump/bdump/accpross_cjq0_1970272.trc Warning: Errors detected in file /dbms/tdbaaccp/accpross/admin/dump/bdump/accpross_cjq0_1970272.trc > /dbms/tdbaaccp/accpross/admin/dump/bdump/accpross_cjq0_1970272.trc > Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production > With the Partitioning, OLAP and Data Mining options > ORACLE_HOME = /dbms/tdbaaccp/ora10g/home > System name: AIX > Node name: pl003 > Release: 3 > Version: 5 > Machine: 00CB560D4C00 > Instance name: accpross > Redo thread mounted by this instance: 1 > Oracle process number: 10 > Unix process pid: 1970272, image: oracle@pl003 (CJQ0) > > *** 2008-04-01 06:01:21.210 > *** SERVICE NAME:(SYS$BACKGROUND) 2008-04-01 06:00:48.099 > *** SESSION ID:(217.1) 2008-04-01 06:00:48.099 > Waited for process J001 to initialize for 60 seconds > *** 2008-04-01 06:01:21.210 > Dumping diagnostic information for J001: > OS pid = 3645448 > loadavg : 1.28 1.18 1.16 > swap info: free_mem = 107.12M rsv = 24.00M > alloc = 3749.61M avail = 6144.00M swap_free = 2394.39M > F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD > 240001 A tdbaaccp 3645448 1 8 64 20 7566c510 91844 05:59:48 - 0:00 ora_j001_accpross > open: Permission denied > 3645448: ora_j001_accpross > 0x00000001000f81e0 sskgpwwait(??, ??, ??, ??, ??) + ?? > 0x00000001000f5c54 skgpwwait(??, ??, ??, ??, ??) + 0x94 > 0x000000010010ba00 ksliwat(??, ??, ??, ??, ??, ??, ??, ??) + 0x640 > 0x0000000100116744 kslwaitns_timed(??, ??, ??, ??, ??, ??, ??, ??) + 0x24 > 0x0000000100170374 kskthbwt(0x0, 0x0, 0x7000000, 0x7000000, 0x16656c, 0x1, 0xfffffff, 0x7000000) + 0x214 > 0x0000000100116884 kslwait(??, ??, ??, ??, ??, ??) + 0x84 > 0x00000001021d4fcc kkjsexe() + 0x32c > 0x00000001021d5d58 kkjrdp() + 0x478 > 0x00000001041c8bd0 opirip(??, ??, ??) + 0x4f0 > 0x0000000102ab4ba8 opidrv(??, ??, ??) + 0x448 > 0x000000010409df30 sou2o(??, ??, ??, ??) + 0x90 > 0x0000000100000870 opimai_real(??, ??) + 0x150 > 0x00000001000006d8 main(??, ??) + 0x98 > 0x0000000100000360 __start() + 0x90 > *** 2008-04-01 06:01:26.792 >>>> DATABASE CALLED PRODROSS 05:15:00 - Check for changes since lastscan in file: /dbms/tdbaprod/prodross/admin/dump/bdump/prodross_cjq0_2068516.trc Warning: Errors detected in file /dbms/tdbaprod/prodross/admin/dump/bdump/prodross_cjq0_2068516.trc > /dbms/tdbaprod/prodross/admin/dump/bdump/prodross_cjq0_2068516.trc > Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production > With the Partitioning, OLAP and Data Mining options > ORACLE_HOME = /dbms/tdbaprod/ora10g/home > System name: AIX > Node name: pl101 > Release: 3 > Version: 5 > Machine: 00CB85FF4C00 > Instance name: prodross > Redo thread mounted by this instance: 1 > Oracle process number: 10 > Unix process pid: 2068516, image: oracle@pl101 (CJQ0) > > *** 2008-04-01 05:13:52.362 > *** SERVICE NAME:(SYS$BACKGROUND) 2008-04-01 05:11:46.862 > *** SESSION ID:(217.1) 2008-04-01 05:11:46.861 > Waited for process J000 to initialize for 60 seconds > *** 2008-04-01 05:13:52.362 > Dumping diagnostic information for J000: > OS pid = 1855710 > loadavg : 1.08 1.15 1.20 > swap info: free_mem = 63.91M rsv = 24.00M > alloc = 2110.61M avail = 6144.00M swap_free = 4033.39M > F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD > 240001 A tdbaprod 1855710 1 4 66 22 1cb2f5400 92672 05:10:46 - 0:00 ora_j000_prodross > open: Permission denied > 1855710: ora_j000_prodross > 0x00000001000f81e0 sskgpwwait(??, ??, ??, ??, ??) + ?? > 0x00000001000f5c54 skgpwwait(??, ??, ??, ??, ??) + 0x94 > 0x000000010010ba00 ksliwat(??, ??, ??, ??, ??, ??, ??, ??) + 0x640 > 0x0000000100116744 kslwaitns_timed(??, ??, ??, ??, ??, ??, ??, ??) + 0x24 > 0x0000000100170374 kskthbwt(0x0, 0x0, 0x7000000, 0x7000000, 0x15aab2, 0x1, 0xfffffff, 0x7000000) + 0x214 > 0x0000000100116884 kslwait(??, ??, ??, ??, ??, ??) + 0x84 > 0x00000001021d4fcc kkjsexe() + 0x32c > 0x00000001021d5d58 kkjrdp() + 0x478 > 0x00000001041c8bd0 opirip(??, ??, ??) + 0x4f0 > 0x0000000102ab4ba8 opidrv(??, ??, ??) + 0x448 > 0x000000010409df30 sou2o(??, ??, ??, ??) + 0x90 > 0x0000000100000870 opimai_real(??, ??) + 0x150 > 0x00000001000006d8 main(??, ??) + 0x98 > 0x0000000100000360 __start() + 0x90 > *** 2008-04-01 05:13:59.017 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodroca/admin/dump/bdump/prodroca_cjq0_757946.trc Warning: Errors detected in file /dbms/tdbaprod/prodroca/admin/dump/bdump/prodroca_cjq0_757946.trc > OS pid = 1867996 > loadavg : 1.00 1.09 1.17 > swap info: free_mem = 66.71M rsv = 24.00M > alloc = 2087.91M avail = 6144.00M swap_free = 4056.09M > F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD > 240001 A tdbaprod 1867996 1 3 65 22 1078c5400 92656 05:44:06 - 0:00 ora_j000_prodroca > open: Permission denied > 1867996: ora_j000_prodroca > 0x00000001000f81e0 sskgpwwait(??, ??, ??, ??, ??) + ?? > 0x00000001000f5c54 skgpwwait(??, ??, ??, ??, ??) + 0x94 > 0x000000010010ba00 ksliwat(??, ??, ??, ??, ??, ??, ??, ??) + 0x640 > 0x0000000100116744 kslwaitns_timed(??, ??, ??, ??, ??, ??, ??, ??) + 0x24 > 0x0000000100170374 kskthbwt(0x0, 0x0, 0x7000000, 0x7000000, 0x15ab10, 0x1, 0xfffffff, 0x7000000) + 0x214 > 0x0000000100116884 kslwait(??, ??, ??, ??, ??, ??) + 0x84 > 0x00000001021d4fcc kkjsexe() + 0x32c > 0x00000001021d5d58 kkjrdp() + 0x478 > 0x00000001041c8bd0 opirip(??, ??, ??) + 0x4f0 > 0x0000000102ab4ba8 opidrv(??, ??, ??) + 0x448 > 0x000000010409df30 sou2o(??, ??, ??, ??) + 0x90 > 0x0000000100000870 opimai_real(??, ??) + 0x150 > 0x00000001000006d8 main(??, ??) + 0x98 > 0x0000000100000360 __start() + 0x90 > *** 2008-04-01 05:46:23.398 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodtrid/admin/dump/bdump/prodtrid_mmon_921794.trc Warning: Errors detected in file /dbms/tdbaprod/prodtrid/admin/dump/bdump/prodtrid_mmon_921794.trc > /dbms/tdbaprod/prodtrid/admin/dump/bdump/prodtrid_mmon_921794.trc > Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production > With the Partitioning, OLAP and Data Mining options > ORACLE_HOME = /dbms/tdbaprod/ora10g/home > System name: AIX > Node name: pl101 > Release: 3 > Version: 5 > Machine: 00CB85FF4C00 > Instance name: prodtrid > Redo thread mounted by this instance: 1 > Oracle process number: 11 > Unix process pid: 921794, image: oracle@pl101 (MMON) > > *** 2008-04-01 06:01:39.797 > *** SERVICE NAME:(SYS$BACKGROUND) 2008-04-01 06:01:39.385 > *** SESSION ID:(106.1) 2008-04-01 06:01:39.385 > Waited for process m000 to initialize for 60 seconds > *** 2008-04-01 06:01:39.797 > Dumping diagnostic information for m000: 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodrman/admin/dump/bdump/alert_prodrman.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodrman/admin/dump/udump/sbtio.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodroca/admin/dump/bdump/alert_prodroca.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodroca/admin/dump/udump/sbtio.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodross/admin/dump/bdump/alert_prodross.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodross/admin/dump/udump/sbtio.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodslot/admin/dump/bdump/alert_prodslot.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodslot/admin/dump/udump/sbtio.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodtrid/admin/dump/bdump/alert_prodtrid.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/prodtrid/admin/dump/udump/sbtio.log 06:01:42 - Check for changes since lastscan in file: /dbms/tdbaprod/ora10g/home/network/log/listener.log File /dbms/tdbaprod/ora10g/home/network/log/listener.log is changed, but no errors detected Note 1: ------- Q: Hi, we're running oracle 10 on AIX 5.3 TL04. We're experiencing some troubles with paging space. We've got 7 GB real mem and 10 GB paging space, and smoetimes the paging space occupation increases and it "freezes" the server (no telnet nor console connection). We've seen oracle has shown this error: CODE *** 2007-06-18 11:16:49.696 Dump diagnostics for process q002 pid 786600 which did not start after 120 seconds: (spawn_time:x10BF1F175 now:x10BF3CB36 diff:x1D9C1) *** 2007-06-18 11:16:54.668 Dumping diagnostic information for q002: OS pid = 786600 loadavg : 0.07 0.27 0.28 swap info: free_mem = 9.56M rsv = 40.00M alloc = 4397.23M avail = 10240.00M swap_free = 5842.77M skgpgpstack: fgets() timed out after 60 seconds skgpgpstack: pclose() timed out after 60 seconds ERROR: process 786600 is not alive *** 2007-06-18 11:19:41.152 *** 2007-06-18 11:27:36.403 Process startup failed, error stack: ORA-27300: OS system dependent operation:fork failed with status: 12 ORA-27301: OS failure message: Not enough space ORA-27302: failure occurred at: skgpspawn3 So we think it's oracle's fault, but we're not sure. We're AIX guys, not oracle, so we're not sure about this. Can anyone confirm if this is caused by oracle? A: Looks like a bug. We are running on a Windows 2003 Server Standard edition. I had the same problem. Server was not responding anymore after the following errors: ORA-27300: OS system dependent operation:spcdr:9261:4200 failed with status: 997 ORA-27301: OS failure message: Overlapped I/O operation is in progress. ORA-27302: failure occurred at: skgpspawn And later: O/S-Error: (OS 1450) Insufficient system resources exist to complete the requested service. We are running the latest patchset 10.2.0.2 because of a big problem in 10.2.0.1 (wrong parsing causes client memory problems. Procobol., plsql developer ect crash because oracle made mistakes skipping the parse process, goto direct execute and return corrupted data to the client. Tomorrow I will rise a level 1 TAR indicating we had a crach. Server is now running normaly. A: Oracle finally admit there was a bug: BUG 5607984 - ORACLE DOES NOT CLOSE TCP CONNECTIONS. REMAINS IN CLOSE_WAIT STATE. [On Windows 32-bit]. The patch 10 (patch number 5639232) is supposed to solve the problem for 10.2.0.2.0. We applied it monday morning and everything is fine up to now. This bug is also supposed to be solved in the 10.2.0.3.0 patchset that is availlable on the Metalink site. Note 2: ------- Q: question: ----------------------------------------------------------- my bdump received two error message traces this morning. One of the trace displays a lot of detail, mainly as: *** SESSION ID:(822.1) 2007-02-11 00:35:06.147 Waited for process J000 to initialize for 60 seconds *** 2007-02-11 00:35:20.276 Dumping diagnostic information for J000: OS pid = 811172 loadavg : 0.55 0.42 0.44 swap info: free_mem = 3.77M rsv = 24.50M alloc = 2418.36M avail = 6272.00M swap_free = 3853.64M F S UID PID PPID C PRI NI ADDR SZ WCHAN STIME TTY TIME CMD 240001 A oracle 811172 1 0 60 20 5bf12400 86396 00:34:32 - 0:00 ora_j000_BAAN open: The file access permissions do not allow the specified action. Then whole bunch of the pointers and something like this "0x0000000100055800 kghbshrt(??, ??, ??, ??, ??, ??) + 0x80" how do I find out what really went wrong? This error occured after I did an export pump of the DB, about 10 minutes later. This is first time I sae such and the export pump has been for a year. My system is Oracle 10g R2 on AIX 5.3L Note 3: ------- At least here you have an explanation about the Oracle processes: pmon The process monitor performs process recovery when a user process fails. PMON is responsible for cleaning up the cache and freeing resources that the process was using. PMON also checks on the dispatcher processes (described later in this table) and server processes and restarts them if they have failed. mman Used for internal database tasks. dbw0 The database writer writes modified blocks from the database buffer cache to the datafiles. Oracle Database allows a maximum of 20 database writer processes (DBW0-DBW9 and DBWa-DBWj). The initialization parameter DB_WRITER_PROCESSES specifies the number of DBWn processes. The database selects an appropriate default setting for this initialization parameter (or might adjust a user specified setting) based upon the number of CPUs and the number of processor groups. lgwr The log writer process writes redo log entries to disk. Redo log entries are generated in the redo log buffer of the system global area (SGA), and LGWR writes the redo log entries sequentially into a redo log file. If the database has a multiplexed redo log, LGWR writes the redo log entries to a group of redo log files. ckpt At specific times, all modified database buffers in the system global area are written to the datafiles by DBWn. This event is called a checkpoint. The checkpoint process is responsible for signalling DBWn at checkpoints and updating all the datafiles and control files of the database to indicate the most recent checkpoint. smon The system monitor performs recovery when a failed instance starts up again. In a Real Application Clusters database, the SMON process of one instance can perform instance recovery for other instances that have failed. SMON also cleans up temporary segments that are no longer in use and recovers dead transactions skipped during system failure and instance recovery because of file-read or offline errors. These transactions are eventually recovered by SMON when the tablespace or file is brought back online. reco The recoverer process is used to resolve distributed transactions that are pending due to a network or system failure in a distributed database. At timed intervals, the local RECO attempts to connect to remote databases and automatically complete the commit or rollback of the local portion of any pending distributed transactions. cjq0 Job Queue Coordinator (CJQ0) Job queue processes are used for batch processing. The CJQ0 process dynamically spawns job queue slave processes (J000...J999) to run the jobs. d000 Dispatchers are optional background processes, present only when the shared server configuration is used. s000 Dunno. qmnc Queue monitor background process A queue monitor process which monitors the message queues. Used by Oracle Streams Advanced Queuing. mmon Performs various manageability-related background tasks. mmnl Performs frequent and light-weight manageability-related tasks, such as session history capture and metrics computation. j000 A job queue slave. (See cjq0) Addition: --------- Sep 13, 2006 Oracle Background Processes, incl. 10gR2 ------------------- --New in 10gR2 ------------------- PSP0 (new in 10gR2) - Process SPawner - to create and manage other Oracle processes. NOTE: There is no documentation currently in the Oracle Documentation set on this process. LNS1(new in 10gR2) - a network server process used in a Data Guard (primary) database. Further explaination From "What's New in Oracle Data Guard?" in the Oracle® Data Guard Concepts and Administration 10g Release 2 (10.2) "During asynchronous redo transmission, the network server (LNSn) process transmits redo data out of the online redo log files on the primary database and no longer interacts directly with the log writer process. This change in behavior allows the log writer (LGWR) process to write redo data to the current online redo log file and continue processing the next request without waiting for inter-process communication or network I/O to complete." ------------------- --New in 10gR1 ------------------- MMAN - Memory MANager - it serves as SGA Memory Broker and coordinates the sizing of the memory components, which keeps track of the sizes of the components and pending resize operations. Used by Automatic Shared Memory Management feature. RVWR -Recovery Writer - which is responsible for writing flashback logs which stores pre-image(s) of data blocks. It is used by Flashback database feature in 10g, which provides a way to quickly revert an entire Oracle database to the state it was in at a past point in time. - This is different from traditional point in time recovery. - One can use Flashback Database to back out changes that: - Have resulted in logical data corruptions. - Are a result of user error. - This feature is not applicable for recovering the database in case of media failure. - The time required for flashbacking a database to a specific time in past is DIRECTLY PROPORTIONAL to the number of changes made and not on the size of the database. Jnnn - Job queue processes which are spawned as needed by CJQ0 to complete scheduled jobs. This is not a new process. CTWR - Change Tracking Writer (CTWR) which works with the new block changed tracking features in 10g for fast RMAN incremental backups. MMNL - Memory Monitor Light process - which works with the Automatic Workload Repository new features (AWR) to write out full statistics buffers to disk as needed. MMON - Memory MONitor (MMON) process - is associated with the Automatic Workload Repository new features used for automatic problem detection and self-tuning. MMON writes out the required statistics for AWR on a scheduled basis. M000 - MMON background slave (m000) processes. CJQn - Job Queue monitoring process - which is initiated with the job_queue_processes parameter. This is not new. RBAL - It is the ASM related process that performs rebalancing of disk resources controlled by ASM. ARBx - These processes are managed by the RBAL process and are used to do the actual rebalancing of ASM controlled disk resources. The number of ARBx processes invoked is directly influenced by the asm_power_limit parameter. ASMB - is used to provide information to and from the Cluster Synchronization Services used by ASM to manage the disk resources. It is also used to update statistics and provide a heartbeat mechanism. Changes about Queue Monitor Processes The QMON processes are optional background processes for Oracle Streams Advanced Queueing (AQ) which monitor and maintain all the system and user owned AQ objects. These optional processes, like the job_queue processes, does not cause the instance to fail on process failure. They provide the mechanism for message expiration, retry, and delay, maintain queue statistics, remove processed messages from the queue table and maintain the dequeue IOT. QMNx - Pre-10g QMON Architecture The number of queue monitor processes is controlled via the dynamic initialisation parameter AQ_TM_PROCESSES. If this parameter is set to a non-zero value X, Oracle creates that number of QMNX processes starting from ora_qmn0_ (where is the identifier of the database) up to ora_qmnX_ ; if the parameter is not specified or is set to 0, then QMON processes are not created. There can be a maximum of 10 QMON processes running on a single instance. For example the parameter can be set in the init.ora as follows aq_tm_processes=1 or set dynamically via alter system set aq_tm_processes=1; QMNC & Qnnn - 10g QMON Architecture Beginning with release 10.1, the architecture of the QMON processes has been changed to an automatically controlled coordinator slave architecture. The Queue Monitor Coordinator, ora_qmnc_, dynamically spawns slaves named, ora_qXXX_, depending on the system load up to a maximum of 10 in total. For version 10.01.XX.XX onwards it is no longer necessary to set AQ_TM_PROCESSES when Oracle Streams AQ or Streams is used. However, if you do specify a value, then that value is taken into account. However, the number of qXXX processes can be different from what was specified by AQ_TM_PROCESSES. If AQ_TM_PROCESSES is not specified in versions 10.1 and above, QMNC only runs when you have AQ objects in your database. 19.78: ORA-00600: internal error code, arguments: [13080], [], [], [], [], [], [], []: ====================================================================================== When running statement ALTER TABLE ENABLE CONSTAINT this ORA-00600 error appears. 19.79: WARNING: inbound connection timed out (ORA-3136): ======================================================== Note 1: Q: WARNING: inbound connection timed out (ORA-3136) this error appearing in Alert log . Please explain following:--------------- 1.How to overcome this error? 2.Is there any adverse effect in long run? 3.Is it require to SHUTDOWN the DATABASE to solve it. A: A good dicussion at freelist.ora http://www.freelists.org/archives/oracle-l/08-2005/msg01627.html In 10gR2, SQLNET.INBOUND_CONNECT_TIMEOUT the parameters were set to have a default of 60 (seconds). Set the parameters SQLNET.INBOUND_CONNECT_TIMEOUT and INBOUND_CONNECT_TIMEOUT_listenername to 0 (indefinite). A: What the error is telling you is that a connection attempt was made, but the session authentication was not provided before SQLNET.INBOUND_CONNECT_TIMEOUT seconds. As far as adverse effects in the long run, you have a user or process that is unable to connect to the database. So someone is unhappy about the database/application. Before setting SQLNET.INBOUND_CONNECT_TIMEOUT, verify that there is not a firewall or Network Address Translation (NAT) between the client and server. Those are common cause for ORA-3136. Q: Subject: WARNING: inbound connection timed out (ORA-3136) I have been getting like 50 of these error message a day in my alert_log the past couple of days. Anybody know what they mean? WARNING: inbound connection timed out (ORA-3136) A: Yep this is annoying, especially if you have alert log monitors :(. I had these when I first went to 10G... make these changes to get rid of them: Listener.ora: INBOUND_CONNECT_TIMEOUT_=0 .. for every listener Sqlnet.ora: SQLNET.INBOUND_CONNECT_TIMEOUT=0 Then the errors stop... Note 2: SQLNET.INBOUND_CONNECT_TIMEOUT Purpose Use the SQLNET.INBOUND_CONNECT_TIMEOUT parameter to specify the time, in seconds, for a client to connect with the database server and provide the necessary authentication information. If the client fails to establish a connection and complete authentication in the time specified, then the database server terminates the connection. In addition, the database server logs the IP address of the client and an ORA-12170: TNS:Connect timeout occurred error message to the sqlnet.log file. The client receives either an ORA-12547: TNS:lost contact or an ORA-12637: Packet receive failed error message. . Without this parameter, a client connection to the database server can stay open indefinitely without authentication. Connections without authentication can introduce possible denial-of-service attacks, whereby malicious clients attempt to flood database servers with connect requests that consume resources. To protect both the database server and the listener, Oracle Corporation recommends setting this parameter in combination with the INBOUND_CONNECT_TIMEOUT_listener_name parameter in the listener.ora file. When specifying values for these parameters, consider the following recommendations: Set both parameters to an initial low value. Set the value of the INBOUND_CONNECT_TIMEOUT_listener_name parameter to a lower value than the SQLNET.INBOUND_CONNECT_TIMEOUT parameter. For example, you can set INBOUND_CONNECT_TIMEOUT_listener_name to 2 seconds and INBOUND_CONNECT_TIMEOUT parameter to 3 seconds. If clients are unable to complete connections within the specified time due to system or network delays that are normal for the particular environment, then increment the time as needed. 19.80 How to insert special symbols: ==================================== Note 1: ------- Q: Hi, Is there anyone who knows how to insert a value containing "&" into a table? sth like this: insert into test_tab (test_field) values ('&test'); I tried ''&test' and many more but none of them works:-( As far as I know Oracle tries to bind a value when it encounters '&sth'... thanks in advance A: Try: set define off Then execute your insert. 19.81: SGA POLICY: Cache below reserve getting from component1: =============================================================== 19.82: AUTO SGA: Not free: ========================== Q: Hi, We have 10gr2 on windows server 2003 standard edition. The below errors are generated in mman trace files every now and then. AUTO SGA: Not free 0x2DFE78A8, 4, 1, 0 AUTO SGA: Not free 0x2DFE78A8, 4, 1, 0 AUTO SGA: Not free 0x2DFE795C, 4, 1, 0 AUTO SGA: Not free 0x2DFE7A10, 4, 1, 0 AUTO SGA: Not free 0x2DFE7AC4, 4, 1, 0 AUTO SGA: Not free 0x2DFE7B78, 4, 1, 0 AUTO SGA: Not free 0x2DFE7C2C, 4, 1, 0 AUTO SGA: Not free 0x2DFE7CE0, 4, 1, 0 AUTO SGA: Not free 0x2DFE7D94, 4, 1, 0 AUTO SGA: Not free 0x2DFF2708, 4, 1, 0 metalink doesnt give much info either.( BUG : 5201883 for your reference ) did anybody happened to have come across this issue and probably resolved it. Any comments are appreciated. A: This can be safely ignored.Since ASMM(Automatic Shared Memory Management) is enabled at instance level, you might be hitting this bug. Check Metalink note: 394026.1 Adding the Metalink note. A: As stated in the bug description, either 1) ignore the messages and delete generated trace files periodically and/or 2) wait for patchset 10.2.0.4 19.83: ORA-06512: ================= Error: ORA-06512: at line Cause: This error message indicates the line number in the PLSQL code that the error resulted. Action: For example, if you had the following PLSQL code: declare v_number number(2); begin v_number := 100; end; You would receive the following error message: ORA-06502 nummeric or value error. number precision too large ORA=06512 at line 4 So it tells you at which line the error is. 19.84: ORA-06512: at "SYS.DBMS_CDC_UTILITY": ============================================ Errormessage in trace file: > /dbms/tdbaplay/playroca/admin/dump/udump/playroca_ora_1667142.trc > Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production > With the Partitioning, OLAP and Data Mining options > ORACLE_HOME = /dbms/tdbaplay/ora10g/home > System name: AIX > Node name: pl003 > Release: 3 > Version: 5 > Machine: 00CB560D4C00 > Instance name: playroca > Redo thread mounted by this instance: 1 > Oracle process number: 27 > Unix process pid: 1667142, image: oracleplayroca@pl003 > > *** SERVICE NAME:(SYS$USERS) 2008-05-28 12:40:22.399 > *** SESSION ID:(530.27892) 2008-05-28 12:40:22.399 > oracle.jdbc.driver.OracleSQLException: ORA-01405: fetched column value is NULL > ORA-06512: at "SYS.DBMS_CDC_UTILITY", line 226 > ORA-06512: at line 1 > at oracle.jdbc.driver.T2SConnection.check_error(T2SConnection.java:153) > at oracle.jdbc.driver.T2SCallableStatement.checkError(T2SCallableStatement.java:92) > at oracle.jdbc.driver.T2SCallableStatement.executeForRows(T2SCallableStatement.java:449) > at oracle.jdbc.driver.OracleStatement.doExecuteWithTimeout(OracleStatement.java:1294) > at oracle.jdbc.driver.OraclePreparedStatement.executeInternal(OraclePreparedStatement.java:3514) > at oracle.jdbc.driver.OraclePreparedStatement.execute(OraclePreparedStatement.java:3620) > at oracle.jdbc.driver.OracleCallableStatement.execute(OracleCallableStatement.java:5261) > at oracle.CDC.SubscriptionWindow.qccsero(SubscriptionWindow.java:759) > at oracle.CDC.SubscriptionWindow.extendWindow(SubscriptionWindow.java:337) > at oracle.CDC.SubscribeApi.extendWindow(SubscribeApi.java:428) 19.85: TNS-12519: TNS:no appropriate service handler found: =========================================================== Note 1: ------- >> Hi all, >> >> we have an Oracle 10 (version details below) on Linux (RHEL 4). When >> connecting via JDBC we get intermittend ORA-12519 (reflected as >> TNS-12519 in listener.log). sqlldr also has a problem, although at >> the moment I can't exactly determine whether it's the same (I'm >> guessing it is because the happen about the same time). >> >> Research on the web revealed that a too low value for "processes" >> might be the reason. (The other possible cause I found was non >> matching versions of DB and client but this is not the case here.) >> So we increased DB param "parallel_max_servers" to 200. Since the >> error still showed up we went up to 400. It's been quiet since the >> last change of this parameter on Tuesday but some minutes ago I got >> an email notification that the error occurred again. >> >> I rather not want to increase the value by trial and error since we >> have only 36 sessions on the database right now and there seems to be >> a discrepancy between parameter "processes" (at 150 now, the value is >> derived from "parallel_max_servers") and the actual # of processes. >> Also the system is not much utilized and there's enough free >> resources (CPU wise and memory wise). So I'd like to first find out >> what is causing this error before I take further measures. >> >> I checked the alert log but there were no significant entries. I >> checked job scheduling to check whether there might be a job that >> eats up connections, but no. I guess switching on some trace might >> be helpful but at the moment I don't have an idea which one would be >> appropriate. Any ideas? Thanks for any insights! >> > > I fought this battle earlier this month. > The problem is that the more recent version listeners "count" > the incominng connection requests. When the count would exceed the > processes value the ORA-12519 error is raised. The problem is that > the listener does not really count the disconnetions. I learns > of them only periodically. If/when you have many, many short lived > connections you can see this error. I had found this article which basically states the same: http://forums.oracle.com/forums/thread.jspa?threadID=360226&tstart=0 Note 2: ------- This is an issue your DBA needs to address, it is not due to your client application -- Here is a Metalink note I found - for further information you will need to contact support. Article-ID: Note 240710.1 Title: Intermittent TNS-12516 or TNS-12519 Errors Connecting Via Net Symptom(s) ~~~~~~~~~~ Client connections may intermittently fail with either of the following errors: TNS-12516 TNS:listener could not find instance with matching protocol stack TNS-12519 TNS:no appropriate service handler found Additionally, a TNS-12520 error may appear in the listener log. TNS:listener could not find available handler for requested type of server The output of the lsnrctl services command may show that the service handler is in a "blocked" state. e.g. '"DEDICATED" established:1 refused:0 state:blocked' Change(s) ~~~~~~~~~~ None necessarily. Perhaps increase in load. Cause ~~~~~~~ By way of instance registration, PMON is responsible for updating the listener with information about a particular instance such as load and dispatcher information. Maximum load for dedicated connections is determined by the PROCESSES parameter. The frequency at which PMON provides SERVICE_UPDATE information varies according to the workload of the instance. The maximum interval between these service updates is 10 minutes. The listener counts the number of connections it has established to the instance but does not immediately get information about connections that have terminated. Only when PMON updates the listener via SERVICE_UPDATE is the listener informed of current load. Since this can take as long as 10 minutes, there can be a difference between the current instance load according to the listener and the actual instance load. When the listener believes the current number of connections has reached maximum load, it may set the state of the service handler for an instance to "blocked" and begin refusing incoming client connections with either of the following errors: TNS-12516 TNS:listener could not find instance with matching protocol stack TNS-12519 TNS:no appropriate service handler found Additionally, a TNS-12520 error may appear in the listener log. The output of lsnrctl service may show that the service handler is "blocked". e.g. '"DEDICATED" established:1 refused:0 state:blocked' Fix ~~~~ Increase the value for PROCESSES. Note 3: ------- Q: Hi all, I installed 10g Release 2 on Windows 2003 ... My application working fine .... recently i got an error .. TNS-12519: TNS:no appropriate service handler found i checked my listener it is working fine... but after some time connection start working... What was the issue? Regards Mani A: It could be a system load issue. If the problem happens again, try executing "lsnrctl services". If you see something like the following : Service "test10" has 1 instance(s). Instance "test10", status READY, has 1 handler(s) for this service... Handler(s): "DEDICATED" established:0 refused:0 state:blocked LOCAL SERVER then you may want to increase PROCESSES initialization parameter. 19.86. ORA-00313: open failed for members of log group 1 of thread 1: ===================================================================== Note 1: ------- Subject: Loss Of Online Redo Log And ORA-312 And ORA-313 Doc ID: 117481.1 Type: BULLETIN Modified Date : 04-AUG-2008 Status: PUBLISHED Scenario -------- You have a database in archive log mode, shutdown immediate and deleted one of the online redo logs, in this case there are only 2 groups with 1 log member in each. When you try to open the database you receive the following errors: ora-313 open failed for memebers of log group 2 of thread 1. ora-312 online log 2 thread 1 'filename' It is not possible to recover the missing log, so the following needs to be performed - Mount the database and check v$log to see if the deleted log is current; -If the log is not current, simply drop the log group (alter database drop logfile group N). If there are only 2 log groups then it will be necessary to add another group before dropping this one. -If the log is current they should simply perform fake recovery and then open resetlogs connect internal startup mount recover database until cancel; (cancel immediately) alter database open resetlogs; The database will open up as required, providing the log file directory is available. If not available then create it and rerun the resetlogs. This will give error ora-344 unable to recreate online log search words ------------ lost online redo ora-312 ora-313 Recovering After the Loss of Online Redo Log Files: Scenarios ============================================================= If a media failure has affected the online redo logs of a database, then the appropriate recovery procedure depends on the following: The configuration of the online redo log: mirrored or non-mirrored The type of media failure: temporary or permanent The types of online redo log files affected by the media failure: current, active, unarchived, or inactive 1) Recovering After Losing a Member of a Multiplexed Online Redo Log Group --------------------------------------------------------------------------- If the online redo log of a database is multiplexed, and if at least one member of each online redo log group is not affected by the media failure, then the database continues functioning as normal, but error messages are written to the log writer trace file and the alert_SID.log of the database. ACTION PLAN ============= If the hardware problem is temporary, then correct it. The log writer process accesses the previously unavailable online redo log files as if the problem never existed. If the hardware problem is permanent, then drop the damaged member and add a new member by using the following procedure. To replace a damaged member of a redo log group: =============================================== Locate the filename of the damaged member in V$LOGFILE. The status is INVALID if the file is inaccessible: SQL> SELECT GROUP#, STATUS, MEMBER FROM V$LOGFILE WHERE STATUS='INVALID'; GROUP# STATUS MEMBER ------- ----------- --------------------- 0002 INVALID /oracle/oradata/trgt/redo02.log ++ Drop the damaged member. For example, to drop member redo01.log from group 2, issue: SQL > ALTER DATABASE DROP LOGFILE MEMBER '/oracle/oradata/trgt/redo02.log'; ++ Add a new member to the group. For example, to add redo02.log to group 2, issue: SQL> ALTER DATABASE ADD LOGFILE MEMBER '/oracle/oradata/trgt/redo02b.log' TO GROUP 2; If the file you want to add already exists, then it must be the same size as the other group members, and you must specify REUSE. For example: SQL > ALTER DATABASE ADD LOGFILE MEMBER '/oracle/oradata/trgt/redo02b.log' REUSE TO GROUP 2; 2) Losing an Inactive Online Redo Log Group =========================================== If all members of an online redo log group with INACTIVE status are damaged, then the procedure depends on whether you can fix the media problem that damaged the inactive redo log group. If the failure is . . Temporary... then Fix the problem. LGWR can reuse the redo log group when required. If the failure is ... Permanent then the damaged inactive online redo log group eventually halts normal database operation. ACTION PLAN ============ Reinitialize the damaged group manually by issuing the ALTER DATABASE CLEAR LOGFILE You can clear an inactive redo log group when the database is open or closed. The procedure depends on whether the damaged group has been archived. To clear an inactive, online redo log group that has been archived: -------------------------------------------------------------------- If the database is shut down, then start a new instance and mount the database: STARTUP MOUNT Reinitialize the damaged log group. For example, to clear redo log group 2, issue the following statement: ALTER DATABASE CLEAR LOGFILE GROUP 2; Clearing Inactive, Not-Yet-Archived Redo ======================================== Clearing a not-yet-archived redo log allows it to be reused without archiving it. This action makes backups unusable if they were started before the last change in the log, unless the file was taken offline prior to the first change in the log. Hence, if you need the cleared log file for recovery of a backup, then you cannot recover that backup. Also, it prevents complete recovery from backups due to the missing log. To clear an inactive, online redo log group that has not been archived: If the database is shut down, then start a new instance and mount the database: STARTUP MOUNT Clear the log using the UNARCHIVED keyword. For example, to clear log group 2, issue: ALTER DATABASE CLEAR LOGFILE UNARCHIVED GROUP 2; If there is an offline datafile that requires the cleared log to bring it online, then the keywords UNRECOVERABLE DATAFILE are required. The datafile and its entire tablespace have to be dropped because the redo necessary to bring it online is being cleared, and there is no copy of it. For example, enter: ALTER DATABASE CLEAR LOGFILE UNARCHIVED GROUP 2 UNRECOVERABLE DATAFILE; Immediately back up the whole database with an operating system utility, so that you have a backup you can use for complete recovery without relying on the cleared log group. For example, enter: % cp /disk1/oracle/dbs/*.f /disk2/backup Back up the database's control file with the ALTER DATABASE statement. For example, enter: ALTER DATABASE BACKUP CONTROLFILE TO '/oracle/dbs/cf_backup.f'; Failure of CLEAR LOGFILE Operation ---------------------------------------- The ALTER DATABASE CLEAR LOGFILE statement can fail with an I/O error due to media failure when it is not possible to: * Relocate the redo log file onto alternative media by re-creating it under the currently configured redo log filename * Reuse the currently configured log filename to re-create the redo log file because the name itself is invalid or unusable (for example, due to media failure) In these cases, the ALTER DATABASE CLEAR LOGFILE statement (before receiving the I/O error) would have successfully informed the control file that the log was being cleared and did not require archiving. The I/O error occurred at the step in which the CLEAR LOGFILE statement attempts to create the new redo log file and write zeros to it. This fact is reflected in V$LOG.CLEARING_CURRENT. Note 2: ------- Subject: RMAN Recovery Until Time Failed When Redo-Logs Missed - ORA-00313, ORA-00312 AND ORA-27037 Doc ID: 550077.1 Type: PROBLEM Modified Date : 05-AUG-2008 Status: PUBLISHED In this Document Symptoms Cause Solution References -------------------------------------------------------------------------------- Applies to: Oracle Server - Enterprise Edition - Version: 9.2.0.6 to 10.2.0.3 This problem can occur on any platform. Symptoms o Testing rman in this scenario: - Taken a RMAN Backup for Database and Archiving logs -- successful. - All other files (datafiles, Online redolog files ) are deleted. - CURRENT CONTROLFILE (with rman-repository data) is still available, and Current Controlfile is used for RMAN Point In Time Recovery (Restore/Recover) RMAN Failed with Errors: ---- ORA-00283: recovery session canceled due to errors RMAN-11003: failure during parse/execution of SQL statement: alter database recover logfile 'D:\ORACLE\PRODUCT\10.2.0\FLASH_RECOVERY_AREA\L102\ARCHIVELOG\2008_01_22\O1_MF_1_12_3SCBO34D_.ARC' ORA-00283: recovery session canceled due to errors ORA-00313: open failed for members of log group 1 of thread 1 ORA-00312: online log 1 thread 1: 'D:\ORACLE\ORADATA\L102\REDO01.LOG' ORA-27041: unable to open file OSD-04002: unable to open file O/S-Error: (OS 2) The system cannot find the file specified. ---- Cause Bug:3902506 INCOMPLETE RECOVERY USING RMAN ENDED WITH ORA-313, ORA-312 AND ORA-27037 Bug:3991656 PLEASE ADD DESCRIPTION USING ONLINE REDO LOG WHEN USE UNTIL SCN AND UNTIL TIME If you use "UNTIL TIME" or "UNTIL SCN" clause without online redo log, you should use BACKUPED CONTROL FILES. . When using 'UNTIL SEQUENCE' clause for incomplete recovery of database, archive logs are restored and used for recovery with ALTER DATABASE RECOVER LOGFILE command. However, when UNTIL TIME or UNTIL CHANGE is used, even if the archive log is available, it is not used with 'ALTER DATABASE RECOVER LOGFILE' but instead RMAN tries to use the online redo log. . Solution -------------------- Solution/Workaround -------------------- o use BACKUP CONTROL FILE o use UNTIL SEQUENCE - instead of UNTIL TIME, or UNTIL SCN Example ------- SQL> select current_scn from v$database ; CURRENT_SCN ----------- 37079844 SQL> select sysdate from dual ; SYSDATE -------------------- 22-JAN-2008:16:06:35 SQL> select * from v$log ; GROUP# SEQUENCE# BYTES MEMBERS ARC STATUS FIRST_CHANGE# FIRST_TIME ------ ---------- --------- ------- --- -------- ------------- ---- 1 13 10485760 1 YES INACTIVE 37061094 22-JAN-2008:09:34:42 3 15 10485760 1 NO CURRENT 37075396 22-JAN-2008:14:33:21 2 14 10485760 1 YES INACTIVE 37068065 22-JAN-2008:12:00:28 *)1 SQL> shutdown immediate; Database closed. Database dismounted. ORACLE instance shut down. ---------------------- Recovery Actions done: ---------------------- o RMAN Online Backup for Database and Archivelogs was taken successfull o Then CURRENT CONTROLFILE will be used for the RMAN Point In Time RECOVERY o The UNTIL TIME CLAUSE is used to point to a time which is actually in the 'ONLINE REDOLOG' which is already Status ARCHIVED : YES *)1 # Mount the instance SQL> Startup MOUNT -- mounted using CURRENT COntrolfile # o RMAN RMAN> run { 2> set until time "to_date('22-JAN-2008:12:10:00', 'DD-MON-YYYY:HH24:MI:SS')" ; 3> restore database ; 4> recover database ; 5> } # RESTORE with UNTIL SCN: where SCN is pointing to an Online REdolog and is already Archived : YES ! RMAN> run { set until scn 37068100 ; # where SCN is pointing to an Online REdolog restore database check readonly ; # which is already Archived : YES ! recover database ; } ------- Errors: with UNTIL TIME, and UNTIL SCN ------- ... Restore finished successfull ... failed at reocver ... RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of recover command at 01/22/2008 16:27:46 ORA-00283: recovery session canceled due to errors RMAN-11003: failure during parse/execution of SQL statement: alter database recover logfile 'D:\ORACLE\PRODUCT\10.2.0\FLASH_RECOVERY_AREA\L102\ARCHIVELOG\2008_01_22\O1_MF_1_12_3SCBO34D_.ARC' ORA-00283: recovery session canceled due to errors ORA-00313: open failed for members of log group 1 of thread 1 ORA-00312: online log 1 thread 1: 'D:\ORACLE\ORADATA\L102\REDO01.LOG' ORA-27041: unable to open file OSD-04002: unable to open file O/S-Error: (OS 2) The system cannot find the file specified. -------------------------------------------- * Using a Backup COntrolfile is a Workaround -------------------------------------------- Hint: ----- If using the Curretn COntrolfile, you can mark this current controlfile to be a Backup controlfile by starting a Cancel based reocvery in SQL plus To change status of "Current Controlfile" to be a "BACKUP Controlfile" SQL> recover database until cancel using BACKUP CONTROLFILE ; .. then type: CANCEL # This will update a FLag in the Current Controfliel to mark it # as a Backup Controlfile # Then the RMAN should succeed with recovery, -------- Example: -------- SQL> recover database until cancel using backup controlfile ; ORA-279: change 37061094 generated at 01/22/2008 09:34:42 needed for thread 1 ORA-00289: suggestion : D:\ORACLE\...\ARCHIVELOG\2008_01_22\O1_MF_1_13_%_.ARC ORA-00280: change 37061094 for thread 1 is in sequence #13 Specify log: {=suggested | filename | AUTO | CANCEL} CANCEL Media recovery cancelled. SQL> -- -- Now the "Current Controlfile" is marked to be a "Backup Controlfile" -- => Then Run the RMAN RESTORE/RECOVER will succeed when using a BACKUP CONTROLFILE Note 3: ------- Subject: STARTING DATABASE, GETS ERROR, ORA-313. Doc ID: 1006148.6 Type: PROBLEM Modified Date : 20-JAN-2009 Status: PUBLISHED Problem Description: ==================== Database gives ORA-313 upon startup. ORA-00313: cannot open online '%s' (log #^ %s, log sequence # %s) Cause: The online log cannot be opened. Action: Restore online log. Solution Description: ===================== The Tape backup software had locked the logfile so Oracle could not update it. Everything worked fine once the tape backup job was terminated. Explanation: ============ The online logs are opened by the database at the time of startup. If these online logs are not available (maybe they have been deleted by mistake) or have been locked by a process, the above error will occur. The solution is to find out why they are not available and make them available. In this particular case, the tape backup software had locked the logfile. Terminating the tape backup job removed the lock on the online log files and made them available to oracle. Note: The online redo logfiles is another name for redo logfiles. Note 4: ------- Subject: ORA-00313 at Startup After a New Redo Log Memeber is Added Doc ID: 1005110.6 Type: PROBLEM Modified Date : 02-JUL-2007 Status: PUBLISHED Problem Description: ==================== A new redo log member is added to the database. However, if a shutdown and a subsequent startup is issued BEFORE the status of the log member as per V$LOGFILE goes from 'INVALID' (which is expected) to blank entry (i.e., 'IN USE'), you get an ORA-00313 and a trace file is generated. This scenario takes place after adding a redo log member to the database. For example: SQLDBA> ALTER DATABASE ADD LOGFILE MEMBER '' TO GROUP ; If you shutdown the database before the new member is used at the next startup, then an ORA-00313 is written to the alert.log file and a trace file (filename 'lgwr_xxxxx.trc') is also dumped. For the command above, such a trace file shows the following: *** SESSION ID:(3.1) ORA-00313: open failed for members of log group 1 of thread 1 The entire database, however, is still working just fine. Solution Description: ===================== This error message is meant to have useful information reported at startup. Since the new log member is still labeled as 'INVALID', it still does not have all file headers properly initialized, hence its contents cannot be trusted for recovery of any kind. However, once a log switch operation switches into the redo log group, the new redo log member will be then initialized and ready to write redo entries copied from the log buffer. What Development should have done was to just warn the DBA in the alert.log file. A trace file (lgwr_xxxxx.trc) is really not necessary. To accomplish this, however, Development has to come up with a new routine that only dumps warning messages of this nature to the alert.log file and NOT to the trace ===================== 20. DATABASE TRACING: ===================== 20.2 Oracle 10g: ================ 20.2.1 Tracing a session in 10g: -------------------------------- The current state of database and instance trace is reported in the data dictionary view DBA_ENABLED_TRACES. SQL> desc DBA_ENABLED_TRACES Name Null? Type ----------------------------------------- -------- ---------------------------- TRACE_TYPE VARCHAR2(21) PRIMARY_ID VARCHAR2(64) QUALIFIER_ID1 VARCHAR2(48) QUALIFIER_ID2 VARCHAR2(32) WAITS VARCHAR2(5) BINDS VARCHAR2(5) INSTANCE_NAME VARCHAR2(16) Note 1: 10g tracing quick start: -------------------------------- Oracle’s released a few new facilities to help with tracing in 10g, here’s a real quick wrap up of the most significant: >>>> Using the new client identifier: You can tag database sessions with a session identifier that can later be used to identify sessions to trace. You can set the identifier like this: begin dbms_session.set_identifier('GUY1'); end; You can set this from a login trigger if you don’t have access to the source code. To set trace on for a matching client id, you use DBMS_MONITOR.CLIENT_ID_TRACE_ENABLE: BEGIN DBMS_MONITOR.client_id_trace_enable (client_id => 'GUY1', waits => TRUE, binds => FALSE ); END; You can add waits and or bind variables to the trace file using the flags shown. >>>> Tracing by Module and/or action: Many Oracle-aware applications set Module and action properties and you can use these to enable tracing as well. The serv_mod_act_trace_enable method allows you to set the tracing on for sessions matching particular service, module, actions and (for clusters) instance identifiers. You can see current values for these usng the following query: SELECT DISTINCT instance_name, service_name, module, action FROM gv$session JOIN gv$instance USING (inst_id); INSTANCE_NAME SERVICE_NA MODULE ACTION ---------------- ---------- ------------------------------ ------------ ghrac11 SYS$USERS ghrac11 ghrac1 SQLNav5.exe ghrac11 ghrac1 Spotlight On Oracle, classic 4.0 ghrac13 SYS$USERS racgimon@mel601416.melquest.de v.mel.au.qsft (TNS ghrac13 ghrac1 Spotlight On Oracle, classic 4.0 ghrac12 ghrac1 SQL*Plus ghrac12 SYS$USERS racgimon@mel601416.melquest.de v.mel.au.qsft (TNS So to generate traces for all SQL*plus sessions that connect to the cluster from any instance, I could issue the following command: BEGIN DBMS_MONITOR.serv_mod_act_trace_enable (service_name => 'ghrac1', module_name => 'SQL*Plus', action_name => DBMS_MONITOR.all_actions, waits => TRUE, binds => FALSE, instance_name => NULL ); END; / >>>> Tracing using sid and serial DBMS_MONITOR can enable traces for specific sid and serial as you would expect: SELECT instance_name, SID, serial#, module, action FROM gv$session JOIN gv$instance USING (inst_id) WHERE username = 'SYSTEM'; INSTANCE_NAME SID SERIAL# MODULE ACTION ---------------- ---------- ---------- ------------ ------------ ghrac11 184 13179 SQL*Plus ghrac11 181 3353 SQLNav5.exe ghrac13 181 27184 SQL*Plus ghrac13 180 492 SQL*Plus ghrac12 184 18601 SQL*Plus BEGIN dbms_monitor.session_trace_enable (session_id => 180, serial_num => 492, waits => TRUE, binds => TRUE ); END; / BEGIN BEGIN dbms_monitor.session_trace_enable (session_id => 123, serial_num => 1826, waits => TRUE, binds => TRUE ); END; / BEGIN dbms_monitor.session_trace_enable (session_id => 124, serial_num => 914, waits => FALSE, binds => FALSE ); END; / The sid and serial need to be current now – unlike the other methods, this does not setup a permanent trace request (simply because the sid and serial# will never be repeated). Also, you need to issue this from the same instance if you are in a RAC cluster. Providing NULLs for sid and serial# traces the current session. >>>> Finding and analyzing the trace: This hasn’t changed much in 10g; the traces are in the USER_DUMP_DEST directory, and you can analyze them using tkprof. The trcsess utility is a new additional that allows you to generate a trace based on multiple input files and several other conditions. trcsess [output=] [session=] [clientid=] [service=] [action=] [module=] output= To generate a single trace file combining all the entries from the SQL*Plus sessions I traced earlier, then to feed them into tkprof for analysis, I would issue the following commands: [oracle@mel601416 udump]$ trcsess module='SQL*Plus' *.trc output=sqlplus.trc [oracle@mel601416 udump]$ tkprof sqlplus.trc sqlplus.prf TKPROF: Release 10.2.0.1.0 - Production on Wed Sep 27 14:47:51 2006 Note 2: ------- Setting Up Tracing with DBMS_MONITOR The DBMS_MONITOR package has routines for enabling and disabling statistics aggregation as well as for tracing by session ID, or tracing based upon a combination of service name, module name, and action name. (These three are associated hierarchically: you can't specify an action without specifying the module and the service name, but you can specify only the service name, or only the service name and module name.) The module and action names, if available, come from within the application code. For example, Oracle E-Business Suite applications provide module and action names in the code, so you can identify these by name in any of the Oracle Enterprise Manager pages. (PL/SQL developers can embed calls into their applications by using the DBMS_APPLICATION_INFO package to set module and action names.) Note that setting the module, action, and other paramters such as client_id no longer causes a round-trip to the database —these routines now piggyback on all calls from the application. The service name is determined by the connect string used to connect to a service. User sessions not associated with a specific service are handled by sys$users (sys$background is the default service for the background processes). Since we have a service and a module name, we can turn on tracing for this module as follows: SQL> exec dbms_monitor.serv_mod_act_trace_enable (service_name=>'testenv', module_name=>'product_update'); PL/SQL procedure successfully completed. We can turn on tracing for the client: SQL> exec dbms_monitor.client_id_trace_enable (client_id=>'kimberly'); PL/SQL procedure successfully completed. Note that all of these settings are persistent—all sessions associated with the service and module will be traced, not just the current sessions. To trace the SQL based on the session ID, look at the Oracle Enter-prise Manager Top Sessions page, or query the V$SESSION view as you likely currently do. SQL> select sid, serial#, username from v$session; SID SERIAL# USERNAME ------ ------- ------------ 133 4152 SYS 137 2418 SYSMAN 139 53 KIMBERLY 140 561 DBSNMP 141 4 DBSNMP . . . 168 1 169 1 170 1 28 rows selected. With the session ID (SID) and serial number, you can use DBMS_MONITOR to enable tracing for just this session: SQL> exec dbms_monitor.session_trace_enable(139); exec dbms_monitor.session_trace_enable(81); PL/SQL procedure successfully completed. The serial number defaults to the current serial number for the SID (unless otherwise specified), so if that's the session and serial number you want to trace, you need not look any further. Also, by default, WAITS are set to true and BINDS to false, so the syntax above is effectively the same as the following: SQL> exec dbms_monitor.session_trace_enable(session_id=>139, serial_num=>53, waits=>true, binds=>false); Note that WAITS and BINDS are the same parameters that you might have set in the past using DBMS_SUPPORT and the 10046 event. If you're working in a production environment, at this point you'd rerun the errant SQL or application, and the trace files would be created accordingly. Note 3: DBMS_MONITOR: --------------------- The DBMS_MONITOR package let you use PL/SQL for controlling additional tracing and statistics gathering. The chapter contains the following topics: Subprogram Description CLIENT_ID_STAT_DISABLE Procedure Disables statistic gathering previously enabled for a given Client Identifier CLIENT_ID_STAT_ENABLE Procedure Enables statistic gathering for a given Client Identifier CLIENT_ID_TRACE_DISABLE Procedure Disables the trace previously enabled for a given Client Identifier globally for the database CLIENT_ID_TRACE_ENABLE Procedure Enables the trace for a given Client Identifier globally for the database DATABASE_TRACE_DISABLE Procedure Disables SQL trace for the whole database or a specific instance DATABASE_TRACE_ENABLE Procedure Enables SQL trace for the whole database or a specific instance SERV_MOD_ACT_STAT_DISABLE Procedure Disables statistic gathering enabled for a given combination of Service Name, MODULE and ACTION SERV_MOD_ACT_STAT_ENABLE Procedure Enables statistic gathering for a given combination of Service Name, MODULE and ACTION SERV_MOD_ACT_TRACE_DISABLE Procedure Disables the trace for ALL enabled instances for a or a given combination of Service Name, MODULE and ACTION name globally SERV_MOD_ACT_TRACE_ENABLE Procedure Enables SQL tracing for a given combination of Service Name, MODULE and ACTION globally unless an instance_name is specified SESSION_TRACE_DISABLE Procedure Disables the previously enabled trace for a given database session identifier (SID) on the local instance SESSION_TRACE_ENABLE Procedure Enables the trace for a given database session identifier (SID) on the local instance ---------------------------------------------------------------------------------------------------------------------- -- CLIENT_ID_STAT_ENABLE Procedure This procedure enables statistic gathering for a given Client Identifier. Statistics gathering is global for the database and persistent across instance starts and restarts. That is, statistics are enabled for all instances of the same database, including restarts. Statistics are viewable through V$CLIENT_STATS views. Syntax DBMS_MONITOR.CLIENT_ID_STAT_ENABLE( client_id IN VARCHAR2); Parameters Table 60-3 CLIENT_ID_STAT_ENABLE Procedure Parameters Parameter Description client_id The Client Identifier for which statistic aggregation is enabled. Examples To enable statistic accumulation for a client with a given client ID: EXECUTE DBMS_MONITOR.CLIENT_ID_STAT_ENABLE('janedoe'); EXECUTE DBMS_MONITOR.CLIENT_ID_STAT_ENABLE('edp$jvl'); EXECUTE DBMS_MONITOR.CLIENT_ID_STAT_DISABLE('edp$jvl'); -- CLIENT_ID_STAT_DISABLE Procedure This procedure will disable statistics accumulation for all instances and remove the accumulated results from V$CLIENT_STATS view enabled by the CLIENT_ID_STAT_ENABLE Procedure. Syntax DBMS_MONITOR.CLIENT_ID_STAT_DISABLE( client_id IN VARCHAR2); Parameters Parameter Description client_id The Client Identifier for which statistic aggregation is disabled. Examples To disable accumulation: EXECUTE DBMS_MONITOR.CLIENT_ID_STAT_DISABLE('janedoe'); ---------------------------------------------------------------------------------------------------------------------- -- CLIENT_ID_TRACE_DISABLE Procedure This procedure will disable tracing enabled by the CLIENT_ID_TRACE_ENABLE Procedure. Syntax DBMS_MONITOR.CLIENT_ID_TRACE_DISABLE( client_id IN VARCHAR2); Parameters Table 60-4 CLIENT_ID_TRACE_DISABLE Procedure Parameters Parameter Description client_id The Client Identifier for which SQL tracing is disabled. Examples EXECUTE DBMS_MONITOR.CLIENT_ID_TRACE_DISABLE ('janedoe'); edp$jvl -- CLIENT_ID_TRACE_ENABLE Procedure This procedure will enable the trace for a given client identifier globally for the database. Syntax DBMS_MONITOR.CLIENT_ID_TRACE_ENABLE( client_id IN VARCHAR2, waits IN BOOLEAN DEFAULT TRUE, binds IN BOOLEAN DEFAULT FALSE); Parameters Table 60-5 CLIENT_ID_TRACE_ENABLE Procedure Parameters Parameter Description client_id Database Session Identifier for which SQL tracing is enabled. waits If TRUE, wait information is present in the trace. binds If TRUE, bind information is present in the trace. Usage Notes The trace will be written to multiple trace files because more than one Oracle shadow process can work on behalf of a given client identifier. The tracing is enabled for all instances and persistent across restarts. Examples EXECUTE DBMS_MONITOR.CLIENT_ID_TRACE_ENABLE('janedoe', TRUE,FALSE); EXECUTE DBMS_MONITOR.CLIENT_ID_TRACE_ENABLE('albert'); EXECUTE DBMS_MONITOR.CLIENT_ID_TRACE_DISABLE ('albert'); ---------------------------------------------------------------------------------------------------------------------- -- SERV_MOD_ACT_STAT_DISABLE Procedure This procedure will disable statistics accumulation and remove the accumulated results from V$SERV_MOD_ACT_STATS view. Statistics disabling is persistent for the database. That is, service statistics are disabled for instances of the same database (plus dblinks that have been activated as a result of the enable). Syntax DBMS_MONITOR.SERV_MOD_ACT_STAT_DISABLE( service_name IN VARCHAR2, module_name IN VARCHAR2, action_name IN VARCHAR2 DEFAULT ALL_ACTIONS); Parameters Table 60-8 SERV_MOD_ACT_STAT_DISABLE Procedure Parameters Parameter Description service_name Name of the service for which statistic aggregation is disabled. module_name Name of the MODULE. An additional qualifier for the service. It is a required parameter. action_name Name of the ACTION. An additional qualifier for the Service and MODULE name. Omitting the parameter (or supplying ALL_ACTIONS constant) means enabling aggregation for all Actions for a given Server/Module combination. In this case, statistics are aggregated on the module level. -- SERV_MOD_ACT_STAT_ENABLE Procedure This procedure enables statistic gathering for a given combination of Service Name, MODULE and ACTION. Calling this procedure enables statistic gathering for a hierarchical combination of Service name, MODULE name, and ACTION name on all instances for the same database. Statistics are accessible by means of the V$SERV_MOD_ACT_STATS view. Syntax DBMS_MONITOR.SERV_MOD_ACT_STAT_ENABLE( service_name IN VARCHAR2, module_name IN VARCHAR2, action_name IN VARCHAR2 DEFAULT ALL_ACTIONS); Parameters Table 60-9 SERV_MOD_ACT_STAT_ENABLE Procedure Parameters Parameter Description service_name Name of the service for which statistic aggregation is enabled. module_name Name of the MODULE. An additional qualifier for the service. It is a required parameter. action_name Name of the ACTION. An additional qualifier for the Service and MODULE name. Omitting the parameter (or supplying ALL_ACTIONS constant) means enabling aggregation for all Actions for a given Server/Module combination. In this case, statistics are aggregated on the module level. Usage Notes Enabling statistic aggregation for the given combination of Service/Module/Action names is slightly complicated by the fact that the Module/Action values can be empty strings which are indistinguishable from NULLs. For this reason, we adopt the following conventions: A special constant (unlikely to be a real action names) is defined: ALL_ACTIONS constant VARCHAR2 := '###ALL_ACTIONS'; Using ALL_ACTIONS for a module specification means that aggregation is enabled for all actions with a given module name, while using NULL (or empty string) means that aggregation is enabled for an action whose name is an empty string. Examples To enable statistic accumulation for a given combination of Service name and MODULE: EXECUTE DBMS_MONITOR.SERV_MOD_ACT_STAT_ENABLE( 'APPS1','PAYROLL'); To enable statistic accumulation for a given combination of Service name, MODULE and ACTION: EXECUTE DBMS_MONITOR.SERV_MOD_ACT_STAT_ENABLE('APPS1','GLEDGER','DEBIT_ENTRY'); If both of the preceding commands are issued, statistics are accumulated as follows: For the APPS1 service, because accumulation for each Service Name is the default. For all actions in the PAYROLL Module. For the DEBIT_ENTRY Action within the GLEDGER Module. ---------------------------------------------------------------------------------------------------------------------- -- DATABASE_TRACE_ENABLE Procedure This procedure enables SQL trace for the whole database or a specific instance. Syntax DBMS_MONITOR.DATABASE_TRACE_ENABLE( waits IN BOOLEAN DEFAULT TRUE, binds IN BOOLEAN DEFAULT FALSE, instance_name IN VARCHAR2 DEFAULT NULL); Parameters Table 60-7 DATABASE_TRACE_ENABLE Procedure Parameters Parameter Description waits If TRUE, wait information will be present in the trace binds If TRUE, bind information will be present in the trace instance_name If set, restricts tracing to the named instance EXECUTE dbms_monitor.database_trace_enable EXECUTE dbms_monitor.database_trace_disable -- DATABASE_TRACE_DISABLE Procedure This procedure disables SQL trace for the whole database or a specific instance. Syntax DBMS_MONITOR.DATABASE_TRACE_DISABLE( instance_name IN VARCHAR2 DEFAULT NULL); Parameters Table 60-6 DATABASE_TRACE_DISABLE Procedure Parameters Parameter Description instance_name Disables tracing for the named instance ---------------------------------------------------------------------------------------------------------------------- SERV_MOD_ACT_TRACE_DISABLE Procedure This procedure will disable the trace at ALL enabled instances for a given combination of Service Name, MODULE, and ACTION name globally. Syntax DBMS_MONITOR.SERV_MOD_ACT_TRACE_DISABLE( service_name IN VARCHAR2, module_name IN VARCHAR2, action_name IN VARCHAR2 DEFAULT ALL_ACTIONS, instance_name IN VARCHAR2 DEFAULT NULL); Parameters Table 60-10 SERV_MOD_ACT_TRACE_DISABLE Procedure Parameters Parameter Description service_name Name of the service for which tracing is disabled. module_name Name of the MODULE. An additional qualifier for the service. action_name Name of the ACTION. An additional qualifier for the Service and MODULE name. instance_name If set, this restricts tracing to the named instance_name. Usage Notes Specifying NULL for the module_name parameter means that statistics will no longer be accumulated for the sessions which do not set the MODULE attribute. Examples To enable tracing for a Service named APPS1: EXECUTE DBMS_MONITOR.SERV_MOD_ACT_TRACE_ENABLE('APPS1', DBMS_MONITOR.ALL_MODULES, DBMS_MONITOR.ALL_ACTIONS,TRUE, FALSE,NULL); To disable tracing specified in the previous step: EXECUTE DBMS_MONITOR.SERV_MOD_ACT_TRACE_DISABLE('APPS1'); To enable tracing for a given combination of Service and MODULE (all ACTIONs): EXECUTE DBMS_MONITOR.SERV_MOD_ACT_TRACE_ENABLE('APPS1','PAYROLL', DBMS_MONITOR.ALL_ACTIONS,TRUE,FALSE,NULL); To disable tracing specified in the previous step: EXECUTE DBMS_MONITOR.SERV_MOD_ACT_TRACE_DISABLE('APPS1','PAYROLL'); -------------------------------------------------------------------------------- SERV_MOD_ACT_TRACE_ENABLE Procedure This procedure will enable SQL tracing for a given combination of Service Name, MODULE and ACTION globally unless an instance_name is specified. Syntax DBMS_MONITOR.SERV_MOD_ACT_TRACE_ENABLE( service_name IN VARCHAR2, module_name IN VARCHAR2 DEFAULT ANY_MODULE, action_name IN VARCHAR2 DEFAULT ANY_ACTION, waits IN BOOLEAN DEFAULT TRUE, binds IN BOOLEAN DEFAULT FALSE, instance_name IN VARCHAR2 DEFAULT NULL); Parameters Table 60-11 SERV_MOD_ACT_TRACE_ENABLE Procedure Parameters Parameter Description service_name Name of the service for which tracing is enabled. module_name Name of the MODULE. An optional additional qualifier for the service. action_name Name of the ACTION. An optional additional qualifier for the Service and MODULE name. waits If TRUE, wait information is present in the trace. binds If TRUE, bind information is present in the trace. instance_name If set, this restricts tracing to the named instance_name. Usage Notes The procedure enables a trace for a given combination of Service, MODULE and ACTION name. The specification is strictly hierarchical: Service Name or Service Name/MODULE, or Service Name, MODULE, and ACTION name must be specified. Omitting a qualifier behaves like a wild-card, so that not specifying an ACTION means all ACTIONs. Using the ALL_ACTIONS constant achieves the same purpose. This tracing is useful when an application MODULE and optionally known ACTION is experiencing poor service levels. By default, tracing is enabled globally for the database. The instance_name parameter is provided to restrict tracing to named instances that are known, for example, to exhibit poor service levels. Tracing information is present in multiple trace files and you must use the trcsess tool to collect it into a single file. Specifying NULL for the module_name parameter means that statistics will be accumulated for the sessions which do not set the MODULE attribute. Examples To enable tracing for a Service named APPS1: EXECUTE DBMS_MONITOR.SERV_MOD_ACT_TRACE_ENABLE('APPS1', DBMS_MONITOR.ALL_MODULES, DBMS_MONITOR.ALL_ACTIONS,TRUE, FALSE,NULL); To enable tracing for a given combination of Service and MODULE (all ACTIONs): EXECUTE DBMS_MONITOR.SERV_MOD_ACT_TRACE_ENABLE('APPS1','PAYROLL', DBMS_MONITOR.ALL_ACTIONS,TRUE,FALSE,NULL); -------------------------------------------------------------------------------- SESSION_TRACE_DISABLE Procedure This procedure will disable the trace for a given database session at the local instance. Syntax DBMS_MONITOR.SESSION_TRACE_DISABLE( session_id IN BINARY_INTEGER DEFAULT NULL, serial_num IN BINARY_INTEGER DEFAULT NULL); Parameters Table 60-12 SESSION_TRACE_DISABLE Procedure Parameters Parameter Description session_id Name of the service for which SQL trace is disabled. serial_num Serial number for this session. Usage Notes If serial_num is NULL but session_id is specified, a session with a given session_id is no longer traced irrespective of its serial number. If both session_id and serial_num are NULL, the current user session is no longer traced. It is illegal to specify NULL session_id and non-NULL serial_num. In addition, the NULL values are default and can be omitted. Examples To enable tracing for a client with a given client session ID: EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(7,4634, TRUE, FALSE); To disable tracing specified in the previous step: EXECUTE DBMS_MONITOR.SESSION_TRACE_DISABLE(7,4634);; -------------------------------------------------------------------------------- SESSION_TRACE_ENABLE Procedure This procedure enables a SQL trace for the given Session ID on the local instance Syntax DBMS_MONITOR.SESSION_TRACE_ENABLE( session_id IN BINARY_INTEGER DEFAULT NULL, serial_num IN BINARY_INTEGER DEFAULT NULL, waits IN BOOLEAN DEFAULT TRUE, binds IN BOOLEAN DEFAULT FALSE) Parameters Table 60-13 SESSION_TRACE_ENABLE Procedure Parameters Parameter Description session_id Database Session Identifier for which SQL tracing is enabled. Specifying NULL means that my current session should be traced. serial_num Serial number for this session. Specifying NULL means that any session which matches session_id (irrespective of serial number) should be traced. waits If TRUE, wait information is present in the trace. binds If TRUE, bind information is present in the trace. Usage Notes The procedure enables a trace for a given database session, and is still useful for client/server applications. The trace is enabled only on the instance to which the caller is connected, since database sessions do not span instances. This tracing is strictly local to an instance. If serial_num is NULL but session_id is specified, a session with a given session_id is traced irrespective of its serial number. If both session_id and serial_num are NULL, the current user session is traced. It is illegal to specify NULL session_id and non-NULL serial_num. In addition, the NULL values are default and can be omitted. Examples To enable tracing for a client with a given client session ID: EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(7,4634, TRUE, FALSE); To disable tracing specified in the previous step: EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(82,30962); EXECUTE DBMS_MONITOR.SESSION_TRACE_DISABLE(82,30962); Either EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(5); or EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(5, NULL); traces the session with session ID of 5, while either EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(); or EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(NULL, NULL); traces the current user session. Also, EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(NULL, NULL, TRUE, TRUE); traces the current user session including waits and binds. The same can be also expressed using keyword syntax: EXECUTE DBMS_MONITOR.SESSION_TRACE_ENABLE(binds=>TRUE); Note 4: ------- End-to-End Tracing A common approach to diagnosing performance problems is to enable sql_trace to trace database calls and then analyze the output later using a tool such as tkprof. However, the approach has a serious limitation in databases with shared server architecture. In this configuration, several shared server processes are created to service the requests from the users. When user BILL connects to the database, the dispatcher passes the connection to an available shared server. If none is available, a new one is created. If this session starts tracing, the calls made by the shared server process are traced. Now suppose that BILL's session becomes idle and LORA's session becomes active. At that point the shared server originally servicing BILL is assigned to LORA's session. At this point, the tracing information emitted is not from BILL's session, but from LORA's. When LORA's session becomes inactive, this shared server may be assigned to another active session, which will have completely different information. In 10g, this problem has been effectively addressed through the use of end-to-end tracing. In this case, tracing is not done only by session, but by an identifiable name such as a client identifier. A new package called DBMS_MONITOR is available for this purpose. For instance, you may want to trace all sessions with the identifier account_update. To set up the tracing, you would issue: exec DBMS_MONITOR.CLIENT_ID_TRACE_ENABLE('account_update'); This command enables tracing on all sessions with the identifier account_update. When BILL connects to the database, he can issue the following to set the client identifier: exec DBMS_SESSION.SET_IDENTIFIER ('account_update') Tracing is active on the sessions with the identifier account_update, so the above session will be traced and a trace file will be generated on the user dump destination directory. If another user connects to the database and sets her client identifier to account_update, that session will be traced as well, automatically, without setting any other command inside the code. All sessions with the client identifier account_update will be traced until the tracing is disabled by issuing: exec DBMS_MONITOR.CLIENT_ID_TRACE_DISABLE('account_update'); The resulting trace files can be analyzed by tkprof. However, each session produces a different trace file. For proper problem diagnosis, we are interested in the consolidated trace file; not individual ones. How do we achieve that? Simple. Using a tool called trcsess, you can extract information relevant to client identifier account_update to a single file that you can run through tkprof. In the above case, you can go in the user dump destination directory and run: trcsess output=account_update_trc.txt clientid=account_update * This command creates a file named account_update_trc.txt that looks like a regular trace file but has information on only those sessions with client identifier account_update. This file can be run through tkprof to get the analyzed output. Contrast this approach with the previous, more difficult method of collecting trace information. Furthermore, tracing is enabled and disabled by some variable such as client identifier, without calling alter session set sql_trace = true from that session. Another procedure in the same package, SERV_MOD_ACT_TRACE_ENABLE, can enable tracing in other combinations such as for a specific service, module, or action, which can be set by dbms_application_info package. Note 5: ------- Generating SQL Trace Files Oracle Tips by Burleson Consulting The following Tip is from the outstanding book "Oracle PL/SQL Tuning: Expert Secrets for High Performance Programming" by Dr. Tim Hall, Oracle ACE of the year, 2006: There are numerous ways to enable, disable and vary the contents of this trace. The following methods have been available for several versions of the database. -- All versions. SQL> ALTER SESSION SET sql_trace=TRUE; SQL> ALTER SESSION SET sql_trace=FALSE; SQL> EXEC DBMS_SESSION.set_sql_trace(sql_trace => TRUE); SQL> EXEC DBMS_SESSION.set_sql_trace(sql_trace => FALSE); SQL> ALTER SESSION SET EVENTS '10046 trace name context forever, level 8'; SQL> ALTER SESSION SET EVENTS '10046 trace name context off'; SQL> EXEC DBMS_SYSTEM.set_sql_trace_in_session(sid=>123, serial#=>1234, sql_trace=>TRUE); SQL> EXEC DBMS_SYSTEM.set_sql_trace_in_session(sid=>123, serial#=>1234, sql_trace=>FALSE); SQL> EXEC DBMS_SYSTEM.set_ev(si=>123, se=>1234, ev=>10046, le=>8, nm=>' '); SQL> EXEC DBMS_SYSTEM.set_ev(si=>123, se=>1234, ev=>10046, le=>0, nm=>' '); -- All versions, requires DBMS_SUPPORT package to be loaded. SQL> EXEC DBMS_SUPPORT.start_trace(waits=>TRUE, binds=>FALSE); SQL> EXEC DBMS_SUPPORT.stop_trace; SQL> EXEC DBMS_SUPPORT.start_trace(sid=>123, serial=>1234, waits=>TRUE, binds=>FALSE); SQL> EXEC DBMS_SUPPORT.stop_trace(sid=>123, serial=>1234); The dbms_support package is not present by default, but can be loaded as the SYS user by executing the @$ORACLE_HOME/rdbms/admin/dbmssupp.sql script. For methods that require tracing levels, the following are valid values: 0 - No trace. Like switching sql_trace off. 2 - The equivalent of regular sql_trace. 4 - The same as 2, but with the addition of bind variable values. 8 - The same as 2, but with the addition of wait events. 12 - The same as 2, but with both bind variable values and wait events. The same combinations are possible for those methods with boolean parameters for waits and binds. With the advent of Oracle 10g, the SQL tracing options have been centralized and extended using the dbms_monitor package. The examples below show a few possible variations for enabling and disabling SQL trace in Oracle 10g. -- Oracle 10g SQL> EXEC DBMS_MONITOR.session_trace_enable; SQL> EXEC DBMS_MONITOR.session_trace_enable(waits=>TRUE, binds=>FALSE); SQL> EXEC DBMS_MONITOR.session_trace_disable; SQL> EXEC DBMS_MONITOR.session_trace_enable(session_id=>1234, serial_num=>1234); SQL> EXEC DBMS_MONITOR.session_trace_enable(session_id =>1234, serial_num=>1234, waits=>TRUE, binds=>FALSE); SQL> EXEC DBMS_MONITOR.session_trace_disable(session_id=>1234, serial_num=>1234); SQL> EXEC DBMS_MONITOR.client_id_trace_enable(client_id=>'tim_hall'); SQL> EXEC DBMS_MONITOR.client_id_trace_enable(client_id=>'tim_hall', waits=>TRUE, binds=>FALSE); SQL> EXEC DBMS_MONITOR.client_id_trace_disable(client_id=>'tim_hall'); SQL> EXEC DBMS_MONITOR.serv_mod_act_trace_enable(service_name=>'db10g', module_name=>'test_api', action_name=>'running'); SQL> EXEC DBMS_MONITOR.serv_mod_act_trace_enable(service_name=>'db10g', module_name=>'test_api', action_name=>'running', waits=>TRUE, binds=>FALSE); SQL> EXEC DBMS_MONITOR.serv_mod_act_trace_disable(service_name=>'db10g', module_name=>'test_api', action_name=>'running'); The package provides the conventional session level tracing along with two new variations. First, tracing can be enabled on multiple sessions based on the value of the client_identifier column of the v$session view, set using the dbms_session package. Second, tracing can be activated for multiple sessions based on various combinations of the service_name, module, action columns in the v$session view, set using the dbms_application_info package, along with the instance_name in RAC environments. With all the possible permutations and default values, this provides a high degree of flexibility. trcsess Activating trace on multiple sessions means that trace information is spread throughout many trace files. For this reason Oracle 10g introduced the trcsess utility, allowing trace information from multiple trace files to be identified and consolidated into a single trace file. The trcsess usage is listed below. trcsess [output=] [session=] [clientid=] [service=] [action=] [module=] output= output destination default being standard output. session= session to be traced. Session id is a combination of session Index & session serial number e.g. 8.13. clientid= clientid to be traced. service= service to be traced. action= action to be traced. module= module to be traced. Space separated list of trace files with wild card '*' supported. With all these options, the consolidated trace file can be as broad or as specific as needed. tkprof The SQL trace files produced by the methods discussed previously can be read in their raw form, or they can be translated by the tkprof utility into a more human readable form. The output below lists the usage notes from the tkprof utility in Oracle 10g. $ tkprof Usage: tkprof tracefile outputfile [explain= ] [table= ] [print= ] [insert= ] [sys= ] [sort= ] table=schema.tablename Use 'schema.tablename' with 'explain=' option. explain=user/password Connect to ORACLE and issue EXPLAIN PLAN. print=integer List only the first 'integer' SQL statements. aggregate=yes|no insert=filename List SQL statements and data inside INSERT statements. sys=no TKPROF does not list SQL statements run as user SYS. record=filename Record non-recursive statements found in the trace file. waits=yes|no Record summary for any wait events found in the trace file. sort=option Set of zero or more of the following sort options: prscnt number of times parse was called prscpu cpu time parsing prsela elapsed time parsing prsdsk number of disk reads during parse prsqry number of buffers for consistent read during parse prscu number of buffers for current read during parse prsmis number of misses in library cache during parse execnt number of execute was called execpu cpu time spent executing exeela elapsed time executing exedsk number of disk reads during execute exeqry number of buffers for consistent read during execute execu number of buffers for current read during execute exerow number of rows processed during execute exemis number of library cache misses during execute fchcnt number of times fetch was called fchcpu cpu time spent fetching fchela elapsed time fetching fchdsk number of disk reads during fetch fchqry number of buffers for consistent read during fetch fchcu number of buffers for current read during fetch fchrow number of rows fetched userid userid of user that parsed the cursor $ The waits parameter was only added in Oracle 9i, so prior to this version wait information had to be read from the raw trace file. The values of bind variables must be read from the raw files as they are not displayed in the tkprof output. 20.2 OLDER ORACLE Versions 8,8i,9i: =================================== 20.2.1 Trace a session: ----------------------- Examples: --------- exec DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(sid, serial#, TRUE); exec DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(23, 54071, TRUE); DBMS_SYSTEM has some mysterious and apparently dangerous procedures in it. Obtaining any information about SET_EV and READ_EV was very difficult and promises to be more difficult in the future since the package header is no longer exposed in Oracle 8.0. In spite of Oracle's desire to keep DBMS_SYSTEM "under wraps," I feel strongly that the SET_SQL_TRACE_IN_SESSION procedure is far too valuable to be hidden away in obscurity. DBAs and developers need to find out exactly what is happening at runtime when a user is experiencing unusual performance problems, and the SQL trace facility is one of the best tools available for discovering what the database is doing during a user's session. This is especially useful when investigating problems with software packages where source code (including SQL) is generally unavailable. So how can we get access to the one program in DBMS_SYSTEM we want without exposing those other dangerous elements to the public? The answer, of course, is to build a package of our own to encapsulate DBMS_SYSTEM and expose only what is safe. In the process, we can make DBMS_SYSTEM easier to use as well. Those of us who are "keyboard-challenged" (or just plain lazy) would certainly appreciate not having to type a procedure name with 36 characters. I've created a package called trace to cover DBMS_SYSTEM and provide friendlier ways to set SQL tracing on or off in other user's sessions. Here is the package specification: */ Filename on companion disk: trace.sql */* CREATE OR REPLACE PACKAGE trace IS type rr_rec is record ( v_sid number, v_serial number ); r_rec rr_rec; /* || Exposes DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION || with easier to call programs || || Author: John Beresniewicz, Savant Corp || Created: 07/30/97 || || Compilation Requirements: || SELECT on SYS.V_$SESSION || EXECUTE on SYS.DBMS_SYSTEM (or create as SYS) || || Execution Requirements: || */ /* turn SQL trace on by session id */ PROCEDURE Xon(sid_IN IN NUMBER); /* turn SQL trace off by session id */ PROCEDURE off(sid_IN IN NUMBER); /* turn SQL trace on by username */ PROCEDURE Xon(user_IN IN VARCHAR2); /* turn SQL trace off by username */ PROCEDURE off(user_IN IN VARCHAR2); END trace; The trace package provides ways to turn SQL tracing on or off by session id or username. One thing that annoys me about DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION is having to figure out and pass a session serial number into the procedure. There should always be only one session per sid at any time connected to the database, so trace takes care of figuring out the appropriate serial number behind the scenes. Another improvement (in my mind) is replacing the potentially confusing BOOLEAN parameter sql_trace with two distinct procedures whose names indicate what is being done. Compare the following commands, either of which might be used to turn SQL tracing off in session 15 using SQL*Plus: SQL> execute trace.off(sid_IN=>15); SQL> execute SYS.DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(15,4567,FALSE); The first method is both more terse and easier to understand. The xon and off procedures are both overloaded on the single IN parameter, with versions accepting either the numeric session id or a character string for the session username. Allowing session selection by username may be easier than by sids. Why? Because sids are transient and must be looked up at runtime, whereas username is usually permanently associated with an individual. Beware, though, that multiple sessions may be concurrently connected under the same username, and invoking trace.xon by username will turn tracing on in all of them. Let's take a look at the trace package body: /* Filename on companion disk: trace.sql */* CREATE OR REPLACE PACKAGE BODY trace IS /* || Use DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION to turn tracing on || or off by either session id or username. Affects all sessions || that match non-NULL values of the user and sid parameters. */ PROCEDURE set_trace (sqltrace_TF BOOLEAN ,user IN VARCHAR2 DEFAULT NULL ,sid IN NUMBER DEFAULT NULL) IS BEGIN /* || Loop through all sessions that match the sid and user || parameters and set trace on in those sessions. The NVL || function in the cursor WHERE clause allows the single || SELECT statement to filter by either sid OR user. */ FOR sid_rec IN (SELECT sid,serial# FROM v$session S WHERE S.type='USER' AND S.username = NVL(UPPER(user),S.username) AND S.sid = NVL(sid,S.sid) ) LOOP SYS.DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION (sid_rec.sid, sid_rec.serial#, sqltrace_TF); END LOOP; END set_trace; /* || The programs exposed by the package all simply || call set_trace with different parameter combinations. */ PROCEDURE Xon(sid_IN IN NUMBER) IS BEGIN set_trace(sqltrace_TF => TRUE, sid => sid_IN); END Xon; PROCEDURE off(sid_IN IN NUMBER) IS BEGIN set_trace(sqltrace_TF => FALSE, sid => sid_IN); END off; PROCEDURE Xon(user_IN IN VARCHAR2) IS BEGIN set_trace(sqltrace_TF => TRUE, user => user_IN); END Xon; PROCEDURE off(user_IN IN VARCHAR2) IS BEGIN set_trace(sqltrace_TF => FALSE, user => user_IN); END off; END trace; All of the real work done in the trace package is contained in a single private procedure called set_trace. The public procedures merely call set_trace with different parameter combinations. This is a structure that many packages exhibit: private programs with complex functionality exposed through public programs with simpler interfaces. One interesting aspect of set_trace is the cursor used to get session identification data from V_$SESSION. I wanted to identify sessions for tracing by either session id or username. I could have just defined two cursors on V_$SESSION with some conditional logic deciding which cursor to use, but that just did not seem clean enough. After all, less code means fewer bugs. The solution I arrived at: make use of the NVL function to have a single cursor effectively ignore either the sid or the user parameter when either is passed in as NULL. Since set_trace is always called with either sid or user, but not both, the NVLs act as a kind of toggle on the cursor. I also supplied both the sid and user parameters to set_trace with the default value of NULL so that only the parameter being used for selection needs be passed in the call. Once set_trace was in place, the publicly visible procedures were trivial. A final note about the procedure name "xon": I wanted to use the procedure name "on," but ran afoul of the PL/SQL compiler since ON is a reserved word in SQL and PL/SQL. You can also try: Alter system set sql_trace=true; Setting sql_trace=true is a prerequisite when using tk prof. -- TRACING a session: ----------------------- Enable tracing a session to generate a tarce file. This file can be formatted with TKPROF 6.1. The following INIT.ORA parameters must be set: #SQL_TRACE = TRUE USER_DUMP_DEST = TIMED_STATISTICS = TRUE MAX_DUMP_FILE_SIZE = 6.2 To enable the SQL trace facility for your current session, enter: ALTER SESSION SET SQL_TRACE = TRUE; or use DBMS_SUPPORT.START_TRACE_IN_SESSION( SID , SERIAL# ); DBMS_SUPPORT.STOP_TRACE_IN_SESSION( SID , NULL ); DBMS_SYSTEM.SET_SQL_TRACE_IN_SESSION(sid, serial#, TRUE); DBMS_SUPPORT.START_TRACE_IN_SESSION(86,43326); To enable the SQL trace facility for your instance, set the value of the SQL_TRACE initialization parameter to TRUE. Statistics will be collected for all sessions. Once the SQL trace facility has been enabled for the instance, you can disable it for an individual session by entering: ALTER SESSION SET SQL_TRACE = FALSE; 6.3 Examples of TKPROF TKPROF ora53269.trc ora 53269.prf SORT = (PRSDSK, EXEDSK, FCHDSK) PRINT = 10 To analyze the sql statements: 1. tkprof ora_11598.trc myfilename 2. tkprof ora_11598.trc /tmp/myfilename 3. tkprof ora_11598.trc /tmp/myfilename explain=ap/ap 4. tkprof ora_23532.trc myfilename explain=po/po sort=execpu 7 STATSPACK: ------------ Statspack is a set of SQL, PL/SQL, and SQL*Plus scripts that allow the collection, automation, storage, and viewing of perfoRMANce data (see Table 2). The installation script (statscre.sql) calls several other scripts in order to create the entire Statspack environment. (Note: You should run only the installation script, not the base scripts that statscre.sql invokes.) All the scripts you need for installing and running Statspack are in the ORACLE_HOME/rdbms/admin directory for UNIX platforms and in %ORACLE_HOME%\rdbms\admin for Microsoft Windows NT systems. The simplest interactive way to take a snapshot is to log in to SQL*Plus as the owner perfstat and execute the statspack.snap procedure: SQL> connect perfstat/perfstat SQL> execute statspack.snap; You can use dbms_job to automate statistics collection. The file statsauto.sql contains an example of how to do this, scheduling a snapshot every hour. When you create a job by using dbms_job, Oracle assigns the job a unique number that you can use for changing or removing the job. In order to use dbms_job to schedule snapshots automatically, you must set the job_queue_processes initialization parameter to greater than 0 in the init.ora file: # Set to enable the job-queue process to start. # This allows dbms_job to schedule automatic # statistics collection, using Statspack job_queue_processes=1 Change the interval of statistics collection by using the dbms_job.interval procedure: execute dbms_job.interval(, 'SYSDATE+(1/48)'); In this case, SYSDATE+(1/48)' causes the statistics to be gathered each 1/48 day-every half hour. To stop and remove the automatic-collection job: execute dbms_job.remove(); Install Statspack: CREATE USER perfstat identified by perfstat default tableSpace TOOLS temporary tableSpace TEMP; GRANT CREATE SeSSion to PERFSTAT; GRANT connect to PERFSTAT; GRANT reSource to PERFSTAT; GRANT unlimited tableSpace to PERFSTAT; sqlplus sys -- -- Install Statspack -- Enter tablespace names when prompted -- @?/rdbms/admin/spcreate.sql -- -- Drop Statspack -- Reverse of spcreate.sql -- -- @?/rdbms/admin/spdrop.sql -- The spcreate.sql install script automatically calls 3 other scripts needed: spcusr - creates the user and grants privileges spctab - creates the tables spcpkg - creates the package Check each of the three output files produced (spcusr.lis, spctab.lis, spcpkg.lis) by the installation to ensure no errors were encountered, before continuing on to the next step. Using Statspack (gathering data): sqlplus perfstat -- -- Take a perfoRMANce snapshot -- execute statspack.snap; -- -- Get a list of snapshots -- column snap_time format a21 SELECT snap_id,to_char(snap_time,'MON dd, yyyy hh24:mm:ss') snap_time FROM sp$snapshot; -- NOTE: To include important timing information set the init.ora parameter timed_statistics to true. To examine the change in instancewide statistics between two time periods, the SPREPORT.SQL file is run while connected to the PERFSTAT user. The SPREPORT.SQL command file is located in the rdbms/admin directory of the Oracle home. You are prompted for the following: The beginning snapshot ID The ending snapshot ID The name of the report text file to be created =========== 21. Overig: =========== 20.1 NLS: ========= Bij Server: 1. characterset specificatie bij CREATE DATABASE 2. De Sever kan wel meerdere locale in runtime laden uit files gespecificeerd in $ export ORA_NLSxx=$ORACLE_HOME/ocommon/nls/admin/data 3. NLS init.ora parameters t.b.v. de user sessions. If clients using different character sets will access the database, then choose a superset that includes all client character sets. Otherwise, character conversions may be necessary at the cost of increased overhead and potential data loss. client: 1. client heeft lokaal een NLS environment setting 2. client connect naar database, een session wordt gevormd, en de NLS enviroment wordt gemaakt aan de hAND van de NLS init.ora parameters. Is bij de clent de NLS_LANG environment variable gezet, dan communiceerd de client dat naar de server session. Hierdoor zijn beide hetzelfde. Is er geen NLS_LANG, dan gelden de init.ora NLS parameters voor de server session 3. De session NLS kan worden verANDert via ALTER SESSION. Dit heeft alleen effect op de PL/SQL en SQL statements executed op de server init.ora parameters bij server : invloed op sessions op server environment variables bij client : locale bij client, overrides session alter session statement : verANDert de session, overides init.ora expliciet in SQL statement : overides alles Voorbeeld van override: in init.ora: NLS_SORT=ENGLISH bij client: ALTER SESSION SET NLS_SORT=FRENCH; Examples: --------- Example 1: ---------- ALTER SESSION SET nls_date_format = 'dd/mm/yy' ALTER SESSION SET NLS_DATE_FORMAT = 'DD-MON-YYYY' ALTER SESSION SET NLS_LANGUAGE='ENGLISH'; ALTER SESSION SET NLS_LANGUAGE='NEDERLANDS'; export NLS_NUMERIC_CHARACTERS=',.' ALTER SESSION SET NLS_NUMERIC_CHARACTERS=',.' ALTER SESSION SET NLS_TERRITORY=France; ALTER SESSION SET NLS_TERRITORY=America; In SQL functions: NLS parameters can be used explicitly to hardcode NLS behavior within a SQL function. Doing so will override the default values that are set for the session in the initialization parameter file, set for the client with environment variables, or set for the session by the ALTER SESSION statement. For example: TO_CHAR(hiredate, 'DD/MON/YYYY', 'nls_date_language = FRENCH') SELECT last_name FROM employees WHERE hire_date > TO_DATE('01-JAN-1999','DD-MON-YYYY', 'NLS_DATE_LANGUAGE = AMERICAN'); Example 2: ---------- SQL> ALTER SESSION SET NLS_NUMERIC_CHARACTERS=',.' 2 ; Session altered. SQL> select * from ap2; NAME SAL ---------- ---------- ap 12,53 piet 89,7 SQL> ALTER SESSION SET NLS_NUMERIC_CHARACTERS='.,'; Session altered. SQL> select * from ap2; NAME SAL ---------- ---------- ap 12.53 piet 89.7 priority: --------- 1. expliciet in SQL 2. ALTER SESSION 3. environment variable 4. init.ora NLS parameters, te zetten via: NLS_CALENDAR init.ora, env, alter session NLS_COMP init.ora, env, alter session NLS_CREDIT - env - NLS_CURRENCY init.ora, env, alter session NLS_DATE_FORMAT init.ora, env, alter session NLS_DATE_LANGUAGE init.ora, env, alter session NLS_DEBIT - env - NLS_ISO_CURRENCY init.ora, env, alter session NLS_LANG - env - NLS_LANGUAGE init.ora, - , alter session NLS_LIST_SEPERATOR - env - NLS_MONETARY_CHARACTERS - env - NLS_NCHAR - env - NLS_NUMMERIC_CHARACTERS init.ora, env, alter session NLS_SORT init.ora, env, alter session NLS_TERRITORY init.ora, - , alter session NLS_DUAL_CURRENCY init.ora, env, alter session DATA DICTIONARY VIEWS: ---------------------- Applications can check the session, instance, and database NLS parameters by querying the following data dictionary views: NLS_SESSION_PARAMETERS shows the NLS parameters and their values for the session that is querying the view. It does not show information about the character set. NLS_INSTANCE_PARAMETERS shows the current NLS instance parameters that have been explicitly set and the values of the NLS instance parameters. NLS_DATABASE_PARAMETERS shows the values of the NLS parameters that were used when the database was created. Example: -------- SQL> desc ap1; Name Null? Type ----------------------------------------- -------- ------------- NAME VARCHAR2(10) SAL VARCHAR2(10) SQL> select * from ap1; NAME SAL ---------- ---------- ap 12,53 piet 89,7 SQL> desc ap2; Name Null? Type ----------------------------------------- -------- ---------------------------- NAME VARCHAR2(10) SAL NUMBER SQL> select * from ap2; NAME SAL ---------- ---------- ap 12.53 piet 89.7 SQL> insert into ap2 2 select * from ap1; select * from ap1 * ERROR at line 2: ORA-01722: invalid number SQL> ALTER SESSION SET NLS_NUMERIC_CHARACTERS=',.'; Session altered. SQL> insert into ap2 2 select * from ap1; 2 rows created. 20.2 More on AL32UTF8, AL16UTF16, UTF8: ======================================= 1) What is the National Character Set? -------------------------------------- The National Character set (NLS_NCHAR_CHARACTERSET) is a character set which is defined in addition to the (normal) database character set and is used for data stored in NCHAR, NVARCHAR2 and NCLOB columns. Your current value for the NLS_NCHAR_CHARACTERSET can be found with this select: select value from NLS_DATABASE_PARAMETERS where parameter='NLS_NCHAR_CHARACTERSET'; You cannot have more than 2 charactersets defined in Oracle: The NLS_CHARACTERSET is used for CHAR, VARCHAR2, CLOB columns; The NLS_NCHAR_CHARACTERSET is used for NCHAR, NVARCHAR2, NCLOB columns. NLS_NCHAR_CHARACTERSET is defined when the database is created and specified with the CREATE DATABASE command. The NLS_NCHAR_CHARACTERSET defaults to AL16UTF16 if nothing is specified. From 9i onwards the NLS_NCHAR_CHARACTERSET can have only 2 values: UTF8 or AL16UTF16 who are Unicode charactersets. See Note 260893.1 Unicode character sets in the Oracle database for more info about the difference between them. Al lot of people think that they *need* to use the NLS_NCHAR_CHARACTERSET to have UNICODE support in Oracle, this is not true, NLS_NCHAR_CHARACTERSET (NCHAR, NVARCHAR2) is in 9i always Unicode but you can perfectly use "normal" CHAR and VARCHAR2 columns for storing unicode in a database who has a AL32UTF8 / UTF8 NLS_CHARACTERSET. See also point 15. When trying to use another NATIONAL characterset, the CREATE DATABASE command will fail with "ORA-12714 invalid national character set specified". The character set identifier is stored with the column definition itself. 2) Which datatypes use the National Character Set? -------------------------------------------------- There are three datatypes which can store data in the national character set: NCHAR - a fixed-length national character set character string. The length of the column is ALWAYS defined in characters (it always uses CHAR semantics) NVARCHAR2 - a variable-length national character set character string. The length of the column is ALWAYS defined in characters (it always uses CHAR semantics) NCLOB - stores national character set data of up to four gigabytes. Data is always stored in UCS2 or AL16UTF16, even if the NLS_NCHAR_CHARACTERSET is UTF8. This has very limited impact, for more info about this please see: Note 258114.1 Possible action for CLOB/NCLOB storage after 10g upgrade and if you use DBMS_LOB.LOADFROMFILE see Note 267356.1 Character set conversion when using DBMS_LOB If you don't know what CHAR semantics is, then please read Note 144808.1 Examples and limits of BYTE and CHAR semantics usage If you use N-types, DO use the (N'...') syntax when coding it so that Literals are denoted as being in the national character set by prepending letter 'N', for example: create table test(a nvarchar2(100)); insert into test values(N'this is a NLS_NCHAR_CHARACTERSET string'); 3) How to know if I use N-type columns? --------------------------------------- This select list all tables containing a N-type column: select distinct OWNER, TABLE_NAME from DBA_TAB_COLUMNS where DATA_TYPE in ('NCHAR','NVARCHAR2', 'NCLOB'); On a 9i database created without (!) the "sample" shema you will see these rows (or less) returned: OWNER TABLE_NAME ------------------------------ ------------------------------ SYS ALL_REPPRIORITY SYS DBA_FGA_AUDIT_TRAIL SYS DBA_REPPRIORITY SYS DEFLOB SYS STREAMS$_DEF_PROC SYS USER_REPPRIORITY SYSTEM DEF$_LOB SYSTEM DEF$_TEMP$LOB SYSTEM REPCAT$_PRIORITY 9 rows selected. These SYS and SYSTEM tables may contain data if you are using: * Fine Grained Auditing -> DBA_FGA_AUDIT_TRAIL * Advanced Replication -> ALL_REPPRIORITY, DBA_REPPRIORITY, USER_REPPRIORITY DEF$_TEMP$LOB , DEF$_TEMP$LOB and REPCAT$_PRIORITY * Advanced Replication or Deferred Transactions functionality -> DEFLOB * Oracle Streams -> STREAMS$_DEF_PROC If you do have created the database with the DBCA and included the sample shema then you will see typically: OWNER TABLE_NAME ------------------------------------------------------------ OE BOMBAY_INVENTORY OE PRODUCTS OE PRODUCT_DESCRIPTIONS OE SYDNEY_INVENTORY OE TORONTO_INVENTORY PM PRINT_MEDIA SYS ALL_REPPRIORITY SYS DBA_FGA_AUDIT_TRAIL SYS DBA_REPPRIORITY SYS DEFLOB SYS STREAMS$_DEF_PROC SYS USER_REPPRIORITY SYSTEM DEF$_LOB SYSTEM DEF$_TEMP$LOB SYSTEM REPCAT$_PRIORITY 15 rows selected. The OE and PM tables contain just sample data and can be dropped if needed. 4) Should I worry when I upgrade from 8i or lower to 9i or 10g? --------------------------------------------------------------- * When upgrading from version 7: The National Character Set did not exist in version 7, so you cannot have N-type columns. Your database will just have the -default- AL16UTF16 NLS_NCHAR_CHARACTERSET declaration and the standard sys/system tables. So there is nothing to worry about... * When upgrading from version 8 and 8i: - If you have only the SYS / SYSTEM tables listed in point 3) then you don't have USER data using N-type columns. Your database will just have the -default- AL16UTF16 NLS_NCHAR_CHARACTERSET declaration after the upgrade and the standard sys/system tables. So there is nothing to worry about... We recommend that you follow this note: Note 159657.1 Complete Upgrade Checklist for Manual Upgrades from 8.X / 9.0.1 to Oracle9i - If you have more tables then the SYS / SYSTEM tables listed in point 3) (and they are also not the "sample" tables) then there are two possible cases: * Again, the next to points are *only* relevant when you DO have n-type USER data * a) Your current 8 / 8i NLS_NCHAR_CHARACTERSET is in this list: JA16SJISFIXED , JA16EUCFIXED , JA16DBCSFIXED , ZHT32TRISFIXED KO16KSC5601FIXED , KO16DBCSFIXED , US16TSTFIXED , ZHS16CGB231280FIXED ZHS16GBKFIXED , ZHS16DBCSFIXED , ZHT16DBCSFIXED , ZHT16BIG5FIXED ZHT32EUCFIXED Then the new NLS_NCHAR_CHARACTERSET will be AL16UTF16 and your data will be converted to AL16UTF16 during the upgrade. We recommend that you follow this note: Note 159657.1 Complete Upgrade Checklist for Manual Upgrades from 8.X / 9.0.1 to Oracle9i b) Your current 8 / 8i NLS_NCHAR_CHARACTERSET is UTF8: Then the new NLS_NCHAR_CHARACTERSET will be UTF8 and your data not be touched during the upgrade. We still recommend that you follow this note: Note 159657.1 Complete Upgrade Checklist for Manual Upgrades from 8.X / 9.0.1 to Oracle9i c) Your current 8 / 8i NLS_NCHAR_CHARACTERSET is NOT in the list of point a) and is NOT UTF8: Then your will need to export your data and drop it before upgrading. We recommend that you follow this note: Note 159657.1 Complete Upgrade Checklist for Manual Upgrades from 8.X / 9.0.1 to Oracle9i For more info about the National Character Set in Oracle8 see Note 62107.1 5) The NLS_NCHAR_CHARACTERSET is NOT changed to UTF8 or AL16UTF16 after upgrading to 9i. ---------------------------------------------------------------------------------------- That may happen if you have not set the ORA_NLS33 environment parameter correctly to the 9i Oracle_Home during the upgrade. Note 77442.1 ORA_NLS (ORA_NLS32, ORA_NLS33, ORA_NLS10) Environment Variables explained. We recommend that you follow this note for the upgrade: Note 159657.1 Complete Upgrade Checklist for Manual Upgrades from 8.X / 9.0.1 to Oracle9i Strongly consider to restore your backup and do the migration again or log a TAR, refer to this note and ask to assign the TAR to the NLS/globalization team. That team can then assist you further. However please do note that not all situations can be corrected, so you might be asked to do the migration again... 6) Can I change the AL16UTF16 to UTF8 / I hear that there are problems with AL16UTF16. -------------------------------------------------------------------------------------- a) If you do *not* use N-types then there is NO problem at all with AL16UTF16 because you are simply not using it and we strongly advice you the keep the default AL16UTF16 NLS_NCHAR_CHARACTERSET. b) If you *do* use N-types then there will be a problem with 8i clients and lower accessing the N-type columns (note that you will NOT have a problem selecting from "normal" non-N-type columns). More info about that is found there: Note 140014.1 ALERT Oracle8/8i to Oracle9i/10g using New "AL16UTF16" National Character Set Note 236231.1 New Character Sets Not Supported For Use With Developer 6i And Older Versions If this is a situation you find yourself in we recommend to simply use UTF8 as NLS_NCHAR_CHARACTERSET or create a second 9i db using UTF8 as NCHAR and use this as "inbetween" between the 8i and the 9i db you can create views in this new database that do a select from the AL16UTF16 9i db the data will then be converted from AL16UTF16 to UTF8 in the "inbetween" database and that can be read by oracle 8i This is one of the 2 reasons why you should use UTF8 as NLS_NCHAR_CHARACTERSET. If you are NOT using N-type columns with pre-9i clients then there is NO reason to go to UTF8. c) If you want to change to UTF8 because you are using transportable tablespaces from 8i database then check if are you using N-types in the 8i database that are included in the tablespaces that you are transporting. select distinct OWNER, TABLE_NAME from DBA_TAB_COLUMNS where DATA_TYPE in ('NCHAR','NVARCHAR2', 'NCLOB'); If yes, then you have the second reason to use UTF8 as as NLS_NCHAR_CHARACTERSET. If not, then leave it to AL16UTF16 and log a tar for the solution of the ORA-19736 and refer to this document. d) You are in one of the 2 situations where it's really needed to change from AL16UTF16 to UTF8, log a tar so that we can assist you. provide: 1) the output from: select distinct OWNER, TABLE_NAME, COLUMN_NAME, CHAR_LENGTH from DBA_TAB_COLUMNS where DATA_TYPE in ('NCHAR','NVARCHAR2', 'NCLOB'); 2) a CSSCAN output IMPORTANT: Please *DO* install the version 1.2 or higher from TechNet for you version. http://technet.oracle.com/software/tech/globalization/content.html and use this. copy all scripts and executables found in the zip file you downloaded to your oracle_home overwriting the old versions. Then run csminst.sql using these commands and SQL statements: cd $ORACLE_HOME/rdbms/admin set oracle_sid= sqlplus "sys as sysdba" SQL>set TERMOUT ON SQL>set ECHO ON SQL>spool csminst.log SQL> START csminst.sql Check the csminst.log for errors. Then run CSSCAN csscan FULL=Y FROMNCHAR=AL16UTF16 TONCHAR=UTF8 LOG=Ncharcheck CAPTURE=Y ( note the usage of fromNchar and toNchar ) Upload the 3 resulting files and the output of the select while creating the tar important: Do NOT use the N_SWITCH.SQL script, this will corrupt you NCHAR data !!!!!! 7) Is the AL32UTF8 problem the same as the AL16UTF16 / do I need the same patches? ---------------------------------------------------------------------------------- No, they may look similar but are 2 different issues. For information about the possible AL32UTF8 issue please see Note 237593.1 Problems connecting to AL32UTF8 databases from older versions (8i and lower) 8) But I still want as NLS_NCHAR_CHARACTERSET, like I had in 8(i)! --------------------------------------------------------------------------------- This is simply not possible. From 9i onwards the NLS_NCHAR_CHARACTERSET can have only 2 values: UTF8 or AL16UTF16. Both UTF8 and AL16UTF16 are unicode charactersets, so they can store whatever you had as NLS_NCHAR_CHARACTERSET in 8(i). If you are not using N-types then keep the default AL16UTF16 or use UTF8, it doesn't matter if you don't use the types. There is one condition in which this "limitation" can have a undisired affect, when you are importing an Oracle8i Transportable Tablespace into Oracle9i you can run into a ORA-19736 (as wel with AL16UTF16 as with UTF8). In that case log a TAR, refer to this note and ask to assign the TAR to the NLS/globalization team. That team can then assist you to work around this issue. 9) Do i need to set NLS_LANG to AL16UTF16 when creating/using the NLS_NCHAR_CHARACTERSET ? ------------------------------------------------------------------------------------------ As clearly stated in Note 158577.1 NLS_LANG Explained (How does Client-Server Character Conversion Work?) point "1.2 What is this NLS_LANG thing anyway?" * NLS_LANG is used to let Oracle know what characterset you client's OS is USING so that Oracle can do (if needed) conversion from the client's characterset to the database characterset. NLS_LANG is a CLIENT parameter has has no influance on the database side. 10) I try to use AL32UTF8 as NLS_NCHAR_CHARACTERSET but it fails with ORA-12714 ------------------------------------------------------------------------------- From 9i onwards the NLS_NCHAR_CHARACTERSET can have only 2 values: UTF8 or AL16UTF16. UTF8 is possible so that you can use it (when needed) for 8.x backwards compatibility. In all other conditions AL16UTF16 is the preferred and best value. AL16UTF16 has the same unicode revision as AL23UTF8, so there is no need for AL32UTF8 as NLS_NCHAR_CHARACTERSET. 11) I have the message "( possible ncharset conversion )" during import. ------------------------------------------------------------------------ in the import log you see something similar to this: Import: Release 9.2.0.4.0 - Production on Fri Jul 9 11:02:42 2004 Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved. Connected to: Oracle9i Enterprise Edition Release 9.2.0.4.0 - 64bit Production JServer Release 9.2.0.4.0 - Production Export file created by EXPORT:V08.01.07 via direct path import done in WE8ISO8859P1 character set and AL16UTF16 NCHAR character set export server uses WE8ISO8859P1 NCHAR character set (possible ncharset conversion) This is normal and is not a error condition. - If you do not use N-types then this is a pure informative message. - But even in the case that you use N-types like NCHAR or NCLOB then this is not a problem: * the database will convert from the "old" NCHAR characterset to the new one automatically. (and - unlike the "normal" characterset - the NLS_LANG has no impact on this conversion during exp/imp) * AL16UTF16 or UTF8 (the only 2 possible values in 9i) are unicode characterset and so can store any character... So no data loss is to be expected. 12) Can i use AL16UTF16 as NLS_CHARACTERSET ? ---------------------------------------------- No, AL16UTF16 can only be used as NLS_NCHAR_CHARACTERSET in 9i and above. Trying to create a database with a AL16UTF16 NLS_CHARACTERSET will fail. 13) I'm inserting in a Nchar or Nvarchar2 col but it comes back as ? or ? ... -------------------------------------------------------------------------------------------------- see point 13 in Note 227330.1 Character Sets & Conversion - Frequently Asked Questions 14) Do i need to change the NLS_NCHAR_CHARACTERSET in 8i to UTF8 BEFORE upgrading to 9i/10g? -------------------------------------------------------------------------------------------- No, see point 4) in this note. 15) Having a UTF8 NLS_CHARACTERSET db is there a advantage to use AL16UTF16 N-types ? ------------------------------------------------------------------------------------- there migth be 2 reasons: a) one possible advantage is storage (disk space). UTF8 uses 1 up to 3 bytes, AL16UTF16 always 2 bytes. If you have a lot of non-western data (cyrillic, Chinese, Japanese, Hindi languages..) then i can be advantageous to use N-types for those columns. For western data (english, french, spanish, dutch, german, portuguese etc...) UTF8 will use in most cases less disk space then AL16UTF16. Note 260893.1 Unicode character sets in the Oracle database This is not true for (N)CLOB, they are both encoded a internal fixed-width Unicode character set Note 258114.1 Possible action for CLOB/NCLOB storage after 10g upgrade so they will use the same amount of disk space. b) other possible advantage is extending the limits of CHAR semantics For a single-byte character set encoding, the character and byte length are the same. However, multi-byte character set encodings do not correspond to the bytes, making sizing the column more difficult. Hence the reason why CHAR semantics was introduced. However, we still have some physical underlying byte based limits and development has choosen to allow the full usage of the underlying limits. This results in the following table giving the maximum amount of CHARarcters occupying the MAX datalength that can be stored for a cer datatype in 9i and up. The MAX colum is the MAXIMUM amount of CHARACTERS that can be stored occupying the MAXIMUM data len seen that UTF8 and AL32UTF8 are VARRYING charactersets this means that a string of X chars can be X to X*3 (or X*4 for AL32) bytes. The MIN col is the maximum size that you can *define* and that Oracle can store if all data is the MINIMUM datalength (1 byte for AL32UTF8 and UTF8) for that characet. N-types (NVARCHAR2, NCHAR) are *always* defined in CHAR semantics, you cannot define them in BYTE. all numbers are CHAR definitions UTF8 (1 to 3 bytes) AL32UTF8 (1 to 4 bytes) AL16UTF16 ( 2 bytes) MIN MAX MIN MAX MIN MAX CHAR 2000 666 2000 500 N/A N/A VARCHAR2 4000 1333 4000 1000 N/A N/A NCHAR 2000 666 N/A N/A 1000 1000 NVARCHAR2 4000 1333 N/A N/A 2000 2000 (N/A means not possible) This means that if you try to store more then 666 characters that occupy 3 bytes in UTF8 in a CHAR UTF8 colum you still will get a ORA-01401: inserted value too large for column (or from 10g onwards: ORA-12899: value too large for column ) error, even if you have defined the colum as CHAR (2000 CHAR) so here it might be a good idea to define that column as NCHAR that will raise the MAX to 1000 char's ... Note 144808.1 Examples and limits of BYTE and CHAR semantics usage Disadvantages using N-types: * You might have some problems with older clients if using AL16UTF16 see point 6) b) in this note * Be sure that you use (AL32)UTF8 as NLS_CHARACTERSET , otherwise you will run into point 13 of this note. * Do not expect a higher *performance* by using AL16UTF16, it might be faster on some systems, but that has more to do with I/O then with the database kernel. * If you use N-types, DO use the (N'...') syntax when coding it so that Literals are denoted as being in the national character set by prepending letter 'N', for example: create table test(a nvarchar2(100)); insert into test values(N'this is NLS_NCHAR_CHARACTERSET string'); Normally you will choose to use VARCHAR (using a (AL32)UTF8 NLS_CHARACTERSET) for simplicity, to avoid confusion and possible other limitations who might be imposed by your application or programming language to the usage of N-types. 16) I have a message running DBUA (Database Upgrade Assistant) about NCHAR type when upgrading from 8i . AL16UTF16 The default Oracle character set for the SQL NCHAR data type, which is used for the national character set. It encodes Unicode data in the UTF-16 encoding. AL32UTF8 An Oracle character set for the SQL CHAR data type, which is used for the database character set. It encodes Unicode data in the UTF-8 encoding. Unicode Unicode is a universal encoded character set that allows you information from any language to be stored by using a single character set. Unicode provides a unique code value for every character, regardless of the platform, program, or language. Unicode database A database whose database character set is UTF-8. Unicode code point A 16-bit binary value that can represent a unit of encoded text for processing and interchange. Every point between U+0000 and U+FFFF is a code point. Unicode datatype A SQL NCHAR datatype (NCHAR, NVARCHAR2, and NCLOB). You can store Unicode characters in columns of these datatypes even if the database character set is not Unicode. unrestricted multilingual support The ability to use as many languages as desired. A universal character set, such as Unicode, helps to provide unrestricted multilingual support because it supports a very large character repertoire, encompassing most modern languages of the world. UTFE A Unicode 3.0 UTF-8 Oracle database character set with 6-byte supplementary character support. It is used only on EBCDIC platforms. UTF8 The UTF8 Oracle character set encodes characters in one, two, or three bytes. It is for ASCII-based platforms. The UTF8 character set supports Unicode 3.0. Although specific supplementary characters were not assigned code points in Unicode until version 3.1, the code point range was allocated for supplementary characters in Unicode 3.0. Supplementary characters are treated as two separate, user-defined characters that occupy 6 bytes. UTF-8 The 8-bit encoding of Unicode. It is a variable-width encoding. One Unicode character can be 1 byte, 2 bytes, 3 bytes, or 4 bytes in UTF-8 encoding. Characters from the European scripts are represented in either 1 or 2 bytes. Characters from most Asian scripts are represented in 3 bytes. Supplementary characters are represented in 4 bytes. UTF-16 The 16-bit encoding of Unicode. It is an extension of UCS-2 and supports the supplementary characters defined in Unicode 3.1 by using a pair of UCS-2 code points. One Unicode character can be 2 bytes or 4 bytes in UTF-16 encoding. Characters (including ASCII characters) from European scripts and most Asian scripts are represented in 2 bytes. Supplementary characters are represented in 4 bytes. wide character A fixed-width character format that is useful for extensive text processing because it allows data to be processed in consistent, fixed-width chunks. Wide characters are intended to support internal character processing Oracle started supporting Unicode based character sets in Oracle7. Here is a summary of the Unicode character sets supported in Oracle: +------------+---------+-----------------+ | Charset | RDBMS | Unicode version | +------------+---------+-----------------+ | AL24UTFFSS | 7.2-8.1 | 1.1 | | | | | | UTF8 | 8.0-10g | 2.1 (8.0-8.1.7) | | | | 3.0 (8.1.7-10g) | | | | | | UTFE | 8.0-10g | 2.1 (8.0-8.1.7) | | | | 3.0 (8.1.7-10g) | | | | | | AL32UTF8 | 9.0-10g | 3.0 (9.0) | | | | 3.1 (9.2) | | | | 3.2 (10.1) | | | | | | AL16UTF16 | 9.0-10g | 3.0 (9.0) | | | | 3.1 (9.2) | | | | 3.2 (10.1) | +------------+---------+-----------------+ AL24UTFFSS AL24UTFFSS was the first Unicode character set supported by Oracle. Is was introduced in Oracle 7.2. The AL24UTFFSS encoding scheme was based on the Unicode 1.1 standard, which is now obsolete. AL24UTFFSS has been de-supported from Oracle9i. The migration path for existing AL24UTFFSS databases is to upgrade the database to 8.0 or 8.1, then upgrade the character set to UTF8 before upgrading the database further to 9i or 10g. [NOTE:234381.1] Changing AL24UTFFSS to UTF8 - AL32UTF8 with ALTER DATABASE CHARACTERSET UTF8 UTF8 was the UTF-8 encoded character set in Oracle8 and 8i. It followed the Unicode 2.1 standard between Oracle 8.0 and 8.1.6, and was upgraded to Unicode version 3.0 for versions 8.1.7, 9i and 10g. To maintain compatibility with existing installations this character set will remain at Unicode 3.0 in future Oracle releases. Although specific supplementary characters were not assigned to Unicode until version 3.1, the allocation for these characters were already defined in 3.0. So if supplementary characters are inserted in a UTF8 database, it will not corrupt the actual data inside the database. They will be treated as 2 separate undefined characters, occupying 6 bytes in storage. We recommend that customers switch to AL32UTF8 for full supplementary character support. UTFE This is the UTF8 database character set for the EDCDIC platforms. It has the same properties as UTF8 on ASCII based platforms. The EBCDIC Unicode transformation format is documented in Unicode Technical Report #16 UTF-EBCDIC. Which can be found at http://www.unicode.org/unicode/reports/tr16/ AL32UTF8 This is the UTF-8 encoded character set introduced in Oracle9i. AL32UTF8 is the database character set that supports the latest version (3.2 in 10g) of the Unicode standard. It also provides support for the newly defined supplementary characters. All supplementary characters are stored as 4 bytes. AL32UTF8 was introduced because when UTF8 was designed (in the times of Oracle8) there was no concept of supplementary characters, therefore UTF8 has a maximum of 3 bytes per character. Changing the design of UTF8 would break backward compatibility, so a new character set was introduced. The introduction of surrogate pairs should mean that no significant architecture changes are needed in future versions of the Unicode standard, so the plan is to keep enhancing AL32UTF8 as necessary to support future version of the Unicode standard, for example work is now underway to make sure we support Unicode 4.0 in AL32UTF8 in the release after 10.1. AL16UTF16 This is the first UTF-16 encoded character set in Oracle. It was introduced in Oracle9i as the default national character set (NLS_NCHAR_CHARACTERSET). AL16UTF16 supports the latest version (3.2 in 10g) of the Unicode standard. It also provides support for the newly defined supplementary characters. All supplementary characters are stored as 4 bytes. As with AL32UTF8, the plan is to keep enhancing AL16UTF16 as necessary to support future version of the Unicode standard. AL16UTF16 cannot be used as a database character set (NLS_CHARACTERSET), only as the national character set (NLS_NCHAR_CHARACTERSET). The database character set is used to identify and to hold SQL, SQL metadata and PL/SQL source code. It must have either single byte 7-bit ASCII or single byte EBCDIC as a subset, whichever is native to the deployment platform. Therefore, it is not possible to use a fixed-width, multi-byte character set (such as AL16UTF16) as the database character set. Trying to create a database with AL16UTF16 a characterset in 9i and up will give "ORA-12706: THIS CREATE DATABASE CHARACTER SET IS NOT ALLOWED". Further reading --------------- All the above information is taken from the white paper "Oracle Unicode database support". The paper itself contains much more information and is available from: http://otn.oracle.com/tech/globalization/pdf/TWP_Unicode_10gR1.pdf References ---------- The following URLs contain a complete list of hex values and character descriptions for every Unicode character: Unicode Version 3.2: http://www.unicode.org/Public/3.2-Update/UnicodeData-3.2.0.txt Unicode Version 3.1: http://www.unicode.org/Public/3.1-Update/UnicodeData-3.1.0.txt Unicode Version 3.0: http://www.unicode.org/Public/3.0-Update/UnicodeData-3.0.0.txt Unicode Versions 2.x: http://www.unicode.org/unicode/standard/versions/enumeratedversions.html Unicode Version 1.1: http://www.unicode.org/Public/1.1-Update/UnicodeData-1.1.5.txt A description of the file format can be found at: http://www.unicode.org/Public/UNIDATA/UnicodeData.html For a glossarry of unicode terms, see: http://www.unicode.org/glossary/ On above locations you can find the unicode standard, all characters are there referenced with their UCS-2 codepoint Some further notes: =================== Note 1: ------- Thanks for the detailed reply. > > >Furthermore the use of NLS columns on a utf8 database (al32utf8 would be > better by the way) is > >subject to questions. Correct me if I'm wrong but I believe that most > >asian character sets can be translated into utf8 without loosing any > >information. The only exception to this statement is for surrogate pairs > >and that's the only difference between al32utf8 and utf8 in Oracle. > >al32utf8 supports surrogate pairs. > > I found from Oracle documentation that UTF8 supports surrogate pairs but > requires 6 bytes for surrogate pairs. I should have clarified : the jdbc drivers don't support these 6-bytes utf8 surrogate pairs. That's the reason why we introduced al32utf8 as one of the native character set (ascii, isolatin1, utf8, al32utf8, ucs2, al24utffss). Note 2: ------- > AL32UTF8 > The AL32UTF8 character set encodes characters in one to three bytes. > Surrogate > pairs require four bytes. It is for ASCII-based platforms. > > UTF8 > The UTF8 character set encodes characters in one to three bytes. Surrogate > pairs > require six bytes. It is for ASCII-based platforms. > > AL32UTF8 > --------- > Advantages > ---------- > 1. Surrogate pair Unicode characters > are stored in the standard 4 bytes > representation, and there is no > data conversion upon retrieval > and insertion of those surrogate > characters. Also, the storage for > those characters requires less disk > space than that of the same > characters encoded in UTF8. > > Disadvantages > ------------- > 1. You cannot specify the length of SQL CHAR > types in the number of characters (Unicode > code points) for surrogate characters. For > example, surrogate characters are treated as > one code point rather than the standard of two > code points. > 2. The binary order for SQL CHAR columns is > different from that of SQL NCHAR columns > when the data consists of surrogate pair > Unicode characters. As a result, CHAR columns > NCHAR columns do not always have the same > sort for identical strings. > > UTF8 > ---- > Advantages > ---------- > 1. You can specify the length of SQL > CHAR types as a number of > characters. > 2. The binary order on the SQL CHAR > columns is always the same as > that of the SQL NCHAR columns > when the data consists of the same > surrogate pair Unicode characters. > As a result, CHAR columns and > NCHAR columns have the same > sort for identical strings. > > Disadvantages > ------------- > 1. Surrogate pair Unicode characters are stored > as 6 bytes instead of the 4 bytes defined by the > Unicode standard. As a result, Oracle has to > convert data for those surrogate characters. > > I dont understand the 1st disadvantage of AL32UTF8 encoding !! If surrogate > characters are considered 1 codepoint, then if I declare a CHAR column as of > length 40 characters (codepoints) , then I can enter 40 surrogate > characters. Note 3: ------- Universal Character Sets ==================== Character Set Name Description Comments Language, Country or Region ================= ===================================== ========= ========================== AL16UTF16 Unicode 3.1 UTF-16Universal character set MB, EURO, FIXED Universal Unicode AL32UTF8 Unicode 3.1 UTF-8 Universal character set MB, ASCII, EURO Universal Unicode UTF8 Unicode 3.0 UTF-8 Universal character set MB, ASCII, EURO Universal Unicode CESU-8 compliant UTFE EBCDIC form of Unicode 3.0UTF-8 MB, EURO Universal Unicode Universal character set Note 4: ------- WE8ISO is a single byte character set. It has 255 characters. Korean data requires a multi-byte character set -- each character could be 1, 2, 3 or more bytes. It is a variable length encoding scheme. It has more then, way more then 255 characters. I don't see it fitting into we8iso unless they use RAW in which case it is just bytes, not characters at all. Note 5: ------- Hi Tom, We migrated our DB 8.1.7 to 9.2.In 8.1.7 we used UTF8 character set.It remains same in 9.2. We know that Oracle 9.2 doesn't have UTF8 but AL32UTF8. Can we keep this UTF8 or have to change to AL32UTF8. If we need to change, may we do it by : alter database character set AL32UTF8 or we must use exp/imp utility? Regards Followup: what do you mean -- utf8 is still a valid character set? Note 6: ------- Hi Tom, We are migrating from oracle 8.1.6 to oracle 9 R2. We have about 14 oracle instance. All instances have WE8ISO88591P1 character set. Our company is expanding globally so we are thinking to use unicode character set with oracle 9. I have few questions on this issue. 1) What is the difference between UTF-8,UTF-16 Is AL32UTF8 and UTF-8 is same character set or they are different? Is UTF-16 and AL16UTF16 is same character set or different ? 2) Which character is super set of all character set? If there is any, Does oracle support that character set? 3) Do we have to change our pl/sql procedure if we move to unicode database ? The reason for this question is our developer is using ascii character for carrage return and line feed like chr(10) and chr(13) and some other ascii character . 4) What is impact on CLOB ? 5) What will be the size of the database? Our production DB size is currently 50GB. What it would be in unicode? Thanks basically utf8 is unicode 3.0 support, utf16 is unicode 3.1 there is no super super "top" set. Your plsql routines may will have to change -- your data model may well have to change. You'll find that in utf, european characters (except ascii -- 7bit data) all take 2 bytes. That varchar2(80) you have in your database? It might only hold 40 characters of eurpean data (or even less of other kinds of data). It is 80 bytes (you can use the new 9i syntax varchar2( N char ) -- it'll allocate in characters, not bytes). So, you could find your 80 character description field cannot hold 80 characters. You might find that x := a || b; fails -- with string to long in your plsql code due to the increased size. You might find that your string intensive routines run slower (substr(x,1,80) is no longer byte 1 .. byte 80 -- Oracle has to look through the string to find where characters start and stop -- it is more complex) chr(10) and chr(13) should work find, they are simple ASCII. On clob -- same impact as on varchar2, same issues. Your database could balloon to 200gb, but it will be somewhere between 50 and 200. As unicode is a VARYING WIDTH encoding scheme, it is impossible to be precise -- it is not a fixed width scheme, so we don't know how big your strings will get to be. 21.3 Oracle Rowid's ------------------- Rowid's: Every table row has an internal rowid which contains information about object_id, block_id, file#. Also you can query on the "logical" number rownum. SQL> SELECT * FROM charlie.xyz; ID NAME --------- -------------------- 1 joop 2 gerrit SQL> SELECT rownum FROM charlie.xyz; ROWNUM --------- 1 2 SQL> SELECT rowid FROM SALES.xyz; ROWID ------------------ AAAI92AAQAAAFXbAAA AAAI92AAQAAAFXbAAB - DBMS_ROWID: DBMS_ROWID. Every row has a rowid. Every row has also an associated logical "rownum" on which you can query. The rowid is an 18 byte structure that stores the location of blockid WHERE the row is in. The old format is the restricted format of Oracle 7 The new format is the extended format of Oracle 8, 8i format: OOOOOOFFFBBBBBRRRR 000000=object_id FFF=relative datafile number BBBBB=block_id RRR=row in block The dbms package DBMS_ROWID has several function to convert FROM the one format to the other. DBMS_ROWID EXAMPLES: -------------------- SELECT DBMS_ROWID.ROWID_TO_EXTENDED(ROWID,null,null,0), DBMS_ROWID.ROWID_TO_RESTRICTED(ROWID,0), rownum FROM CHARLIE.XYZ; SELECT dbms_rowid.rowid_block_number(rowid) FROM emp WHERE ename = 'KING'; SELECT dbms_rowid.rowid_block_number(rowid) FROM TCMLOGDBUSER.EVENTLOG WHERE id = 5; This example returns the ROWID for a row in the EMP table, extracts the data object number FROM the ROWID, using the ROWID_OBJECT function in the DBMS_ROWID package, then displays the object number: DECLARE object_no INTEGER; row_id ROWID; BEGIN SELECT ROWID INTO row_id FROM TCMLOGDBUSER.EVENTLOG WHERE id=5; object_no := dbms_rowid.rowid_object(row_id); dbms_output.put_line('The obj. # is '|| object_no); END; / PL/SQL procedure successfully completed. SQL> set serveroutput on SQL> / The obj. # is 28954 PL/SQL procedure successfully completed. SQL> select * from dba_objects where object_id=28954; OWNER ------------------------------ OBJECT_NAME ----------------------------------------------------------- SUBOBJECT_NAME OBJECT_ID DATA_OBJECT_ID ------------------------------ ---------- -------------- OBJECT_TYPE CREATED LAST_DDL_ TIMESTAMP ------------------ --------- --------- ------------------- STATUS T G S ------- - - - TCMLOGDBUSER EVENTLOG 28954 28954 TABLE 05-DEC-04 05-DEC-04 2004-12-05:22:26:10 VALID N N N 21.4 HETEROGENEOUS SERVICES: ---------------------------- Generic connectivity is intended for low-end data integration solutions requiring the ad hoc query capability to connect from Oracle8i to non-Oracle database systems. Generic connectivity is enabled by Oracle Heterogeneous Services, allowing you to connect to non-Oracle systems with improved performance and throughput. Generic connectivity is implemented as a Heterogeneous Services ODBC agent. An ODBC agent is included as part of your Oracle8i system. To access the non-Oracle data store using generic connectivity, the agent works with an ODBC driver. Oracle8i provides support for the ODBC driver interface. The driver that you use must be on the same machine as the agent. The non-Oracle data stores can reside on the same machine as Oracle8i or a different machine. Agent processes are usually started when a user session makes its first non-Oracle system access through a database link. These connections are made using Oracle's remote data access software, Oracle Net Services, which enables both client-server and server-server communication. The agent process continues to run until the user session is disconnected or the database link is explicitly closed. Multithreaded agents behave slightly differently. They have to be explicitly started and shut down by a database administrator instead of automatically being spawned by Oracle Net Services. Oracle has Generic Connectivity agents for ODBC and OLE DB that enable you to use ODBE and OLEDB drivers to access non-Oracle systems that have an ODBC or an OLE DB interface. Setup: ------ 1. HS datadictonary ------------------- To install the data dictionary tables and views for Heterogeneous Services, you must run a script that creates all the Heterogeneous Services data dictionary tables, views, and packages. On most systems the script is called caths.sql and resides in $ORACLE_HOME/rdbms/admin. Check for the existence of Heterogeneous Services data dictionary views, All normal standard preparations for HS needs to be in place in Oracle 9i. To recap this here, if you must install HS from scratch: - run caths.sql as SYS on Ora9i DB Server. - The HS Agent will be installed as part of 9i DB install. It will be started as part of the listener. - On NT/2000, The agent works with a OLEDB or ODBC driver to connect to target db - The DB Server will connect to the agent through NET8, which is why a tnsnames.ora and a listener.ora entry needs to be setup You van also check on HS installation. Just check on existence of the HS% views in the SYS schema, for example, SYS.HS_FDS_CLASS. 2. tnsnames.ora and listener.ora -------------------------------- To initiate a connection to the non-Oracle system, the Oracle9i server starts an agent process through the Oracle Net listener. For the Oracle9i server to be able to connect to the agent, you must configure tnsnames.ora and listener.ora ------------------------------------------------------------------------------ tnsnames examples: Sybase_sales= (DESCRIPTION= (ADDRESS=(PROTOCOL=tcp) (HOST=dlsun206) -- local machine (PORT=1521) ) (CONNECT_DATA = (SERVICE_NAME=SalesDB) ) (HS = OK) ) TNSNAMES.ORA hsmsql = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = tcp)(host=winhost)(port=1521)) ) -- local machine (CONNECT_DATA = (SID = msql) ) -- needs to match the sid in listener.ora. (HS=OK) ) ) TG4MSQL.WORLD = (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = ukp15340)(PORT = 1528) ) (CONNECT_DATA = (SID = tg4msql) ) (HS = OK) ) ------------------------------------------------------------------------------- listener.ora examples: LISTENER = (ADDRESS_LIST = (ADDRESS= (PROTOCOL=tcp) (HOST = dlsun206) (PORT = 1521) ) ) ... SID_LIST_LISTENER = (SID_LIST = (SID_DESC = (SID_NAME=SalesDB) (ORACLE_HOME=/home/oracle/megabase/9.0.1) (PROGRAM=tg4mb80) (ENVS=LD_LIBRARY_PATH=non_oracle_system_lib_directory) ) ) LISTENER.ORA LISTENER = (DESCRIPTION_LIST = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = winhost)(PORT = 1521)) ) ) SID_LIST_LISTENER = (SID_LIST = (SID_DESC = (SID_NAME = msql) <== needs to match the sid in tnsnames.ora (ORACLE_HOME = E:\Ora816) (PROGRAM = hsodbc) <== hsodbc is the executable ) ) 3. create the initialization file: ---------------------------------- Create the Initialization file. Oracle supplies a sample initialization file named "inithsodbc.ora" which is stored in the $ORACLE_HOME\hs\admin directory. To create an initialization file, copy the appropriate sample file and rename the file to initHS_SID.ora. In this example the sid noted in the listener and tnsnames is msql so our new initialization file is called initmsql.ora. INITMSQL.ORA # HS init parameters # HS_FDS_CONNECT_INFO = msql <= odbc data_source_name HS_FDS_TRACE_LEVEL = 0 <= trace levels 0 - 4 (4 is verbose) HS_FDS_TRACE_FILE_NAME = hsmsql.trc <= trace file name # # Environment variables required for the non-Oracle system # #set = HS_FDS_SHAREABLE_NAME Default value: none Range of values: not applicable HS_FDS_SHAREABLE_NAME: Specifies the full path name to the ODBC library. This parameter is required when you are using generic connectivity to access data from an ODBC provider on a UNIX machine. 4. create a database link: -------------------------- CREATE DATABASE LINK sales USING `Sybase_sales'; Common Errors: -------------- AGTCTL.exe = ORA-28591 unable to access parameter file, ORA-28592 agent SID not set agentctl hsodbc.exe = caths.sql What is the difference between agtctl and lsnrctl dbsnmp_start Error: ORA-28591 Text: agent control utility: unable to access parameter file --------------------------------------------------------------------------- Cause: The agent control utility was unable to access its parameter file. This could be because it could not find its admin directory or because permissions on directory were not correctly set. Action: The agent control utility puts its parameter file in either the directory pointed to by the environment variable AGTCTL_ADMIN or in the directory pointed to by the environment variable TNS_ADMIN. Make sure that at least one of these environment variables is set and that it points to a directory that the agent has access to. SET AGTCTL_ADMIN=\OPT\ORACLE\ORA81\HS\ADMIN Error: ORA-28592 Text: agent control utility: agent SID not set --------------------------------------------------------------------------- Cause: The agent needs to know the value of the AGENT_SID parameter before it can process any commands. If it does not have a value for AGENT_SID then all commands will fail. Action: Issue the command SET AGENT_SID and then retry the command that failed. Error: ------ fix: Set the HS_FDS_TRACE_FILE_NAME to a filename: HS_FDS_TRACE_FILE_NAME = test.log or comment it out: #HS_FDS_TRACE_FILE_NAME Error: incorrect characters ------ Change the HS_LANGUAGE to a correct NLS like AMERICAN_AMERICA.WE8MSWIN1252 Error: ORA-02085 ---------------- HS_FDS_CONNECT_INFO = HS_FDS_TRACE_LEVEL = 0 HS_FDS_TRACE_FILE_NAME = c:\hs.log HS_DB_NAME = exhsodbc -- case sensitive HS_DB_DOMAIN = ch.oracle.com -- case sensitive ERROR: ORA-02085 ---------------- SET GLOBAL_NAMES TRUE ERORR:ORA-02068 and ORA-28511 ----------------------------- LD_LIBRARY_PATH=/u06/home/oracle/support/network/ODBC/lib f the LD_LIBRARY_PATH does not contain the path to the ODBC library, a dd the ODBC library path and start the listener with this environment. LD_LIBRARY_PATH=/u01/app/oracle/product/8.1.7/lib; export LD_LIBRARY_PATH When the listener launches the agent hsodbc, the agent inherits the environment from the listener and needs to have the ODBC library path in order to access the ODBC shareable file. The shareable file is defined in the init.ora file located in the $ORACLE_HOME/hs/admin directory. HS_FDS_SHAREABLE_NAME=/u06/home/oracle/support/network/ODBC/lib/libodbc.so 21.5 SET EVENTS: ---------------- Note 1: ------- - What is a database EVENT and how does one set it? Oracle trace events are useful for debugging the Oracle database server. The following two examples are simply to demonstrate syntax. Refer to later notes on this page for an explanation of what these particular events do. Events can be activated by either adding them to the INIT.ORA parameter file. E.g. event='1401 trace name errorstack, level 12' ... or, by issuing an ALTER SESSION SET EVENTS command: E.g. alter session set events '10046 trace name context forever, level 4'; The alter session method only affects the user's current session, whereas changes to the INIT.ORA file will affect all sessions once the database has been restarted. - What database events can be set? The following events are frequently used by DBAs and Oracle Support to diagnose problems: 10046 trace name context forever, level 4 Trace SQL statements and show bind variables in trace output. 10046 trace name context forever, level 8 This shows wait events in the SQL trace files 10046 trace name context forever, level 12 This shows both bind variable names and wait events in the SQL trace files 1401 trace name errorstack, level 12 1401 trace name errorstack, level 4 1401 trace name processstate Dumps out trace information if an ORA-1401 "inserted value too large for column" error occurs. The 1401 can be replaced by any other Oracle Server error code that you want to trace. 60 trace name errorstack level 10 Show where in the code Oracle gets a deadlock (ORA-60), and may help to diagnose the problem. - The following list of events are examples only. They might be version specific, so please call Oracle before using them: 10210 trace name context forever, level 10 10211 trace name context forever, level 10 10231 trace name context forever, level 10 These events prevent database block corruptions 10049 trace name context forever, level 2 Memory protect cursor 10210 trace name context forever, level 2 Data block check 10211 trace name context forever, level 2 Index block check 10235 trace name context forever, level 1 Memory heap check 10262 trace name context forever, level 300 Allow 300 bytes memory leak for connections - How can one dump internal database structures? The following (mostly undocumented) commands can be used to obtain information about internal database structures. -- Dump control file contents alter session set events 'immediate trace name CONTROLF level 10' / -- Dump file headers alter session set events 'immediate trace name FILE_HDRS level 10' / -- Dump redo log headers alter session set events 'immediate trace name REDOHDR level 10' / -- Dump the system state -- NOTE: Take 3 successive SYSTEMSTATE dumps, with 10 minute intervals alter session set events 'immediate trace name SYSTEMSTATE level 10' / -- Dump the process state alter session set events 'immediate trace name PROCESSSTATE level 10' / -- Dump Library Cache details alter session set events 'immediate trace name library_cache level 10' / -- Dump optimizer statistics whenever a SQL statement is parsed (hint: change statement or flush pool) alter session set events '10053 trace name context forever, level 1' / -- Dump a database block (File/ Block must be converted to DBA address) -- Convert file and block number to a DBA (database block address). Eg: variable x varchar2; exec :x := dbms_utility.make_data_block_address(1,12); print x alter session set events 'immediate trace name blockdump level 50360894' / ALTER SESSION SET EVENTS '1652 trace name errorstack level 1 '; or alter system set events '1652 trace name errorstack level 1 '; alter system set events '1652 trace name errorstack off '; Note 2: ------- Doc ID : Note:218105.1 Content Type: TEXT/PLAIN Subject: Introduction to ORACLE Diagnostic EVENTS Creation Date: 11-NOV-2002 Type: BULLETIN Last Revision Date: 20-NOV-2002 Status: PUBLISHED PURPOSE ------- This document describes the different types of Oracle EVENT that exist to help customers and Oracle Support Services when investigating Oracle RDBMS related issues. This note will only provide information of a general nature. Specific information on the usage of a given event should be provided by Oracle Support Services or the Support related article that is suggesting the use of a given event. This note will not provide that level of detail. SCOPE & APPLICATION ------------------- The information held here is of use to Oracle DBAs, developers and Oracle Support Services. Introduction to ORACLE Diagnostic EVENTS ---------------------------------------- Before proceeding, please review the following note as it contain some important additional information on Events. [NOTE:75713.1] "Important Customer information about using Numeric Events" EVENTS are primarily used to produce additional diagnostic information when insufficient information is available to resolve a given problem. EVENTS are also used to workaround or resolve problems by changing Oracle's behaviour or enabling undocumented features. *WARNING* Do not use an Oracle Diagnostic Event unless directed to do so by Oracle Support Services or via a Support related article on Metalink. Incorrect usage can result in disruptions to the database services. Setting EVENTS -------------- There are a number of ways in which events can be set. How you set an event depends on the nature of the event and the circumstances at the time. As stated above, specific information on how you set a given event should be provided by Oracle Support Services or the Support related article that is suggesting the use of a given event. Most events can be set using more than one of the following methods : o As INIT parameters o In the current session o From another session using a Debug tool INIT Parameters ~~~~~~~~~~~~~~~ Syntax: EVENT = " " Reference: [NOTE:160178.1] How to set EVENTS in the SPFILE Current Session ~~~~~~~~~~~~~~~ Syntax: ALTER SESSION SET EVENTS ' '; From another Session using a Debug tool ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There are a number of debug tools : o ORADEBUG o ORAMBX (VMS only) ORADEBUG : ======== Syntax: Prior to Oracle 9i, SVRMGR> oradebug event Oracle 9i and above : SQL> oradebug event Reference: [NOTE:29786.1] "SUPTOOL: ORADEBUG 7.3+ (Server Manager/SQLPLUS Debug Commands)" [NOTE:1058210.6] "HOW TO ENABLE SQL TRACE FOR ANOTHER SESSION USING ORADEBUG" ORAMBX : on OpenVMS is still available and described under : ====== [NOTE:29062.1] "SUPTOOL: ORAMBX (VMS) - Quick Reference" This note will not enter into additional details on these tools. EVENT Categories ---------------- The most commonly used events fall into one of four categories : o Dump diagnostic information on request o Dump diagnostic information when an error occurs o Change Oracle's behaviour o Produce trace diagnostic information as the instance runs Dump diagnostic information on request (Immediate Dump) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ An immediate dump Event will result in information immediately being written to a trace file. Some common immediate dump Events include : SYSTEMSTATE, ERRORSTACK, CONTROLF, FILE_HDRS and REDOHDR These type of events are typically set in the current session. For example: ALTER SESSION SET EVENTS 'IMMEDIATE trace name ERRORSTACK level 3'; Dump Diagnostic information when an error occurs (On-Error Dump) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The on-error dump Event is similar to the immediate dump Event with the difference being that the trace output is only produced when the given error occurs. You can use virtually any standard Oracle error to trigger this type of event. For example, an ORA-942 "table or view does not exist" error does not include the name of the problem table or view. When this is not obvious from the application (due to its complexity), then it can be difficult to investigate the source of the problem. However, an On-Error dump against the 942 error can help narrow the search. These type of events are typically set as INIT parameters. For example, using the 942 error : EVENT "942 trace name ERRORSTACK level 3" Once established, the next time a session encounters an ORA-942 error, a trace file will be produced that shows (amongst other information) the current SQL statement being executed. This current SQL can now be checked and the offending table or view more easily discovered. Change Oracle's behaviour ~~~~~~~~~~~~~~~~~~~~~~~~~ Instance behaviour can be changed or hidden features can be enabled using these type of Event A common event in this category is 10262 which is discussed in [NOTE:21235.1] EVENT: 10262 "Do not check for memory leaks" These type of events are typically set as INIT parameters. For example: EVENT "10262 trace name context forever, level 4000" Produce trace diagnostic information as the instance runs (Trace Events) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Trace events produce diagnostic information as processes are running. They are used to gather additional information about a problem. A common event in this category is 10046 which is discussed in [NOTE:21154.1] EVENT: 10046 "enable SQL statement tracing (including binds/waits)" These type of events are typically set as INIT parameters. For example: EVENT = "10046 trace name context forever, level 12" Summary ------- EVENT usage and syntax can be very complex and due to the possible impact on the database, great care should be taken when dealing with them. Oracle Support Services (or a Support article) should provide information on the appropriate method to be adopted and syntax to be used when establishing a given event. If it is possible to do so, test an event against a development system prior to doing the same thing on a production system. The misuse of events can lead to a loss of service. RELATED DOCUMENTS ----------------- [NOTE:75713.1] Important Customer information about using Numeric Events [NOTE:21235.1] EVENT: 10262 "Do not check for memory leaks" [NOTE:21154.1] EVENT: 10046 "enable SQL statement tracing (including binds/waits)" [NOTE:160178.1] How to set EVENTS in the SPFILE [NOTE:1058210.6] HOW TO ENABLE SQL TRACE FOR ANOTHER SESSION USING ORADEBUG [NOTE:29786.1] SUPTOOL: ORADEBUG 7.3+ (Server Manager/SQLPLUS Debug Commands) [NOTE:29062.1] SUPTOOL: ORAMBX (VMS) - Quick Reference ====================== 22. DBA% and v$ views ====================== NLS: ---- VIEW_NAME OWNER ------------------------------ ------------------------------ NLS_DATABASE_PARAMETERS SYS NLS_INSTANCE_PARAMETERS SYS NLS_SESSION_PARAMETERS SYS DBA: ---- VIEW_NAME OWNER ------------------------------ ------------------------------ DBA_2PC_NEIGHBORS SYS DBA_2PC_PENDING SYS DBA_ALL_TABLES SYS DBA_ANALYZE_OBJECTS SYS DBA_ASSOCIATIONS SYS DBA_AUDIT_EXISTS SYS DBA_AUDIT_OBJECT SYS DBA_AUDIT_SESSION SYS DBA_AUDIT_STATEMENT SYS DBA_AUDIT_TRAIL SYS DBA_CACHEABLE_OBJECTS SYS DBA_CACHEABLE_TABLES SYS DBA_CACHEABLE_TABLES_BASE SYS DBA_CATALOG SYS DBA_CLUSTERS SYS DBA_CLUSTER_HASH_EXPRESSIONS SYS DBA_CLU_COLUMNS SYS DBA_COLL_TYPES SYS DBA_COL_COMMENTS SYS DBA_COL_PRIVS SYS DBA_CONSTRAINTS SYS DBA_CONS_COLUMNS SYS DBA_CONTEXT SYS DBA_DATA_FILES SYS DBA_DB_LINKS SYS DBA_DEPENDENCIES SYS DBA_DIMENSIONS SYS DBA_DIM_ATTRIBUTES SYS DBA_DIM_CHILD_OF SYS DBA_DIM_HIERARCHIES SYS DBA_DIM_JOIN_KEY SYS DBA_DIM_LEVELS SYS DBA_DIM_LEVEL_KEY SYS DBA_DIRECTORIES SYS DBA_DMT_FREE_SPACE SYS DBA_DMT_USED_EXTENTS SYS DBA_ERRORS SYS DBA_EXP_FILES SYS DBA_EXP_OBJECTS SYS DBA_EXP_VERSION SYS DBA_EXTENTS SYS DBA_FREE_SPACE SYS DBA_FREE_SPACE_COALESCED SYS DBA_FREE_SPACE_COALESCED_TMP1 SYS DBA_FREE_SPACE_COALESCED_TMP2 SYS DBA_FREE_SPACE_COALESCED_TMP3 SYS DBA_IAS_CONSTRAINT_EXP SYS DBA_IAS_GEN_STMTS SYS DBA_IAS_GEN_STMTS_EXP SYS DBA_IAS_OBJECTS SYS DBA_IAS_OBJECTS_BASE SYS DBA_IAS_OBJECTS_EXP SYS DBA_IAS_POSTGEN_STMTS SYS DBA_IAS_PREGEN_STMTS SYS DBA_IAS_SITES SYS DBA_IAS_TEMPLATES SYS DBA_INDEXES SYS DBA_INDEXTYPES SYS DBA_INDEXTYPE_OPERATORS SYS DBA_IND_COLUMNS SYS DBA_IND_EXPRESSIONS SYS DBA_IND_PARTITIONS SYS DBA_IND_SUBPARTITIONS SYS DBA_INTERNAL_TRIGGERS SYS DBA_JAVA_POLICY SYS DBA_JOBS SYS DBA_JOBS_RUNNING SYS DBA_LIBRARIES SYS DBA_LMT_FREE_SPACE SYS DBA_LMT_USED_EXTENTS SYS DBA_LOBS SYS DBA_LOB_PARTITIONS SYS DBA_LOB_SUBPARTITIONS SYS DBA_METHOD_PARAMS SYS DBA_METHOD_RESULTS SYS DBA_MVIEWS SYS DBA_MVIEW_AGGREGATES SYS DBA_MVIEW_ANALYSIS SYS DBA_MVIEW_DETAIL_RELATIONS SYS DBA_MVIEW_JOINS SYS DBA_MVIEW_KEYS SYS DBA_NESTED_TABLES SYS DBA_OBJECTS SYS DBA_OBJECT_SIZE SYS DBA_OBJECT_TABLES SYS DBA_OBJ_AUDIT_OPTS SYS DBA_OPANCILLARY SYS DBA_OPARGUMENTS SYS DBA_OPBINDINGS SYS DBA_OPERATORS SYS DBA_OUTLINES SYS DBA_OUTLINE_HINTS SYS DBA_PARTIAL_DROP_TABS SYS DBA_PART_COL_STATISTICS SYS DBA_PART_HISTOGRAMS SYS DBA_PART_INDEXES SYS DBA_PART_KEY_COLUMNS SYS DBA_PART_LOBS SYS DBA_PART_TABLES SYS DBA_PENDING_TRANSACTIONS SYS DBA_POLICIES SYS DBA_PRIV_AUDIT_OPTS SYS DBA_PROFILES SYS DBA_QUEUES SYS DBA_QUEUE_SCHEDULES SYS DBA_QUEUE_TABLES SYS DBA_RCHILD SYS DBA_REFRESH SYS DBA_REFRESH_CHILDREN SYS DBA_REFS SYS DBA_REGISTERED_SNAPSHOTS SYS DBA_REGISTERED_SNAPSHOT_GROUPS SYS DBA_REPAUDIT_ATTRIBUTE SYS DBA_REPAUDIT_COLUMN SYS DBA_REPCAT SYS DBA_REPCATLOG SYS DBA_REPCAT_REFRESH_TEMPLATES SYS DBA_REPCAT_TEMPLATE_OBJECTS SYS DBA_REPCAT_TEMPLATE_PARMS SYS DBA_REPCAT_TEMPLATE_SITES SYS DBA_REPCAT_USER_AUTHORIZATIONS SYS DBA_REPCAT_USER_PARM_VALUES SYS DBA_REPCOLUMN SYS DBA_REPCOLUMN_GROUP SYS DBA_REPCONFLICT SYS DBA_REPDDL SYS DBA_REPFLAVORS SYS DBA_REPFLAVOR_COLUMNS SYS DBA_REPFLAVOR_OBJECTS SYS DBA_REPGENERATED SYS DBA_REPGENOBJECTS SYS DBA_REPGROUP SYS DBA_REPGROUPED_COLUMN SYS DBA_REPGROUP_PRIVILEGES SYS DBA_REPKEY_COLUMNS SYS DBA_REPOBJECT SYS DBA_REPPARAMETER_COLUMN SYS DBA_REPPRIORITY SYS DBA_REPPRIORITY_GROUP SYS DBA_REPPROP SYS DBA_REPRESOLUTION SYS DBA_REPRESOLUTION_METHOD SYS DBA_REPRESOLUTION_STATISTICS SYS DBA_REPRESOL_STATS_CONTROL SYS DBA_REPSCHEMA SYS DBA_REPSITES SYS DBA_RGROUP SYS DBA_ROLES SYS DBA_ROLE_PRIVS SYS DBA_ROLLBACK_SEGS SYS DBA_RSRC_CONSUMER_GROUPS SYS DBA_RSRC_CONSUMER_GROUP_PRIVS SYS DBA_RSRC_MANAGER_SYSTEM_PRIVS SYS DBA_RSRC_PLANS SYS DBA_RSRC_PLAN_DIRECTIVES SYS DBA_RULESETS SYS DBA_SEGMENTS SYS DBA_SEQUENCES SYS DBA_SNAPSHOTS SYS DBA_SNAPSHOT_LOGS SYS DBA_SNAPSHOT_LOG_FILTER_COLS SYS DBA_SNAPSHOT_REFRESH_TIMES SYS DBA_SOURCE SYS DBA_STMT_AUDIT_OPTS SYS DBA_SUBPART_COL_STATISTICS SYS DBA_SUBPART_HISTOGRAMS SYS DBA_SUBPART_KEY_COLUMNS SYS DBA_SUMMARIES SYS DBA_SUMMARY_AGGREGATES SYS DBA_SUMMARY_DETAIL_TABLES SYS DBA_SUMMARY_JOINS SYS DBA_SUMMARY_KEYS SYS DBA_SYNONYMS SYS DBA_SYS_PRIVS SYS DBA_TABLES SYS DBA_TABLESPACES SYS DBA_TAB_COLUMNS SYS DBA_TAB_COL_STATISTICS SYS DBA_TAB_COMMENTS SYS DBA_TAB_HISTOGRAMS SYS DBA_TAB_MODIFICATIONS SYS DBA_TAB_PARTITIONS SYS DBA_TAB_PRIVS SYS DBA_TAB_SUBPARTITIONS SYS DBA_TEMP_FILES SYS DBA_TRIGGERS SYS DBA_TRIGGER_COLS SYS DBA_TS_QUOTAS SYS DBA_TYPES SYS DBA_TYPE_ATTRS SYS DBA_TYPE_METHODS SYS DBA_UNUSED_COL_TABS SYS DBA_UPDATABLE_COLUMNS SYS DBA_USERS SYS DBA_USTATS SYS DBA_VARRAYS SYS DBA_VIEWS SYS V_$: ---- VIEW_NAME OWNER ------------------------------ ------------------------------ V_$ACCESS SYS V_$ACTIVE_INSTANCES SYS V_$AQ SYS V_$AQ1 SYS V_$ARCHIVE SYS V_$ARCHIVED_LOG SYS V_$ARCHIVE_DEST SYS V_$ARCHIVE_PROCESSES SYS V_$BACKUP SYS V_$BACKUP_ASYNC_IO SYS V_$BACKUP_CORRUPTION SYS V_$BACKUP_DATAFILE SYS V_$BACKUP_DEVICE SYS V_$BACKUP_PIECE SYS V_$BACKUP_REDOLOG SYS V_$BACKUP_SET SYS V_$BACKUP_SYNC_IO SYS V_$BGPROCESS SYS V_$BH SYS V_$BSP SYS V_$BUFFER_POOL SYS V_$BUFFER_POOL_STATISTICS SYS V_$CIRCUIT SYS V_$CLASS_PING SYS V_$COMPATIBILITY SYS V_$COMPATSEG SYS V_$CONTEXT SYS V_$CONTROLFILE SYS V_$CONTROLFILE_RECORD_SECTION SYS V_$COPY_CORRUPTION SYS V_$DATABASE SYS V_$DATAFILE SYS V_$DATAFILE_COPY SYS V_$DATAFILE_HEADER SYS V_$DBFILE SYS V_$DBLINK SYS V_$DB_CACHE_ADVICE SYS V_$DB_OBJECT_CACHE SYS V_$DB_PIPES SYS V_$DELETED_OBJECT SYS V_$DISPATCHER SYS V_$DISPATCHER_RATE SYS V_$DLM_ALL_LOCKS SYS V_$DLM_CONVERT_LOCAL SYS V_$DLM_CONVERT_REMOTE SYS V_$DLM_LATCH SYS V_$DLM_LOCKS SYS V_$DLM_MISC SYS V_$DLM_RESS SYS V_$DLM_TRAFFIC_CONTROLLER SYS V_$ENABLEDPRIVS SYS V_$ENQUEUE_LOCK SYS V_$EVENT_NAME SYS V_$EXECUTION SYS V_$FAST_START_SERVERS SYS V_$FAST_START_TRANSACTIONS SYS V_$FILESTAT SYS V_$FILE_PING SYS V_$FIXED_TABLE SYS V_$FIXED_VIEW_DEFINITION SYS V_$GLOBAL_BLOCKED_LOCKS SYS V_$GLOBAL_TRANSACTION SYS V_$HS_AGENT SYS V_$HS_PARAMETER SYS V_$HS_SESSION SYS V_$INDEXED_FIXED_COLUMN SYS V_$INSTANCE SYS V_$INSTANCE_RECOVERY SYS V_$KCCDI SYS V_$KCCFE SYS V_$LATCH SYS V_$LATCHHOLDER SYS V_$LATCHNAME SYS V_$LATCH_CHILDREN SYS V_$LATCH_MISSES SYS V_$LATCH_PARENT SYS V_$LIBRARYCACHE SYS V_$LICENSE SYS V_$LOADCSTAT SYS V_$LOADISTAT SYS V_$LOADPSTAT SYS V_$LOADTSTAT SYS V_$LOCK SYS V_$LOCKED_OBJECT SYS V_$LOCKS_WITH_COLLISIONS SYS V_$LOCK_ACTIVITY SYS V_$LOCK_ELEMENT SYS V_$LOG SYS V_$LOGFILE SYS V_$LOGHIST SYS V_$LOGMNR_CONTENTS SYS V_$LOGMNR_DICTIONARY SYS V_$LOGMNR_LOGS SYS V_$LOGMNR_PARAMETERS SYS V_$LOG_HISTORY SYS V_$MAX_ACTIVE_SESS_TARGET_MTH SYS V_$MLS_PARAMETERS SYS V_$MTS SYS V_$MYSTAT SYS V_$NLS_PARAMETERS SYS V_$NLS_VALID_VALUES SYS V_$OBJECT_DEPENDENCY SYS V_$OBSOLETE_PARAMETER SYS V_$OFFLINE_RANGE SYS V_$OPEN_CURSOR SYS V_$OPTION SYS V_$PARALLEL_DEGREE_LIMIT_MTH SYS V_$PARAMETER SYS V_$PARAMETER2 SYS V_$PQ_SESSTAT SYS V_$PQ_SLAVE SYS V_$PQ_SYSSTAT SYS V_$PQ_TQSTAT SYS V_$PROCESS SYS V_$PROXY_ARCHIVEDLOG SYS V_$PROXY_DATAFILE SYS V_$PWFILE_USERS SYS V_$PX_PROCESS SYS V_$PX_PROCESS_SYSSTAT SYS V_$PX_SESSION SYS V_$PX_SESSTAT SYS V_$QUEUE SYS V_$RECOVERY_FILE_STATUS SYS V_$RECOVERY_LOG SYS V_$RECOVERY_PROGRESS SYS V_$RECOVERY_STATUS SYS V_$RECOVER_FILE SYS V_$REQDIST SYS V_$RESERVED_WORDS SYS V_$RESOURCE SYS V_$RESOURCE_LIMIT SYS V_$ROLLNAME SYS V_$ROLLSTAT SYS V_$ROWCACHE SYS V_$ROWCACHE_PARENT SYS V_$ROWCACHE_SUBORDINATE SYS V_$RSRC_CONSUMER_GROUP SYS V_$RSRC_CONSUMER_GROUP_CPU_MTH SYS V_$RSRC_PLAN SYS V_$RSRC_PLAN_CPU_MTH SYS V_$SESSION SYS V_$SESSION_CONNECT_INFO SYS V_$SESSION_CURSOR_CACHE SYS V_$SESSION_EVENT SYS V_$SESSION_LONGOPS SYS V_$SESSION_OBJECT_CACHE SYS V_$SESSION_WAIT SYS V_$SESSTAT SYS V_$SESS_IO SYS V_$SGA SYS V_$SGASTAT SYS V_$SHARED_POOL_RESERVED SYS V_$SHARED_SERVER SYS V_$SORT_SEGMENT SYS V_$SORT_USAGE SYS V_$SQL SYS V_$SQLAREA SYS V_$SQLTEXT SYS V_$SQLTEXT_WITH_NEWLINES SYS V_$SQL_BIND_DATA SYS V_$SQL_BIND_METADATA SYS V_$SQL_CURSOR SYS V_$SQL_SHARED_CURSOR SYS V_$SQL_SHARED_MEMORY SYS V_$STATNAME SYS V_$SUBCACHE SYS V_$SYSSTAT SYS V_$SYSTEM_CURSOR_CACHE SYS V_$SYSTEM_EVENT SYS V_$SYSTEM_PARAMETER SYS V_$SYSTEM_PARAMETER2 SYS V_$TABLESPACE SYS V_$TARGETRBA SYS V_$TEMPFILE SYS V_$TEMPORARY_LOBS SYS V_$TEMPSTAT SYS V_$TEMP_EXTENT_MAP SYS V_$TEMP_EXTENT_POOL SYS V_$TEMP_PING SYS V_$TEMP_SPACE_HEADER SYS V_$THREAD SYS V_$TIMER SYS V_$TRANSACTION SYS V_$TRANSACTION_ENQUEUE SYS V_$TYPE_SIZE SYS V_$VERSION SYS V_$WAITSTAT SYS V_$_LOCK SYS ========== 23 TUNING: ========== 1. init.ora settings -------------------- background_dump_dest = /var/opt/oracle/SALES/bdump control_files = ( /oradata/arc/control/ctrl1SALES.ctl , /oradata/temp/control/ctrl2SALES.ctl , /oradata/rbs/control/ctrl3SALES.ctl) db_block_size = 16384 db_name = SALES db_block_buffers = 17500 db_block_checkpoint_batch = 16 db_files = 255 db_file_multiblock_read_count = 10 license_max_users = 170 #core_dump_dest = /var/opt/oracle/SALES/cdump core_dump_dest = /oradata/rbs/cdump distributed_transactions = 40 dml_locks = 1000 job_queue_processes = 2 log_archive_buffers = 20 log_archive_buffer_size = 256 log_archive_dest = /oradata/arc log_archive_format = arcSALES_%s.arc log_archive_start = true log_buffer = 163840 log_checkpoint_interval = 1250 log_checkpoint_timeout = 1800 log_simultaneous_copies = 4 max_dump_file_size = 100240 max_enabled_roles = 50 oracle_trace_enable = true open_cursors = 2000 open_links = 20 processes = 200 remote_os_authent = true rollback_segments = (r1, r2, r3, rbig,rbig2) sequence_cache_entries = 30 sequence_cache_hash_buckets = 23 shared_pool_size = 750M sort_area_retained_size = 15728640 sort_area_size = 15728640 sql_trace = false timed_statistics = true resource_limit = true user_dump_dest = /var/opt/oracle/SALES/udump utl_file_dir = /var/opt/oracle/utl utl_file_dir = /var/opt/oracle/utl/frontend SORT_AREA_SIZE = 65536 (per PGA, max sort area) SORT_AREA_RETAINED_SIZE = 65536 (size after sort) PROCESSES = 100 (alle processes) DB_BLOCK_SIZE = 8192 DB_BLOCK_BUFFERS = 3400 (DB_CACHE_SIZE in Oracle 9i) SHARED_POOL_SIZE = 52428800 LOG_BUFFER = 26215400 4194304 8388608 LARGE_POOL_SIZE = DBWR_IO_SLAVES (DB_WRITER_PROCESSES) DB_WRITER_PROCESSES = 2 LGWR_IO_SLAVES= DB_FILE_MULTIBLOCK_READ_COUNT =16 (minimize io during table scans, it specifies max number of blocks in one io operation during sequential read) BUFFER_POOL_RECYCLE = BUFFER_POOL_KEEP = TIMED_STATISTICES =TRUE (statistics related to time are collected or not) OPTIMIZER_MODE =RULE, CHOOSE, FIRST_ROWS, ALL_ROWS PARALLEL_MIN_SERVERS = 2 (voor Parallel Query, en parallel recovery) PARALLEL_MAX_SERVERS = 4 RECOVERY_PARALLELISM = 2 (set parallel recovery op database niveau) 2. UTLBSTAT and UTLESTAT ------------------------ - if wanted change default tablespace of SYS to TOOLS - set timed_statistics=true - in $ORACLE_HOME/rdbms/admin you find utlbstat.sql and utlestat.sql to create perfoRMANce table and insert baseline: run utlbstat let the database run for some time to gather statistics, run utlestat which drop tables and generate report.txt 3. STATSPACK: ------------- Available as of 8.1.6 installation: - connect internal - @$ORACLE_HOME/rdbms/admin/statscre.sql It will create user PERFSTAT who ownes the new statistics tables You will be prompted for TEMP and DEFAULT tablespaces Gather statistices: - connect perfstat/perfstat - execute statspack.snap Or use DBMS_JOB to schedule the generation of snapshots Create report: - connect perfstat/perfstat - @ORACLE_HOME/rdbms/admin/statsrep.sql This will ask for beginning snapshot id and ending snapshot id. Then you can enter the filename for the report. 4. QUERIES: ----------- -- 4.1 HIT RATIO buffercache SELECT (1-(pr.value/(dbg.value+cg.value)))*100 FROM v$sysstat pr, v$sysstat dbg, v$sysstat cg WHERE pr.name = 'physical reads' AND dbg.name = 'db block gets' AND cg.name = 'consistent gets'; -- 4.2 redo noWait ratio SELECT (req.value*5000)/entries.value FROM v$sysstat req, v$sysstat entries WHERE req.name ='redo log space requests' AND entries.name='redo entries'; -- 4.3 Library cache and shared pool Overview memory: SELECT * FROM V$SGA; Free memory shared pool: SELECT * FROM v$sgastat WHERE name = 'free memory'; How often an object has to be reloaded into the cache once it has been loaded SELECT sum(pins) Executions, sum(reloads) Misses, sum(reloads)/sum(pins) Ratio FROM v$librarycache; SELECT gethits,gets,gethitratio FROM v$librarycache WHERE namespace = 'SQL AREA'; SELECT sum(sharable_mem) FROM v$db_object_cache; -- 4.4 TABLE OR INDEX REBUILD NECCESARY? SELECT substr(segment_name, 1, 30), segment_type, substr(owner, 1, 10), extents, initial_extent, next_extent, max_extents FROM dba_segments WHERE extents > max_extents - 100 AND owner not in ('SYS','SYSTEM'); SELECT index_name, blevel, decode(blevel,0,'OK BLEVEL',1,'OK BLEVEL', 2,'OK BLEVEL',3,'OK BLEVEL',4,'OK BLEVEL','BLEVEL HIGH') OK FROM dba_indexes WHERE owner='SALES'; EXAMPLE OF A SCRIPT THAT YOU MIGHT SCHEDULE ONCE A DAY: ------------------------------------------------------- -- report 1. set linesize 500 set pagesize 500 set serveroutput on set trimspool on spool d:\logs\ exec dbms_output.put_line('DAILY REPORT SALES DATABASE ON SERVER SUPER'); exec dbms_output.put_line('RUNTIME: '||to_char(SYSDATE, 'DD-MM-YYYY;HH24:MI')); exec dbms_output.put_line('Please read all sections carefully, takes only 1 minute.'); exec dbms_output.put_line(' '); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line('SECTION 1: OBJECTS AND USERS'); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('1.1 INVALID OBJECTS AS FOUND RIGHT NOW:'); exec dbms_output.put_line(' '); SELECT substr(object_name, 1. 30), substr(object_type, 1, 20), owner, status FROM dba_objects WHERE status='INVALID'; exec dbms_output.put_line(' '); exec dbms_output.put_line('Remark: If invalid objects are found intervention is required.'); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('1.2 TABLE/INDEX REACHING MAX NO OF EXTENTS:'); exec dbms_output.put_line(' '); SELECT substr(segment_name, 1, 30), segment_type, substr(owner, 1, 10), extents, initial_extent, next_extent, max_extents FROM dba_segments WHERE extents > max_extents - 50 AND owner not in ('SYS','SYSTEM'); exec dbms_output.put_line(' '); exec dbms_output.put_line('Remark: If objects are found intervention is required.'); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('1.3 SKEWED or BAD INDEXES with blevel > 3:'); exec dbms_output.put_line(' '); SELECT index_name, owner, blevel, decode(blevel,0,'OK BLEVEL',1,'OK BLEVEL', 2,'OK BLEVEL',3,'OK BLEVEL',4,'OK BLEVEL','BLEVEL HIGH') OK FROM dba_indexes WHERE owner in ('SALES','FRONTEND') and blevel > 3; exec dbms_output.put_line(' '); exec dbms_output.put_line('Remark: If indexes are found rebuild is required.'); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('1.4. NEW OBJECTS CREATED SINCE YESTERDAY:'); exec dbms_output.put_line(' '); SELECT owner, substr(object_name, 1, 30), object_type, created, last_ddl_time, status FROM dba_objects WHERE created > SYSDATE-5; exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('1.5. NEW ORACLE USERS CREATED SINCE YESTERDAY:'); exec dbms_output.put_line(' '); SELECT substr(username, 1, 20), account_status, default_tablespace, temporary_tablespace, created FROM dba_users WHERE created > SYSDATE -10; exec dbms_output.put_line(' '); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line('SECTION 2: TABLESPACES, DATAFILES, ROLLBACK SEGS'); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('2.1 FREE/USED SPACE OF TABLESPACES RIGHT NOW:'); exec dbms_output.put_line(' '); SELECT Total.name "Tablespace Name", Free_space, (total_space-Free_space) Used_space, total_space FROM (SELECT tablespace_name, sum(bytes/1024/1024) Free_Space FROM sys.dba_free_space GROUP BY tablespace_name ) Free, (SELECT b.name, sum(bytes/1024/1024) TOTAL_SPACE FROM sys.v_$datafile a, sys.v_$tablespace B WHERE a.ts# = b.ts# GROUP BY b.name ) Total WHERE Free.Tablespace_name = Total.name; exec dbms_output.put_line(' '); exec dbms_output.put_line('REMARK: FOR MONTHLY INTERNET BILLING AT LEAST 50MB SPACE MUST'); exec dbms_output.put_line('BE AVAILABLE IN EACH OF THE MANIIN% TABLESPACES. '); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('2.2 STATUS DATABASE FILES RIGHT NOW:'); exec dbms_output.put_line(' '); SELECT substr(file_name, 1, 50), tablespace_name, status FROM dba_data_files; exec dbms_output.put_line(' '); exec dbms_output.put_line('Remark: status of all files should be available '); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('2.3 STATUS ROLLBACK SEGMENTS RIGHT NOW:'); exec dbms_output.put_line(' '); SELECT substr(segment_name, 1, 20), substr(tablespace_name, 1, 20), status, INITIAL_EXTENT, NEXT_EXTENT, MIN_EXTENTS, MAX_EXTENTS, PCT_INCREASE FROM DBA_ROLLBACK_SEGS; exec dbms_output.put_line(' '); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line('SECTION 3: PERFORMANCE STATS SINCE DATABASE STARTUP'); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('3.1 ORACLE MEMORY (SGA LAYOUT):'); exec dbms_output.put_line(' '); SELECT * FROM V$SGA; exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('3.2 FREE MEMORY SHARED POOL:'); exec dbms_output.put_line(' '); SELECT * FROM v$sgastat WHERE name = 'free memory'; exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('3.3 LIBRARY (pl/sql) HIT RATIO:'); exec dbms_output.put_line(' '); SELECT sum(pins) Executions, sum(reloads) Misses, sum(reloads)/sum(pins) Ratio FROM v$librarycache; exec dbms_output.put_line(' '); exec dbms_output.put_line('Remark: above Ratio should be low '); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('3.4 DATABASE BUFFERS HIT RATIO:'); exec dbms_output.put_line(' '); SELECT (1-(pr.value/(dbg.value+cg.value)))*100 FROM v$sysstat pr, v$sysstat dbg, v$sysstat cg WHERE pr.name = 'physical reads' AND dbg.name = 'db block gets' AND cg.name = 'consistent gets'; exec dbms_output.put_line(' '); exec dbms_output.put_line('Remark: above Ratio should be high '); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('3.5 REDO BUFFERS WAITS:'); exec dbms_output.put_line(' '); SELECT (req.value*5000)/entries.value FROM v$sysstat req, v$sysstat entries WHERE req.name ='redo log space requests' AND entries.name='redo entries'; exec dbms_output.put_line(' '); exec dbms_output.put_line('Remark: above Ratio should be very low '); exec dbms_output.put_line(' '); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line('SECTION 4: LOCKS'); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('4.1 OBJECT LOCKS RIGHT NOW:'); exec dbms_output.put_line(' '); SELECT l.object_id object_id, l.session_id session_id, substr(l.oracle_username, 1, 10) username, substr(l.os_user_name, 1, 30) osuser, l.process process, l.locked_mode lockmode, substr(o.object_name, 1, 20) objectname FROM v$locked_object l, dba_objects o WHERE l.object_id=o.object_id; exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('4.2 PERSISTENT LOCKS SINCE YESTERDAY:'); exec dbms_output.put_line(' '); SELECT OBJECT_ID,SESSION_ID,USERNAME,OSUSER,PROCESS,LOCKMODE, OBJECT_NAME, to_char(DATUM, 'DD-MM-YYYY;HH24:MI') FROM PROJECTS.LOCKLIST WHERE DATUM > SYSDATE-2 ORDER BY DATUM; exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('4.3 BLOCKED SESSIONS RIGHT NOW:'); exec dbms_output.put_line(' '); SELECT s.sid sid, substr(s.username, 1, 10) username, substr(s.schemaname, 1, 10) schemaname, substr(s.osuser, 1, 10) osuser, substr(s.program, 1, 30) program, s.command command, l.lmode lockmode, l.block blocked FROM v$session s, v$lock l WHERE s.sid=l.sid and schemaname not in ('SYS','SYSTEM'); exec dbms_output.put_line(' '); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line('SECTION 5: ONLY NEEDED FOR oracle-dba '); exec dbms_output.put_line(' INFO NEEDED FOR RECOVERY '); exec dbms_output.put_line('==================================================='); exec dbms_output.put_line(' '); exec dbms_output.put_line('scn datafiles: '); exec dbms_output.put_line('scn controlfiles: '); exec dbms_output.put_line('latest 20 archived redo: '); exec dbms_output.put_line(' '); exec dbms_output.put_line(' '); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('---------------------------------------------------'); exec dbms_output.put_line('END REPORT 1'); exec dbms_output.put_line('Thanks a lot for reading this report !!!'); exit / ======== 24 RMAN: ======== $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ =============== 24.1: RMAN 10g: =============== 24.1.1 Create the catalog and register target database: ------------------------------------------------------- 10g example: ------------ Suppose an Oracle 10.2 target database is called "TEST10g". Suppose an Oracle 10.2 rman database is called "RMAN". Let's further suppose we have just created the RMAN database. Now we create the RMAN user (schema which will hold, or own, the metadata) create user rman identified by rman default tablespace rman temporary tablespace temp; - give the right permissions: grant connect, resource, recovery_catalog_owner to rman; Set up the catalog and register the target: RMAN> create catalog tablespace "RMAN" recovery catalog created RMAN> exit Recovery Manager complete. C:\oracle>rman catalog=rman/rman@rman target=system/vga88nt@test10g Recovery Manager: Release 10.2.0.1.0 - Production on Wed Feb 27 21:31:02 2008 Copyright (c) 1982, 2005, Oracle. All rights reserved. connected to target database: TEST10G (DBID=899275577) connected to recovery catalog database RMAN> register database; database registered in recovery catalog starting full resync of recovery catalog full resync complete 24.1.2 Backup and recovery examples 10g RMAN: --------------------------------------------- Good Examples using RMAN on 10g: -------------------------------- >>>> Full Backup First we configure several persistant parameters for this instance: RMAN> configure retention policy to recovery window of 5 days; RMAN> configure default device type to disk; RMAN> configure controlfile autobackup on; RMAN> configure channel device type disk format 'C:\Oracle\Admin\W2K2\Backup%d_DB_%u_%s_%p'; Next we perform a complete database backup using a single command: RMAN> run { 2> backup database plus archivelog; 3> delete noprompt obsolete; 4> } The recovery catalog should be resyncronized on a regular basis so that changes to the database structure and presence of new archive logs is recorded. Some commands perform partial and full resyncs implicitly, but if you are in doubt you can perform a full resync using the follwoing command: RMAN> resync catalog; >>>> Restore & Recover The Whole Database If the controlfiles and online redo logs are still present a whole database recovery can be achieved by running the following script: run { shutdown immediate; # use abort if this fails startup mount; restore database; recover database; alter database open; } This will result in all datafiles being restored then recovered. RMAN will apply archive logs as necessary until the recovery is complete. At that point the database is opened. If the tempfiles are still present you can issue a command like like the following for each of them: sql "ALTER TABLESPACE temp ADD TEMPFILE ''C:\Oracle\oradata\W2K2\temp01.dbf'' REUSE"; If the tempfiles are missing they must be recreated as follows: sql "ALTER TABLESPACE temp ADD TEMPFILE ''C:\Oracle\oradata\W2K2\temp01.dbf'' SIZE 100M AUTOEXTEND ON NEXT 64K"; >>>> Restore & Recover A Subset Of The Database A subset of the database can be restored in a similar fashion: run { sql 'ALTER TABLESPACE users OFFLINE IMMEDIATE'; restore tablespace users; recover tablespace users; sql 'ALTER TABLESPACE users ONLINE'; } Recovering a Tablespace in an Open Database The following example takes tablespace TBS_1 offline, restores and recovers it, then brings it back online: run { allocate channel dev1 type 'sbt_tape'; sql "ALTER TABLESPACE tbs_1 OFFLINE IMMEDIATE"; restore tablespace tbs_1; recover tablespace tbs_1; sql "ALTER TABLESPACE tbs_1 ONLINE"; } Recovering Datafiles Restored to New Locations The following example allocates one disk channel and one media management channel to use datafile copies on disk and backups on tape, and restores one of the datafiles in tablespace TBS_1 to a different location: run { allocate channel dev1 type disk; allocate channel dev2 type 'sbt_tape'; sql "ALTER TABLESPACE tbs_1 OFFLINE IMMEDIATE"; set newname for datafile 'disk7/oracle/tbs11.f' to 'disk9/oracle/tbs11.f'; restore tablespace tbs_1; switch datafile all; recover tablespace tbs_1; sql "ALTER TABLESPACE tbs_1 ONLINE"; } >>>> Example backup to sbt: echo " run { allocate channel t1 type 'sbt_tape' parms 'ENV=(tdpo_optfile=/usr/tivoli/tsm/client/oracle/bin64/tdpo.opt)'; allocate channel t2 type 'sbt_tape' parms 'ENV=(tdpo_optfile=/usr/tivoli/tsm/client/oracle/bin64/tdpo.opt)'; backup full database ; backup (spfile) (current controlfile) ; sql 'alter system archive log current'; backup archivelog all delete input ; release channel t1; release channel t2; } >>>> Incomplete Recovery As you would expect, RMAN allows incomplete recovery to a specified time, SCN or sequence number: run { shutdown immediate; startup mount; set until time 'Nov 15 2000 09:00:00'; # set until scn 1000; # alternatively, you can specify SCN # set until sequence 9923; # alternatively, you can specify log sequence number restore database; recover database; alter database open resetlogs; } The incomplete recovery requires the database to be opened using the RESETLOGS option. >>>> Disaster Recovery In a disaster situation where all files are lost you can only recover to the last SCN in the archived redo logs. Beyond this point the recovery would have to make reference to the online redo logs which are not present. Disaster recovery is therefore a type of incomplete recovery. To perform disaster recovery connect to RMAN: C:>rman catalog=rman/rman@w2k1 target=sys/password@w2k2 Once in RMAN do the following: startup nomount; restore controlfile; alter database mount; From SQL*Plus as SYS get the last archived SCN using: SQL> SELECT archivelog_change#-1 FROM v$database; ARCHIVELOG_CHANGE#-1 -------------------- 1048438 1 row selected. SQL>Back in RMAN do the following: run { set until scn 1048438; restore database; recover database; alter database open resetlogs; } If the "until scn" were not set the following type of error would be produced once a redo log was referenced: RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of recover command at 03/18/2003 09:33:19 RMAN-06045: media recovery requesting unknown log: thread 1 scn 1048439 With the database open all missing tempfiles must be replaced: sql "ALTER TABLESPACE temp ADD TEMPFILE ''C:\Oracle\oradata\W2K2\temp01.dbf'' SIZE 100M AUTOEXTEND ON NEXT 64K"; Once the database is fully recovered a new backup should be perfomed. The recovered database will be registered in the catalog as a new incarnation. The current incarnation can be listed and altered using the following commands: list incarnation; reset database to incarnation x;Lists And Reports RMAN has extensive listing and reporting functionality allowing you to monitor you backups and maintain the recovery catalog. Here are a few useful commands: >>>> Restoring a datafile to another location: For example, if you restore datafile ?/oradata/trgt/tools01.dbf to its default location, then RMAN restores the file ?/oradata/trgt/tools01.dbf and overwrites any file that it finds with the same filename. If you run a SET NEWNAME command before you restore a file, then RMAN creates a datafile copy with the name that you specify. For example, assume that you run the following commands: SET NEWNAME FOR DATAFILE '?/oradata/trgt/tools01.dbf' TO '/tmp/tools01.dbf'; RESTORE DATAFILE '?/oradata/trgt/tools01.dbf'; In this case, RMAN creates a datafile copy of ?/oradata/trgt/tools01.dbf named /tmp/tools01.dbf and records it in the repository. To change the name for datafile ?/oradata/trgt/tools01.dbf to /tmp/tools01.dbf in the control file, run a SWITCH command so that RMAN considers the restored file as the current database file. For example: SWITCH DATAFILE '/tmp/tools01.dbf' TO DATAFILECOPY '?/oradata/trgt/tools01.dbf'; The SWITCH command is the RMAN equivalent of the SQL statement ALTER DATABASE RENAME FILE. >>>> Archive logs What is the purpose and are the differences of “ALTER SYSTEM ARCHIVE LOG CURRENT” and “ALTER SYSTEM ARCHIVE LOG ALL” # When the database is open, run the following SQL statement to force Oracle to switch out of the current log and archive it as well as all other unarchived logs: ALTER SYSTEM ARCHIVE LOG CURRENT; # When the database is mounted, open, or closed, you can run the following SQL statement to force Oracle to archive all noncurrent redo logs: ALTER SYSTEM ARCHIVE LOG ALL; A log switch does not mean that the redo is archived. When you execute "'alter system archive log current" you force that the current log to be archived, so it is safe: you are sure to have all the needed archived logs. alter system archive log all: This command will archive all filled redo logs but will not complete current log because it will not be full. >>>> LIST AND REPORT COMMANDS: ============= LIST COMMAND: ============= List commands query the catalog or control file, to determine which backups or copies are available. List commands provide for basic information. Report commands can provide for much more detail. About RMAN Reports Generated by the LIST Command You can control how the output is displayed by using the BY BACKUP and BY FILE options of the LIST command and choosing between the SUMMARY and VERBOSE options. -- Example 1: Query on the incarnations of the target database RMAN> list incarnation of database; RMAN-03022: compiling command: list List of Database Incarnations DB Key Inc Key DB Name DB ID CUR Reset SCN Reset Time ------- ------- -------- ---------------- --- ---------- ---------- 1 2 AIRM 2092303715 YES 1 24-DEC-02 -- Example 2: Query on tablespace backups You can ask for lists of tablespace backups, as shown in the following example: RMAN> list backup of tablespace users; -- Example 3: Query on database backups RMAN> list backup of database; -- Example 4: Query on backup of archivelogs: RMAN> list backup of archivelog all; The primary purpose of the LIST command is to determine which backups are available. For example, you can list: . Backups and proxy copies of a database, tablespace, datafile, archived redo log, or control file . Backups that have expired . Backups restricted by time, path name, device type, tag, or recoverability . Incarnations of a database By default, RMAN lists backups by backup, which means that it serially lists each backup or proxy copy and then identifies the files included in the backup. You can also list backups by file. By default, RMAN lists in verbose mode. You can also list backups in a summary mode if the verbose mode generates too much output. Listing Backups by Backup To list backups by backup, connect to the target database and recovery catalog (if you use one), and then execute the LIST BACKUP command. Specify the desired objects with the listObjList clause. For example, you can enter: LIST BACKUP; # lists backup sets, image copies, and proxy copies LIST BACKUPSET; # lists only backup sets and proxy copies LIST COPY; # lists only disk copies Example: RMAN> LIST BACKUP OF DATABASE; By default the LIST output is detailed, but you can also specify that RMAN display the output in summarized form. Specify the desired objects with the listObjectList or recordSpec clause. If you do not specify an object, then LIST BACKUP displays all backups. After connecting to the target database and recovery catalog (if you use one), execute LIST BACKUP, specifying the desired objects and options. For example: LIST BACKUP SUMMARY; # lists backup sets, proxy copies, and disk copies You can also specify the EXPIRED keyword to identify those backups that were not found during a crosscheck: LIST EXPIRED BACKUP SUMMARY; # Show all backup details list backup; ================ Report commands: ================ RMAN>report schema; Shows the physical structure of the target database. RMAN> report obsolete; RMAN-03022: compiling command: report RMAN-06147: no obsolete backups found -- REPORT COMMAND: -- --------------- About Reports of RMAN Backups Reports enable you to confirm that your backup and recovery strategy is in fact meeting your requirements for database recoverability. The two major forms of REPORT used to determine whether your database is recoverable are: RMAN> REPORT NEED BACKUP; Reports which database files need to be backed up to meet a configured or specified retention policy Use the REPORT NEED BACKUP command to determine which database files need backup under a specific retention policy. With no arguments, REPORT NEED BACKUP reports which objects need backup under the currently configured retention policy. The output for a configured retention policy of REDUNDANCY 1 is similar to this example: REPORT NEED BACKUP; RMAN retention policy will be applied to the command RMAN retention policy is set to redundancy 1 Report of files with less than 1 redundant backups File #bkps Name ---- ----- ----------------------------------------------------- 2 0 /oracle/oradata/trgt/undotbs01.dbf RMAN> REPORT UNRECOVERABLE; Reports which database files require backup because they have been affected by some NOLOGGING operation such as a direct-path insert You can report backup sets, backup pieces and datafile copies that are obsolete, that is, not needed to meet a specified retention policy, by specifying the OBSOLETE keyword. If you do not specify any other options, then REPORT OBSOLETE displays the backups that are obsolete according to the current retention policy, as shown in the following example: RMAN> REPORT OBSOLETE; In the simplest case, you could crosscheck all backups on disk, tape or both, using any one of the following commands: RMAN> CROSSCHECK BACKUP DEVICE TYPE DISK; RMAN> CROSSCHECK BACKUP DEVICE TYPE SBT; RMAN> CROSSCHECK BACKUP; # crosshecks all backups on all devices The REPORT SCHEMA command lists and displays information about the database files. After connecting RMAN to the target database and recovery catalog (if you use one), issue REPORT SCHEMA as shown in this example: RMAN> REPORT SCHEMA; # Show items that beed 7 days worth of # archivelogs to recover completely report need backup days = 7 database; report need backup; # Show/Delete items not needed for recovery report obsolete; delete obsolete; # Show/Delete items not needed for point-in-time # recovery within the last week report obsolete recovery window of 7 days; delete obsolete recovery window of 7 days; RMAN> REPORT OBSOLETE REDUNDANCY 2; RMAN> REPORT OBSOLETE RECOVERY WINDOW OF 5 DAYS; RMAN displays backups that are obsolete according to those retention policies, regardless of the actual configured retention policy. # Show/Delete items with more than 2 newer copies available report obsolete redundancy = 2 device type disk; delete obsolete redundancy = 2 device type disk; # Show datafiles that connot currently be recovered report unrecoverable database; report unrecoverable tablespace 'USERS'; 24.1.3 More on Backup and recovery 10g RMAN: -------------------------------------------- 24.1.3.1 About RMAN Backups: ---------------------------- When you execute the BACKUP command in RMAN, you create one or more backup sets or image copies. By default, RMAN creates backup sets regardless of whether the destination is disk or a media manager. >>>About Image Copies An image copy is an exact copy of a single datafile, archived redo log file, or control file. Image copies are not stored in an RMAN-specific format. They are identical to the results of copying a file with operating system commands. RMAN can use image copies during RMAN restore and recover operations, and you can also use image copies with non-RMAN restore and recovery techniques. To create image copies and have them recorded in the RMAN repository, run the RMAN BACKUP AS COPY command (or, alternatively, configure the default backup type for disk as image copies using CONFIGURE DEVICE TYPE DISK BACKUP TYPE TO COPY before performing a backup). A database server session is used to create the copy, and the server session also performs actions such as validating the blocks in the file and recording the image copy in the RMAN repository. You can also use an operating system command such as the UNIX dd command to create image copies, though these will not be validated, nor are they recorded in the RMAN repository. You can use the CATALOG command to add image copies created with native operating system tools in the RMAN repository. >>>Using RMAN-Created Image Copies If you run a RESTORE command, then by default RMAN restores a datafile or control file to its original location by copying an image copy backup to that location. Image copies are chosen over backup sets because of the extra overhead of reading through an entire backup set in search of files to be restored. However, if you need to restore and recover a current datafile, and if you have an image copy of the datafile available on disk, then you do not actually need to have RMAN copy the image copy back to its old location. You can instead have the database use the image copy in place, as a replacement for the datafile to be restored. The SWITCH command updates the RMAN repository indicate that the image copy should now be treated as the current datafile. Issuing the SWITCH command in this case is equivalent to issuing the SQL statement ALTER DATABASE RENAME FILE. You can then perform recovery on the copy. >>>User-Managed Image Copies RMAN can use image copies created by mechanisms outside of RMAN, such as native operating system file copy commands or third-party utilities that leave image copies of files on disk. These copies are known as user-managed copies or operating system copies. The RMAN CATALOG command causes RMAN to inspect an existing image copy and enter its metadata into the RMAN repository. Once cataloged, these files can be used like any other backup with the RESTORE or SWITCH commands. Some sites store their datafiles on mirrored disk volumes, which permit the creation of image copies by breaking a mirror. After you have broken the mirror, you can notify RMAN of the existence of a new user-managed copy, thus making it a candidate for a backup operation. You must notify RMAN when the copy is no longer available, by using the CHANGE ... UNCATALOG command. In this example, before resilvering the mirror (not including other copies of the broken mirror), you must use a CHANGE ... UNCATALOG command to update the recovery catalog and indicate that this copy is no longer available. >>>Storage of Backups on Disk and Tape RMAN can create backups on disk or a third-party media device such as a tape drive. If you specify DEVICE TYPE DISK, then your backups are created on disk, in the file name space of the target instance that is creating the backup. You can make a backup on any device that can store a datafile. To create backups on non-disk media, such as tape, you must use third-party media management software, and allocate channels with device types, such as SBT, that are supported by that software. >>>Backups of Archived Logs There are several features of RMAN backups specific to backups of archived redo logs. Deletion of Archived Logs After Backups RMAN can delete one or all copies of archived logs from disk after backing them up to backup sets. If you specify the DELETE INPUT option, then RMAN backs up exactly one copy of each specified log sequence number and thread from an archive destination to tape, and then deletes the specific file it backed up while leaving the other copies on disk. If you specify the DELETE ALL INPUT option, then RMAN backs up exactly one copy of each specified log sequence number and thread, and then deletes that log from all archive destinations. Note that there are special considerations related to deletion of archived redo logs in standby database configurations. See Oracle Data Guard Concepts and Administration for details. >>>Backups of Backup Sets The RMAN BACKUP BACKUPSET command backs up previously created backup sets. Only backup sets that were created on device type DISK can be backed up, and they can be backed up to any available device type. Note: RMAN issues an error if you attempt to run BACKUP AS COPY BACKUPSET. The BACKUP BACKUPSET command uses the default disk channel to copy backup sets from disk to disk. To back up from disk to tape, you must either configure or manually allocate a non-disk channel. Uses for Backups of Backup Sets The BACKUP BACKUPSET command is a useful way to spread backups among multiple media. For example, you can execute the following BACKUP command weekly as part of the production backup schedule: # makes backup sets on disk BACKUP DEVICE TYPE DISK AS BACKUPSET DATABASE PLUS ARCHIVELOG; BACKUP DEVICE TYPE sbt BACKUPSET ALL; # copies backup sets on disk to tape In this way, you ensure that all your backups exist on both disk and tape. You can also duplex backups of backup sets, as in this example: BACKUP COPIES 2 DEVICE TYPE sbt BACKUPSET ALL; (Again, control file autobackups are never duplexed.) You can also use BACKUP BACKUPSET to manage backup space allocation. For example, to keep more recent backups on disk and older backups only on tape, you can regularly run the following command: BACKUP DEVICE TYPE sbt BACKUPSET COMPLETED BEFORE 'SYSDATE-7' DELETE INPUT; This command backs up backup sets that were created more than a week ago from disk to tape, and then deletes them from disk. Note that DELETE INPUT here is equivalent to DELETE ALL INPUT; RMAN deletes all existing copies of the backup set. If you duplexed a backup to four locations, then RMAN deletes all four copies of the pieces in the backup set. >>> Restoring Files with RMAN Use the RMAN RESTORE command to restore the following types of files from disk or other media: - Database (all datafiles) - Tablespaces - Control files - Archived redo logs - Server parameter files Because a backup set is in a proprietary format, you cannot simply copy it as you would a backup database file created with an operating system utility; you must use the RMAN RESTORE command to extract its contents. In contrast, the database can use image copies created by the RMAN BACKUP AS COPY command without additional processing. RMAN automates the procedure for restoring files. You do not need to go into the operating system, locate the backup that you want to use, and manually copy files into the appropriate directories. When you issue a RESTORE command, RMAN directs a server session to restore the correct backups to either: - The default location, overwriting the files with the same name currently there - A new location, which you can specify with the SET NEWNAME command To restore a datafile, either mount the database or keep it open and take the datafile to be restored offline. When RMAN performs a restore, it creates the restored files as datafile image copies and records them in the repository. The following table describes the behavior of the RESTORE, SET NEWNAME, and SWITCH commands. >>>Datafile Media Recovery with RMAN The concept of datafile media recovery is the application of online or archived redo logs or incremental backups to a restored datafile in order to update it to the current time or some other specified time. Use the RMAN RECOVER command to perform media recovery and apply logs or incremental backups automatically. RMAN Media Recovery: Basic Steps If possible, make the recovery catalog available to perform the media recovery. If it is not available, or if you do not maintain a recovery catalog, then RMAN uses metadata from the target database control file. If both the control file and recovery catalog are lost, then you can still recover the database --assuming that you have backups of the datafiles and at least one autobackup of the control file. The generic steps for media recovery using RMAN are as follows: -Place the database in the appropriate state: mounted or open. For example, mount the database when performing whole database recovery, or open the database when performing online tablespace recovery. -To perform incomplete recovery, use the SET UNTIL command to specify the time, SCN, or log sequence number at which recovery terminates. Alternatively, specify the UNTIL clause on the RESTORE and RECOVER commands. -Restore the necessary files with the RESTORE command. -Recover the datafiles with the RECOVER command. -Place the database in its normal state. For example, open it or bring recovered tablespaces online. RESTORE DATABASE; RECOVER DATABASE; >>> Corrupt Block recovery Although datafile media recovery is the principal form of recovery, you can also use the RMAN BLOCKRECOVER command to perform block media recovery. Block media recovery recovers an individual corrupt datablock or set of datablocks within a datafile. In cases when a small number of blocks require media recovery, you can selectively restore and recover damaged blocks rather than whole datafiles. For example, you may discover the following messages in a user trace file: ORA-01578: ORACLE data block corrupted (file # 7, block # 3) ORA-01110: data file 7: '/oracle/oradata/trgt/tools01.dbf' ORA-01578: ORACLE data block corrupted (file # 2, block # 235) ORA-01110: data file 2: '/oracle/oradata/trgt/undotbs01.dbf' You can then specify the corrupt blocks in the BLOCKRECOVER command as follows: BLOCKRECOVER DATAFILE 7 BLOCK 3 DATAFILE 2 BLOCK 235; >>> After a Database Restore and Recover, RMAN gives the error: RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of backup command at 03/03/2008 11:13:06 RMAN-06059: expected archived log not found, lost of archived log compromises recoverability ORA-19625: error identifying file /dbms/tdbaeduc/educroca/recovery/archive/arch_1_870_617116679.arch ORA-27037: unable to obtain file status IBM AIX RISC System/6000 Error: 2: No such file or directory Note 1: If you no longer have a particular archivelog file you can let RMAN catalog know this by issuing the following command at the rman prompt after connecting to the rman catalog and the target database - change archivelog all crosscheck ; This will check the archivelog folder and then make the catalog agree with what is actually available. rman> DELETE EXPIRED ARCHVIELOG ; Oracle Error :: RMAN-20011 target database incarnation is not current in recovery catalog Cause the database incarnation that matches the resetlogs change# and time of the mounted target database control file is not the current incarnation of the database Action If "reset database to incarnation " was used to make an old incarnation current then restore the target database from a backup that matches the incarnation and mount it. You will need to do "startup nomount" before you can restore the control file using RMAN. Otherwise use "reset database to incarnation " make the intended incarnation current in the recovery catalog. >>> Note about rman and tape sbt and recovery window: Suppose you have a retention period defined in rman, like for example CONFIGURE RETENTION POLICY TO REDUNDANCY 3 This means that 3 backups needs to be maintained by rman, and other backups are considered "obsolete". But those other backups beyond retention, are not expired or otherwise not usable. If they are still present, you can use them in a recovery. Besides this, it cannot be known beforehand how the tape subsystem will deal with rman commands like "delete obsolete". The tape subsystem has probably its own retention period, and you need much more details about all systems involved, before you know whats going on. ============================================= 24.1.3.2 ABOUT RMAN ERRORS / troubleshooting: ============================================= Err 1: Missing archived redolog: ================================ Problem: If an archived redo is missing, you might get a message similar like this: RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of backup command at 03/05/2008 07:44:35 RMAN-06059: expected archived log not found, lost of archived log compromises recoverability ORA-19625: error identifying file /dbms/tdbaeduc/educroca/recovery/archive/arch_1_817_617116679.arch ORA-27037: unable to obtain file status IBM AIX RISC System/6000 Error: 2: No such file or directory Solution: If archived redo logs are (wrongly) deleted/moved/compressed from disk without being backed up, the rman catalog will not know this has happened, and will keep attempting to backup the missing archived redo logs. That will cause rman archived redo log backups to fail altogether with an error like: RMAN-06059: expected archived log not found, lost of archived log compromises recoverability If you can, you should bring back the missing archved redo logs to their original location and name, and let rman back them up. But if that is impossible, the workaround is to “crosscheck archivelog all”, like: rman < crosscheck archivelog all; You’ll get output like this: validation succeeded for archived log archive log filename=D:REDOARCHARCH_1038.DBF recid=1017 stamp=611103638 for every archived log as they are all checked on disk. That should be the catalog fixed, run an archivelog backup to make sure. Err 2: online redo logs listed as archives: =========================================== Testcase: a 10g 10.2.0.3 shows after recovery with resetlogs the following in v$archived_log. It looks as if it will stay there forever: SEQ# FIRST NEXT NAME DIFF STATUS 814 17311773 17311785 12 D 815 17311785 17354662 42877 D 816 17354662 17354674 12 D 817 17354674 17402531 47857 D 2 17415287 2.81E+14 redo01.log 2.8147E+14 A 0 0 0 redo02.log 0 A 0 0 0 redo03.log 0 A 0 0 0 redo04.log 0 A 1 -->17402532 17415287 redo05.log 12755 A 1 17402532 17404154 1622 D 2 17404154 17404165 11 D FIRST_CHANGE# NEXT_CHANGE# SEQUENCE# RESETLOGS_CHANGE# ------------- ------------ ---------- ----------------- 17311785 17354662 815 1 17354662 17354674 816 1 17354674 17402531 817 1 -->17402532 17404154 1 -->17402532 17404154 17404165 2 17402532 17404165 17415733 3 17402532 We dont know what is going on here. Err 3: Highlevel overview RMAN Error Codes ========================================== RMAN error codes are summarized in the table below. 0550-0999 Command-line interpreter 1000-1999 Keyword analyzer 2000-2999 Syntax analyzer 3000-3999 Main layer 4000-4999 Services layer 5000-5499 Compilation of RESTORE or RECOVER command 5500-5999 Compilation of DUPLICATE command 6000-6999 General compilation 7000-7999 General execution 8000-8999 PL/SQL programs 9000-9999 Low-level keyword analyzer 10000-10999 Server-side execution 11000-11999 Interphase errors between PL/SQL and RMAN 12000-12999 Recovery catalog packages 20000-20999 Miscellaneous RMAN error messages Err 4: RMAN-03009 accompinied with ORA- error: ============================================== Q: Here is my problem; When trying to delete obsolete RMAN backupsets, I get an error: RMAN> change backupset 698, 702, 704, 708 delete; List of Backup Pieces BP Key BS Key Pc# Cp# Status Device Type Piece Name ------- ------- --- --- ----------- ----------- ---------- 698 698 1 1 AVAILABLE SBT_TAPE df_546210555_706_1 702 702 1 1 AVAILABLE SBT_TAPE df_546296605_709_1 704 704 1 1 AVAILABLE SBT_TAPE df_546383776_712_1 708 708 1 1 AVAILABLE SBT_TAPE df_546469964_715_1 Do you really want to delete the above objects (enter YES or NO)? YES RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03009: failure of delete command on ORA_MAINT_SBT_TAPE_1 channel at 03/02/2005 16:27:06 ORA-27191: sbtinfo2 returned error Additional information: 2 What in the world does "Additional information: 2" mean? I can't find any more useful detail than this. A: Oracle Error :: ORA-27191 sbtinfo2 returned error Cause sbtinfo2 returned an error. This happens while retrieving backup file information from the media manager"s catalog. Action This error is returned from the media management software which is linked with Oracle. There should be additional messages which explain the cause of the error. This error usually requires contacting the media management vendor. A: ---> ORA-27191 John Clarke: My guess is that "2" is an O/S return code, and in /usr/sys/include/errno.h, you'll see that error# 2 is "no such file or directory. Accompanied with ORA-27191, I'd guess that your problem is that your tape library doesn't currently have the tape(s) loaded and/or can't find them. Mladen Gogala: Additional information 2 means that OS returned status 2. That is a "file not found" error. In plain Spanglish, you cannot delete files from tape, only from the disk drives. Niall Litchfield: The source error is the ora-27191 error (http://download-west.oracle.com/docs/cd/B14117_01/server.101/b10744/e24280.htm#ORA-27191) which suggests a tape library issue to me. You can search for RMAN errors using the error search page as well http://otn.oracle.com/pls/db10g/db10g.error_search?search=rman-03009, for example A: ---> RMAN-03009 RMAN-03009: failure of delete command on ORA_MAINT_SBT_TAPE_1 channel at date/time RMAN-03009: failure of allocate command on t1 channel at date/time RMAN-03009: failure of backup command on t1 channel at date/time etc.. -> Means most of the time that you have Media Management Library problems -> Can also mean that there is a problem with backup destination (disk not found, no space, tape not loaded etc..) ERR 5: Test your Media Management API: ====================================== Testing the Media Management API On specified platforms, Oracle provides a diagnostic tool called "sbttest". This utility performs a simple test of the tape library by acting as the Oracle database server and attempting to communicate with the media manager. Obtaining the Utility On UNIX, the sbttest utility is located in $ORACLE_HOME/bin. Obtaining Online Documentation For online documentation of sbttest, issue the following on the command line: % sbttest The program displays the list of possible arguments for the program: Error: backup file name must be specified Usage: sbttest backup_file_name # this is the only required parameter <-dbname database_name> <-trace trace_file_name> <-remove_before> <-no_remove_after> <-read_only> <-no_regular_backup_restore> <-no_proxy_backup> <-no_proxy_restore> <-file_type n> <-copy_number n> <-media_pool n> <-os_res_size n> <-pl_res_size n> <-block_size block_size> <-block_count block_count> <-proxy_file os_file_name bk_file_name [os_res_size pl_res_size block_size block_count]> The display also indicates the meaning of each argument. For example, following is the description for two optional parameters: Optional parameters: -dbname specifies the database name which will be used by SBT to identify the backup file. The default is "sbtdb" -trace specifies the name of a file where the Media Management software will write diagnostic messages. Using the Utility Use sbttest to perform a quick test of the media manager. The following table explains how to interpret the output: If sbttest returns... Then... 0 The program ran without error. In other words, the media manager is installed and can accept a data stream and return the same data when requested. non-0 The program encountered an error. Either the media manager is not installed or it is not configured correctly. To use sbttest: Make sure the program is installed, included in your system path, and linked with Oracle by typing sbttest at the command line: % sbttest If the program is operational, you should see a display of the online documentation. Execute the program, specifying any of the arguments described in the online documentation. For example, enter the following to create test file some_file.f and write the output to sbtio.log: % sbttest some_file.f -trace sbtio.log You can also test a backup of an existing datafile. For example, this command tests datafile tbs_33.f of database PROD: % sbttest tbs_33.f -dbname prod Examine the output. If the program encounters an error, it provides messages describing the failure. For example, if Oracle cannot find the library, you see: libobk.so could not be loaded. Check that it is installed properly, and that LD_ LIBRARY_PATH environment variable (or its equivalent on your platform) includes the directory where this file can be found. Here is some additional information on the cause of this error: ld.so.1: sbttest: fatal: libobk.so: open failed: No such file or directory ERR 6: RMAN-12004 ================= Hi, I m facing this problem any pointers will of great help.... 1. RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-00579: the following error occurred at 12/16/2003 02:46:31 RMAN-10035: exception raised in RPC: RMAN-10031: ORA-19624 occurred during call to DBMS_BACKUP_RESTORE.BACKUPPIECECREATE RMAN-03015: error occurred in stored script backup_db_full RMAN-03015: error occurred in stored script backup_del_all_al RMAN-03007: retryable error occurred during execution of command: backup RMAN-12004: unhandled exception during command execution on channel t1 RMAN-10035: exception raised in RPC: ORA-19506: failed to create sequential file, name="l0f93ro5_1_1", parms="" ORA-27028: skgfqcre: sbtbackup returned error ORA-19511: Error received from media manager layer, error text: sbtbackup: Failed to process backup file. RMAN-10031: ORA-19624 occurred during call to DBMS_BACKUP_RESTORE.BACKUPPIECECREATE OR another errorstack RMAN-12004: unhandled exception during command execution on channel disk13 RMAN-10035: exception raised in RPC: ORA-19502: write error on file "/db200_backup/archive_log03/EDPP_ARCH0_21329_1_492222998", blockno 612353 (blocksize=1024) ORA-27072: skgfdisp: I/O error HP-UX Error: 2: No such file or directory Additional information: 612353 RMAN-10031: ORA-19624 occurred during call to DBMS_BACKUP_RESTORE.BACKUPPIECECREATE OR another errorstack RMAN-12004: unhandled exception during command execution on channel ch00 RMAN-10035: exception raised in RPC: ORA-19599: block number 691 is corrupt in controlfile C:\ORACLE\ORA90\DATABASE\SNCFSUMMITDB.ORA RMAN-10031: ORA-19583 occurred during call to DBMS_BACKUP_RESTORE.BACKUPPIECECREATE OR another errorstack Have managed to create a job to backup my db, but I can't restore. I get the following: RMAN-03002: failure during compilation of command RMAN-03013: command type: restore RMAN-03006: non-retryable error occurred during execution of command: IRESTORE RMAN-07004: unhandled exception during command execution on channel BackupTest RMAN-10035: exception raised in RPC: ORA-19573: cannot obtain exclusive enqueue for datafile 1 RMAN-10031: ORA-19583 occurred during call to DBMS_BACKUP_RESTORE.RESTOREBACKUPPIECE Seems to relate to corrupt or missing Oracle files. $$$$ ERR 7: ORA-27211 ================ Q: Continue to get ORA-27211 Failed to load media management library A: had a remarkably similar experience a few months ago with Legato NetWorker and performed all of the steps you listed with the same results. The problem turned out to be very simple. The SA installed the 64-bit version of the Legato Networker client because it is a 64-bit server. However, we were running a 32-bit version of Oracle on it. Installing the 32-bit client solved the problem. A: Cause: User-supplied SBT_LIBRARY or libobk.so could not be loaded. Call to dlopen for media library returned error. See Additional information for error code. Action: Retry the command with proper media library. Or re-install Media management module for Oracle. A: Exact Error Message ORA-27211: Failed to load Media Management Library on HP-UX system Details: Overview: The Oracle return code ORA-27211 implies a failure to load a shared object library into process space. Oracle Recovery Manager (RMAN) backups will fail with a message "ORA-27211: Failed to load Media Management Library" if the SBT_LIBRARY keyword is defined and points to an incorrect library name. The SBT_LIBRARY keyword must be set in the PARMS clause of the ALLOCATE CHANNEL statement in the RMAN script. This keyword is not valid with the SEND command and is new to Oracle 9i. If this value is set, it overrides the default search path for the libobk library. By default, SBT_LIBRARY is not set. Troubleshooting: If an ORA-27211 error is seen for an Oracle RMAN backup, it is necessary to review the Oracle RMAN script and verify if SBT_LIBRARY is either not set or is set correctly. If set, the filename should be libobk.sl for HP-UX 10, 11.00 and 11.11, but libobk.so for HP-UX 11.23 (ia64) clients. Example of an invalid entry for HP-UX 11.23 (ia64) clients: PARMS='SBT_LIBRARY=/usr/openv/netbackup/bin/libobk.sl' Example of a correct entry for HP-UX 11.23 (ia64) clients: PARMS='SBT_LIBRARY=/usr/openv/netbackup/bin/libobk.so' Master Server Log Files: n/a Media Server Log Files: n/a Client Log Files: The RMAN log file on the client will show the following error message: RMAN-00571: =========================================== RMAN-00569: ======= ERROR MESSAGE STACK FOLLOWS ======= RMAN-00571: =========================================== RMAN-03009: failure of allocate command on ch00 channel at 05/21/2005 16:39:17 ORA-19554: error allocating device, device type: SBT_TAPE, device name: ORA-27211: Failed to load Media Management Library Additional information: 25 Resolution: The Oracle return code ORA-27211 implies a failure to load a shared object library into process space. Oracle RMAN backups will fail with a message "ORA-27211: Failed to load Media Management Library" if the SBT_LIBRARY keyword is defined and points to an incorrect library name. To manually set the SBT_LIBRARY path, follow the steps described below: 1. Modify the RMAN ALLOCATE CHANNEL statement in the backup script to reference the HP-UX 11.23 library file directly: PARMS='SBT_LIBRARY=/usr/openv/netbackup/bin/libobk.so' Note: This setting would be added to each ALLOCATE CHANNEL statement. A restart of the Oracle instance is not needed for this change to take affect. 2. Run a test backup or wait for the next scheduled backup of the Oracle database ERR8: More on DBMS_BACKUP_RESTORE: ================================== Note 1: The dbms_backup_restore package is used as a PL/SQL command-line interface for replacing native RMAN commands, and it has very little documentation. The Oracle docs note how to install and configure the dbms_backup_restore package: “The DBMS_BACKUP_RESTORE package is an internal package created by the dbmsbkrs.sql and prvtbkrs.plb scripts. This package, along with the target database version of DBMS_RCVMAN, is automatically installed in every Oracle database when the catproc.sql script is run. This package interfaces with the Oracle database server and the operating system to provide the I/O services for backup and restore operations as directed by RMAN.” The docs also note that “The DBMS_BACKUP_RESTORE package has a PL/SQL procedure to normalize filenames on Windows NT platforms.” Oracle DBA John Parker gives this example of dbms_backup_restore to recover a controlfile: declare devtype varchar2(256); done boolean; begin devtype:=dbms_backup_restore.deviceallocate( type=>'sbt_tape', params=>'ENV=(OB2BARTYPE=Oracle8,OB2APPNAME=rdcs,OB2BARLIST=ORA_RDCS_WEEKLY)', ident=>'t1'); dbms_backup_restore.restoresetdatafile; dbms_backup_restore.restorecontrolfileto('D:\oracle\ora81\dbs\CTL1rdcs.ORA'); dbms_backup_restore.restorebackuppiece( 'ORA_RDCS_WEEKLY.dbf', DONE=>done ); dbms_backup_restore.restoresetdatafile; dbms_backup_restore.restorecontrolfileto('D:\DBS\RDCS\CTL2RDCS.ORA'); dbms_backup_restore.restorebackuppiece( 'ORA_RDCS_WEEKLY.dbf', DONE=>done ); dbms_backup_restore.devicedeallocate('t1'); end; Here are some other examples of using dbms_backup_restore: DECLARE devtype varchar2(256); done boolean; BEGIN devtype := dbms_backup_restore.DeviceAllocate (type => '',ident => 'FUN'); dbms_backup_restore.RestoreSetDatafile; dbms_backup_restore.RestoreDatafileTo(dfnumber => 1,toname => 'D:\ORACLE_BASE\datafiles\SYSTEM01.DBF'); dbms_backup_restore.RestoreDatafileTo(dfnumber => 2,toname => 'D:\ORACLE_BASE\datafiles\UNDOTBS.DBF'); --dbms_backup_restore.RestoreDatafileTo(dfnumber => 3,toname => 'D:\ORACLE_BASE\datafiles\MYSPACE.DBF'); dbms_backup_restore.RestoreBackupPiece(done => done,handle => 'D:\ORACLE_BASE\RMAN_BACKUP\MYDB_DF_BCK05H2LLQP_1_1', params => null); dbms_backup_restore.DeviceDeallocate; END; / --restore archived redolog DECLARE devtype varchar2(256); done boolean; BEGIN devtype := dbms_backup_restore.DeviceAllocate (type => '',ident => 'FUN'); dbms_backup_restore.RestoreSetArchivedLog(destination=>'D:\ORACLE_BASE\achive\'); dbms_backup_restore.RestoreArchivedLog(thread=>1,sequence=>1); dbms_backup_restore.RestoreArchivedLog(thread=>1,sequence=>2); dbms_backup_restore.RestoreArchivedLog(thread=>1,sequence=>3); dbms_backup_restore.RestoreBackupPiece(done => done,handle => 'D:\ORACLE_BASE\RMAN_BACKUP\MYDB_LOG_BCK0DH1JGND_1_1', params => null); dbms_backup_restore.DeviceDeallocate; END; / Note 2: ------- --restore controlfile DECLARE devtype varchar2(256); done boolean; BEGIN devtype := dbms_backup_restore.DeviceAllocate(type => '',ident => 'FUN'); dbms_backup_restore.RestoresetdataFile; dbms_backup_restore.RestoreControlFileto('D:\ORACLE_BASE\controlfiles\CONTROL01.CTL'); dbms_backup_restore.RestoreBackupPiece('D:\ORACLE_BASE\Rman_Backup\MYDB_DF_BCK0BH1JBVA_1_1',done => done); dbms_backup_restore.RestoresetdataFile; dbms_backup_restore.RestoreControlFileto('D:\ORACLE_BASE\controlfiles\CONTROL02.CTL'); dbms_backup_restore.RestoreBackupPiece('D:\ORACLE_BASE\Rman_Backup\MYDB_DF_BCK0BH1JBVA_1_1',done => done); dbms_backup_restore.RestoresetdataFile; dbms_backup_restore.RestoreControlFileto('D:\ORACLE_BASE\controlfiles\CONTROL03.CTL'); dbms_backup_restore.RestoreBackupPiece('D:\ORACLE_BASE\Rman_Backup\MYDB_DF_BCK0BH1JBVA_1_1',done => done); dbms_backup_restore.DeviceDeallocate; END; / --restore datafile DECLARE devtype varchar2(256); done boolean; BEGIN devtype := dbms_backup_restore.DeviceAllocate (type => '',ident => 'FUN'); dbms_backup_restore.RestoreSetDatafile; dbms_backup_restore.RestoreDatafileTo(dfnumber => 1,toname => 'D:\ORACLE_BASE\datafiles\SYSTEM01.DBF'); dbms_backup_restore.RestoreDatafileTo(dfnumber => 2,toname => 'D:\ORACLE_BASE\datafiles\UNDOTBS.DBF'); --dbms_backup_restore.RestoreDatafileTo(dfnumber => 3,toname => 'D:\ORACLE_BASE\datafiles\MYSPACE.DBF'); dbms_backup_restore.RestoreBackupPiece(done => done,handle => 'D:\ORACLE_BASE\RMAN_BACKUP\MYDB_DF_BCK05H2LLQP_1_1', params => null); dbms_backup_restore.DeviceDeallocate; END; / --restore archived redolog DECLARE devtype varchar2(256); done boolean; BEGIN devtype := dbms_backup_restore.DeviceAllocate (type => '',ident => 'FUN'); dbms_backup_restore.RestoreSetArchivedLog(destination=>'D:\ORACLE_BASE\achive\'); dbms_backup_restore.RestoreArchivedLog(thread=>1,sequence=>1); dbms_backup_restore.RestoreArchivedLog(thread=>1,sequence=>2); dbms_backup_restore.RestoreArchivedLog(thread=>1,sequence=>3); dbms_backup_restore.RestoreBackupPiece(done => done,handle => 'D:\ORACLE_BASE\RMAN_BACKUP\MYDB_LOG_BCK0DH1JGND_1_1', params => null); dbms_backup_restore.DeviceDeallocate; END; / ERR 9: RMAN-00554 initialization of internal recovery manager package failed: ============================================================================= connected to target database: PLAYROCA (DBID=575215626) RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-00554: initialization of internal recovery manager package failed RMAN-04004: error from recovery catalog database: ORA-03135: connection lost contact keys: RMAN-00554 RMAN-04004 ORA-03135 ORA-3136 >>>> In alertlog of the rman, we can find: WARNING: inbound connection timed out (ORA-3136) Thu Mar 13 23:09:54 2008 >>>> In Net logs sqlnet.log we can find: Warning: Errors detected in file /dbms/tdbaplay/ora10g/home/network/log/sqlnet.log > *********************************************************************** > Fatal NI connect error 12170. > > VERSION INFORMATION: > TNS for IBM/AIX RISC System/6000: Version 10.2.0.3.0 - Production > TCP/IP NT Protocol Adapter for IBM/AIX RISC System/6000: Version 10.2.0.3.0 - Production > Oracle Bequeath NT Protocol Adapter for IBM/AIX RISC System/6000: Version 10.2.0.3.0 - Production > Time: 18-MAR-2008 23:01:43 > Tracing not turned on. > Tns error struct: > ns main err code: 12535 > TNS-12535: TNS:operation timed out > ns secondary err code: 12606 > nt main err code: 0 > nt secondary err code: 0 > nt OS err code: 0 > Client address: (ADDRESS=(PROTOCOL=tcp)(HOST=57.232.4.123)(PORT=35844)) Note 1: ------- RMAN-00554: initialization of internal recovery manager package failed Is a general error code. You must turn your attention the the codes underneath this one. For example: RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-00554: initialization of internal recovery manager package failed RMAN-06003: ORACLE error from target database: ORA-00210: cannot open the specified control file ORA-00202: control file: '/devel/dev02/dev10g/standbyctl.ctl' RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-00554: initialization of internal recovery manager package failed RMAN-04005: error from target database: ORA-01017: invalid username/password; Note 2: ------- RMAN-04004: error from recovery catalog database: ORA-03135: connection lost contact ERR 10: RMAN-00554 initialization of internal recovery manager package failed: ============================================================================== Starting backup at 17-MAY-08 released channel: t1 released channel: t2 RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of backup command at 05/17/2008 23:30:13 RMAN-06004: ORACLE error from recovery catalog database: RMAN-20242: specification does not match any archive log in the recovery catalog Note 1: ------- Oracle Error :: RMAN-20242 specification does not match any archivelog in the recovery catalog Cause No archive logs in the specified archive log range could be found. Action Check the archive log specifier. Note 2: ------- Some of the common RMAN errors are: RMAN-20242: Specification does not match any archivelog in the recovery catalog. Add to RMAN script: sql 'alter system archive log current'; Note 3: ------- Q: RMAN-20242: specification does not match any archive log in the recovery ca Posted: Feb 12, 2008 7:52 AM Reply A couple of archive log files were deleted from the OS. They still show up in the list of archive logs in Enterprise Manager. I want to fix this because now whenever I try to run a crosscheck command, I get the message: RMAN-20242: specification does not match any archive log in the recovery catalog I also tried to uncatalog those files, but got the same message. Any suggestions on what to do? Thanks! A: hi, from rman run the command list expired archivelog; if ther archives are in this list they will show, then i think you should do a crosscheck archivelog all; then you should be able to delete them. regards Note 4: ------- The RMAN error number would be helpful, but this is a common problem - RMAN-20242 - and is addressed in detail in MetaLink notes. Either the name specification (the one you entered) is wrong, or you could be using mismatched versions between RMAN and the database (don't know since you didn't provide any version details). Note 5: ------- Q: Hi there! We are having problems with an Oracle backup. The compiling of the backup command fails with the error message: RMAN-20242: specification does not match any archivelog in the recovery catalog But RMAN is only supposed to backup any archived logs that are there and then insert them in the catalog... Did anybody experience anything similar? This is 8.1.7 on HP-UX with Legato Networker Thanks, A: If i ask rman to backup archivelogs that are more than 2days old and there are none, thats not an error. That is when i see it the most, now most companies will force a log switch after a set amount of time during the day so in DR, you dont lose days worth of redo that might still be hanging in a redo log if it gets lost. $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ Now we will do some test of RMAN on a testsystem with Oracle 10g R2 Test Case 1: ============ 10g Database test10g: TEST10G: startup mount pfile=c:\oracle\admin\test10g\pfile\init.ora alter database archivelog; archive log start; alter database force logging; alter database add supplemental log data; alter database open; Files and tablespaces: >>> User albert creates table TEST CREATE TABLE test ( id number, name varchar2(10)); insert into test values (1,'test1'); commit; >>> make full RMAN backup BACKUP 1: >>> Some time later, albert inserts second record insert into test values (2,'test2'); commit; >>> make full RMAN backup BACKUP 2: >>> Now investigate some SCN's: SQL> select CHECKPOINT_CHANGE#,CONTROLFILE_SEQUENCE#,ARCHIVE_CHANGE#,CURRENT_SCN from v$database; CHECKPOINT_CHANGE# CONTROLFILE_SEQUENCE# ARCHIVE_CHANGE# CURRENT_SCN ------------------ --------------------- --------------- ----------- 888889 1745 889087 889154 SQL> select CHECKPOINT_CHANGE#,CONTROLFILE_SEQUENCE#,ARCHIVE_CHANGE#,CURRENT_SCN,archivelog_change# from v$database; CHECKPOINT_CHANGE# CONTROLFILE_SEQUENCE# ARCHIVE_CHANGE# CURRENT_SCN ARCHIVELOG_CHANGE# ------------------ --------------------- --------------- ----------- ------------------ 889090 1748 889087 890538 889090 SQL> SELECT DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() from dual; DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() ----------------------------------------- 890599 SQL> select CHECKPOINT_CHANGE#,CONTROLFILE_SEQUENCE#,ARCHIVE_CHANGE#,CURRENT_SCN,archivelog_change# from v$database; CHECKPOINT_CHANGE# CONTROLFILE_SEQUENCE# ARCHIVE_CHANGE# CURRENT_SCN ARCHIVELOG_CHANGE# ------------------ --------------------- --------------- ----------- ------------------ 889090 1748 889087 890678 889090 SQL> select file#,CHECKPOINT_CHANGE#,LAST_CHANGE#,OFFLINE_CHANGE#,ONLINE_CHANGE#,NAME from v$datafile; FILE# CHECKPOINT_CHANGE# LAST_CHANGE# OFFLINE_CHANGE# ONLINE_CHANGE# NAME ---------- ------------------ ------------ --------------- -------------- -------------------------- 1 888936 534906 534907 C:\ORACLE\ORADATA\TEST10G\SYSTEM01.DBF 2 888936 534906 534907 C:\ORACLE\ORADATA\TEST10G\UNDOTBS01.DBF 3 888936 534906 534907 C:\ORACLE\ORADATA\TEST10G\SYSAUX01.DBF 4 888936 534906 534907 C:\ORACLE\ORADATA\TEST10G\USERS01.DBF 5 888936 0 0 C:\ORACLE\ORADATA\TEST10G\EXAMPLE01.DBF 6 888936 0 0 C:\ORACLE\ORADATA\TEST10G\TS_CDC.DBF 6 rows selected. SQL> select RL_SEQUENCE#,RL_FIRST_CHANGE#,RL_NEXT_CHANGE# from V$BACKUP_FILES ...... 151 888889 889090 151 888889 889090 SQL> select SEQUENCE#,FIRST_CHANGE#, STATUS from v$log; SEQUENCE# FIRST_CHANGE# STATUS ---------- ------------- ---------------- 152 889090 CURRENT 150 887780 INACTIVE 151 888889 INACTIVE SQL> select SEQUENCE#,FIRST_CHANGE#,NEXT_CHANGE# from v$log_history ......... 147 880266 882166 148 882166 882431 149 882431 887780 150 887780 888889 151 888889 889090 >>> Some time later, albert inserts third record insert into test values (3,'test3'); commit; >>> shutdown database >>> delete a datafile data file 6: 'C:\ORACLE\ORADATA\TEST10G\TS_CDC.DBF' >>> startup database SQL> alter database open; alter database open * ERROR at line 1: ORA-01157: cannot identify/lock data file 6 - see DBWR trace file ORA-01110: data file 6: 'C:\ORACLE\ORADATA\TEST10G\TS_CDC.DBF' >>> RECOVER WITH RMAN RMAN> RESTORE DATABASE; RMAN> RECOVER DATABASE; >>> logon as albert SQL> select * from test; ID NAME ---------- ---------- 3 test3 1 test1 2 test2 >>> logon as system SQL> select CHECKPOINT_CHANGE#,CONTROLFILE_SEQUENCE#,ARCHIVE_CHANGE#,CURRENT_SCN from v$database; CHECKPOINT_CHANGE# CONTROLFILE_SEQUENCE# ARCHIVE_CHANGE# CURRENT_SCN ------------------ --------------------- --------------- ----------- 891236 1780 889087 891702 SQL> select file#,CHECKPOINT_CHANGE#,LAST_CHANGE#,OFFLINE_CHANGE#,ONLINE_CHANGE#,NAME from v$datafile; FILE# CHECKPOINT_CHANGE# LAST_CHANGE# OFFLINE_CHANGE# ONLINE_CHANGE# NAME ---------- ------------------ ------------ --------------- -------------- -------------------------- 1 891236 534906 534907 C:\ORACLE\ORADATA\TEST10G\SYSTEM01.DBF 2 891236 534906 534907 C:\ORACLE\ORADATA\TEST10G\UNDOTBS01.DBF 3 891236 534906 534907 C:\ORACLE\ORADATA\TEST10G\SYSAUX01.DBF 4 891236 534906 534907 C:\ORACLE\ORADATA\TEST10G\USERS01.DBF 5 891236 0 0 C:\ORACLE\ORADATA\TEST10G\EXAMPLE01.DBF 6 891236 0 0 C:\ORACLE\ORADATA\TEST10G\TS_CDC.DBF 6 rows selected. SQL> select CHECKPOINT_CHANGE#,CONTROLFILE_SEQUENCE#,ARCHIVE_CHANGE#,CURRENT_SCN,archivelog_change# from v$database; CHECKPOINT_CHANGE# CONTROLFILE_SEQUENCE# ARCHIVE_CHANGE# CURRENT_SCN ARCHIVELOG_CHANGE# ------------------ --------------------- --------------- ----------- ------------------ 893124 1785 889087 893131 889090 SQL> select file#,CHECKPOINT_CHANGE#,LAST_CHANGE#,OFFLINE_CHANGE#,ONLINE_CHANGE#,NAME from v$datafil e; FILE# CHECKPOINT_CHANGE# LAST_CHANGE# OFFLINE_CHANGE# ONLINE_CHANGE# NAME ---------- ------------------ ------------ --------------- -------------- -------------------------- 1 893124 534906 534907 C:\ORACLE\ORADATA\TEST10G\SYSTEM01.DBF 2 893124 534906 534907 C:\ORACLE\ORADATA\TEST10G\UNDOTBS01.DBF 3 893124 534906 534907 C:\ORACLE\ORADATA\TEST10G\SYSAUX01.DBF 4 893124 534906 534907 C:\ORACLE\ORADATA\TEST10G\USERS01.DBF 5 893124 0 0 C:\ORACLE\ORADATA\TEST10G\EXAMPLE01.DBF 6 893124 0 0 C:\ORACLE\ORADATA\TEST10G\TS_CDC.DBF 6 rows selected. select THREAD#,SEQUENCE#,FIRST_CHANGE#,NEXT_CHANGE# from v$log_history; .... 1 149 882431 887780 1 150 887780 888889 1 151 888889 889090 1 152 889090 893499 1 153 893499 895665 1 154 895665 896834 1 155 896834 898275 1 156 898275 899008 select THREAD#,SEQUENCE#,FIRST_CHANGE#,NEXT_CHANGE# from v$archived_log; 1 149 882431 887780 1 149 882431 887780 1 150 887780 888889 1 150 887780 888889 1 151 888889 889090 1 151 888889 889090 1 152 889090 893499 1 152 889090 893499 1 153 893499 895665 1 153 893499 895665 1 154 895665 896834 1 154 895665 896834 1 155 896834 898275 1 155 896834 898275 1 156 898275 899008 1 156 898275 899008 END TESTCASE 1: =============== $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ ---------------------------------- The V$RMAN_OUTPUT memory-only view shows the output of a currently executing RMAN job, whereas the V$RMAN_STATUS control file view indicates the status of both executing and completed RMAN jobs. The V$BACKUP_FILES provides access to the information used as the basis of the LIST BACKUP and REPORT OBSOLETE commands. Best views to obtain backup information are: V$RMAN_STATUS V$BACKUP_FILES v$archived_log v$log_history v$database; You can also list backups by querying V$BACKUP_FILES and the RC_BACKUP_FILES recovery catalog view. These views provide access to the same information as the LIST BACKUPSET command. ---------------------------------- Enhanced Reporting: RESTORE PREVIEW The PREVIEW option to the RESTORE command can now tell you which backups will be accessed during a RESTORE operation. ---------------------------------- >> To run RMAN commands interactively, start RMAN and then type commands into the command-line interface. For example, you can start RMAN from the UNIX command shell and then execute interactive commands as follows: % rman TARGET SYS/oracle@trgt CATALOG rman/cat@catdb % rman TARGET=SYS/oracle@trgt CATALOG=rman/cat@catdb ---------------------------------- >> Command files In this example, a sample RMAN script is placed into a command file called commandfile.rcv. You can run this file from the operating system command line and write the output into the log file outfile.txt as follows: % rman TARGET / CATALOG rman/cat@catdb CMDFILE commandfile.rcv LOG outfile.txt ---------------------------------- Run the CONFIGURE DEFAULT DEVICE TYPE command to specify a default device type for automatic channels. For example, you may make backups to tape most of the time and only occasionally make a backup to disk. In this case, configure channels for disk and tape devices, but make sbt the default device type: CONFIGURE DEVICE TYPE DISK PARALLELISM 1; # configure device disk CONFIGURE DEVICE TYPE sbt PARALLELISM 2; # configure device sbt CONFIGURE DEFAULT DEVICE TYPE TO sbt; Now, RMAN will, by default, use sbt channels for backups. For example, if you run the following command: BACKUP TABLESPACE users; RMAN only allocates channels of type sbt during the backup because sbt is the default device. $$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$$ 24.2 Older RMAN stuff: 8,8i,9i: =============================== 24.1 Introduction: ------------------ Recovery Manager (RMAN) is an Oracle tool that allows you to back up, copy, restore, and recover datafiles, control files, and archived redo logs. It is included with the Oracle server and does not require separate installation. You can invoke RMAN as a command line utility from the operating system (O/S) prompt or use the GUI-based Enterprise Manager Backup Manager. RMAN users "server sessions" to automate many of the backup and recovery tasks that were formerly performed manually. For example, instead of requiring you to locate appropriate backups for each datafile, copy them to the correct place using operating system commands, and choose which archived logs to apply, RMAN manages these tasks automatically. RMAN stores metadata about its backup and recovery operations in the recovery catalog, which is a centralized repository of information, or exclusively in the control file. Typically, the recovery catalog is stored in a separate database. If you do not use a recovery catalog, RMAN uses the control file as its repository of metadata. RMAN can be used on a database in archive mode or no archive mode. !!!! But, for open backups, the database MUST BE in ARCHIVE MODE. That's true for Oracle 8, 8i, 9i and 10g. RMAN doesn't do a "begin backup". It is not necessary when you use RMAN. RMAN does an intelligent copy of the database blocks (as opposed to a simple OS copy) and it ensures we do not copy a fractured block. The whole purpose of the begin backup (of the OS type of backup) is to record more info into the redo logs in the event an OS copy copies a "fractured block" - where the head and tail do not match (can happen since we are WRITING to the database at the same time the backup would be reading). When RMAN hits such a block -- it re-reads it to get a clean copy. How to start RMAN? - You can call from unix, or cmd prompt, the RMAN utility: $ rman RMAN> Once started you will see the RMAN> prompt. - Or you can give command line paramaters along with the rman call % rman target sys/sys_pwd@prod1 catalog rman/rman@rcat 24.2 Types of commands, and interactive mode or batch mode: ----------------------------------------------------------- RMAN uses two basic types of commands: stand-alone commands and job commands. - The job commands always appear within the brackets of a run command. - The stand-alone command can be issued right after the RMAN prompt. You can run RMAN in interactive mode or batch mode - examples of interactive mode: RMAN> run { 2> allocate channel d1 type disk; 3> backup database; 4> } RMAN> run { allocate channel c1 type disk; copy datafile 6 to 'F:\oracle\backups\oem01.cpy'; release channel c1; } RMAN> run { allocate channel c1 type disk; backup format 'F:\oracle\backups\oem01.rbu' ( datafile 6 ); release channel c1; } RMAN> run { allocate channel c1 type 'sbt_tape'; restore database; recover database; } Note about 'channel': You must allocate a 'channel" before you execute backup and recovery commands. Each allocated channel establishes a connection from RMAN to a target database by starting a server session on the instance. This server session performs the backup and recovery operations. Only one RMAN session communicates with the allocated server sessions. You can allocate multiple channels, thus allowing a single RMAN command to read or write multiple backups or image copies in parallel. Thus, the number of channels that you allocate affects the degree of parallelism within a command. When backing up to tape you should allocate one channel for each physical device, but when backing up to disk you can allocate as many channels as necessary for maximum throughput. The simplest way to determine whether RMAN encountered an error is to examine its return code. RMAN returns 0 to the operating system if no errors occurred, 1 otherwise. For example, if you are running UNIX and using the C shell, RMAN outputs the return code into a shell variable called $status. The second easiest way is to search the Recovery Manager output for the string RMAN-00569, which is the message number for the error stack banner. All RMAN errors are preceded by this error message. If you do not see an RMAN-00569 message in the output, then there are no errors. - example of batch mode: You can type RMAN commands into a file, and then run the command file by specifying its name on the command line. The contents of the command file should be identical to commands entered at the command line. Suppose the commandfile is called 'b_whole_l0.rcv', then the rman call could be as in the following example: $ rman target / catalog rman/rman@rcat @b_whole_l0.rcv log rman_log.f Another example: c:> rman target xxx/yyy@target rcvcat aaa/bbb@catalog cmdfile bkdb.scr msglog bkdb.log 24.3. Recovery Manager Repository or RMAN Catalog: -------------------------------------------------- Storage of the RMAN Repository in the Recovery Catalog, or exclusively in the target database controlfile: The RMAN repository is the collection of metadata about your target databases that RMAN uses to conduct its backup, recovery, and maintenance operations. You can either create a recovery catalog in which to store this information, or let RMAN store it exclusively in the target database control file. Although RMAN can conduct all major backup and recovery operations using just the control file, some RMAN commands function only when you use a recovery catalog. The recovery catalog is maintained solely by RMAN; the target database never accesses it directly. RMAN propagates information about the database structure, archived redo logs, backup sets, and datafile copies into the recovery catalog from the target database's control file. A single recovery catalog is able to store information for multiple target databases. What is in the recovery catalog? -------------------------------- -Datafile and archived redo log backup sets and backup pieces. -Datafile copies. -Archived redo logs and their copies. -Tablespaces and datafiles on the target database. -Stored scripts, which are named user-created sequences of RMAN and SQL commands. Resynchronization of the Recovery Catalog ----------------------------------------- The recovery catalog obtains crucial RMAN metadata from the target database control file. Resynchronization of the recovery catalog ensures that the metadata that RMAN obtains from the control file stays current. Resynchronizations can be full or partial. In a partial resynchronization, RMAN reads the current control file to update changed data, but does not resynchronize metadata about the database physical schema: datafiles, tablespaces, redo threads, rollback segments (only if the database is open), and online redo logs. In a full resynchronization, RMAN updates all changed records, including schema records. When you issue certain commands in RMAN, the program automatically detects when it needs to perform a full or partial resynchronization and executes the operation as needed. You can also force a full resynchronization by issuing a 'resync catalog' command. It is a good idea to run RMAN once a day or so and issue the resync catalog command to ensure that the catalog stays current. Because the control file employs a circular reuse system, backup and copy records eventually get overwritten. A single recovery catalog is able to store information for multiple target databases. 24.4 Media Manager: ------------------- To utilize tape storage for your database backups, RMAN requires a media manager. A media manager is a utility that loads, labels, and unloads sequential media such as tape drives for the purpose of backing up and recovering data. Note that Oracle does not need to connect to the media management library (MML) software when it backs up to disk. Software that is compliant with the MML interface enables an Oracle server session to issue commands to the media manager to back up or restore a file. The media manager responds to the command by loading, labeling, or unloading the requested tape. 24.5 Backups: ------------- When you execute the backup command, you create one or more backup sets. A backup set, which is a logical construction, contains one or more physical backup pieces. Backup pieces are operating system files that contain the backed up datafiles, control files, or archived redo logs. You cannot split a file across different backup sets or mix archived redo logs and datafiles into a single backup set. A backup set is a complete set of backup pieces that constitute a full or incremental backup of the objects specified in the backup command. Backup sets are in an RMAN-specific format; image copies, in contrast, are available for use without additional processing. So, for example: You can have a backupset 'backupset 1' containing just 1 datafile. You can have a backupset 'backupset 2' containing many datafiles, as blocks. You can have a backupset 'backupset 3' containing archived redologs You can either let RMAN determine a unique name for the backup piece or use the format parameter to specify a name. If you do not specify a filename, RMAN uses the %U substitution variable to guarantee a unique name. The backup command provides substitution variables that allow you to generate unique filenames. 24.6 Starting RMAN Sessions: ---------------------------- Example 1: connect to target database ------------------------------------- $ ORACLE_SID=brdb;export ORACLE_SID $rman RMAN>connect target sys/password RMAN .. connected Example 2: connect to catalog database -------------------------------------- $rman RMAN>connect catalog rman/rman RMAN .. connected Starting and stopping target database $ ORACLE_SID=brdb;export ORACLE_SID $rman RMAN>connect target sys/password RMAN .. connected RMAN>startup -- will start the target database RMAN>shutdown -- will stop the target database Example 3: starting RMAN with command parameters: ------------------------------------------------- $ ORACLE_SID=brdb;export ORACLE_SID $ rman target sys/password@prod1 catalog rman/rman@rcat rman target sys/cactus@playroca catalog rman/cactus@playrman 24.7 Creating the Recovery Catalog: ----------------------------------- - create a database for the Recovery Catalog, for example rcdb - create the user that will hold the catalog, rman with password rman create user rman identified by rman default tablespace rman temporary tablespace temp; - give the right permissions: grant connect, resource, recovery_catalog_owner to rman; - create the catalog in database rcdb In 8.0, to setup Recovery Catalog, you can run $ORACLE_HOME/rdbms/admin/catrman.sql while connected to RMAN database. In 8.1 and later, to setup the Recovery Catalog, use the create catalog command. $ rman RMAN>connect catalog rman/rman RMAN-06008 connected to recovery catalog database RMAN-06428 recovery catalog is not installed RMAN>create catalog tablespace rman; RMAN-06431 recovery catalog created You can expect something like the following to exist in the rcdb database: SQL> select table_name, tablespace_name, owner 2 from dba_tables where owner='RMAN'; TABLE_NAME TABLESPACE_NAME OWNER ------------------------------ ------------------------------ ------ AL DATA RMAN BCB DATA RMAN BCF DATA RMAN BDF DATA RMAN BP DATA RMAN BRL DATA RMAN BS DATA RMAN CCB DATA RMAN CCF DATA RMAN CDF DATA RMAN CKP DATA RMAN CONFIG DATA RMAN DB DATA RMAN DBINC DATA RMAN DF DATA RMAN DFATT DATA RMAN OFFR DATA RMAN ORL DATA RMAN RCVER DATA RMAN RLH DATA RMAN RR DATA RMAN RT DATA RMAN SCR DATA RMAN SCRL DATA RMAN TS DATA RMAN TSATT DATA RMAN XCF DATA RMAN XDF DATA RMAN 28 rows selected. SQL> select view_name, owner 2 from dba_views where owner='RMAN'; 8, 8i: ------ VIEW_NAME OWNER ------------------------------ ----- RC_ARCHIVED_LOG RMAN RC_BACKUP_CONTROLFILE RMAN RC_BACKUP_CORRUPTION RMAN RC_BACKUP_DATAFILE RMAN RC_BACKUP_PIECE RMAN RC_BACKUP_REDOLOG RMAN RC_BACKUP_SET RMAN RC_CHECKPOINT RMAN RC_CONTROLFILE_COPY RMAN RC_COPY_CORRUPTION RMAN RC_DATABASE RMAN RC_DATABASE_INCARNATION RMAN RC_DATAFILE RMAN RC_DATAFILE_COPY RMAN RC_LOG_HISTORY RMAN RC_OFFLINE_RANGE RMAN RC_PROXY_CONTROLFILE RMAN RC_PROXY_DATAFILE RMAN RC_REDO_LOG RMAN RC_REDO_THREAD RMAN RC_RESYNC RMAN RC_STORED_SCRIPT RMAN RC_STORED_SCRIPT_LINE RMAN RC_TABLESPACE RMAN 24 rows selected. The recovery catalog is now installed in the database rcdb. 10g: ---- SQL> select view_name from dba_views where view_name like '%RMAN%'; VIEW_NAME ------------------------------ V_$RMAN_CONFIGURATION GV_$RMAN_CONFIGURATION V_$RMAN_STATUS V_$RMAN_OUTPUT GV_$RMAN_OUTPUT V_$RMAN_BACKUP_SUBJOB_DETAILS V_$RMAN_BACKUP_JOB_DETAILS V_$RMAN_BACKUP_TYPE MGMT$HA_RMAN_CONFIG RC_RMAN_OUTPUT RC_RMAN_BACKUP_SUBJOB_DETAILS RC_RMAN_BACKUP_JOB_DETAILS RC_RMAN_BACKUP_TYPE RC_RMAN_CONFIGURATION RC_RMAN_STATUS 15 rows selected. SQL> select view_name, owner 2 from dba_views where owner='RMAN'; XXXX RC_RMAN_OUTPUT RC_BACKUP_FILES RC_RMAN_BACKUP_SUBJOB_DETAILS RC_RMAN_BACKUP_JOB_DETAILS RC_BACKUP_SET_DETAILS RC_BACKUP_PIECE_DETAILS RC_BACKUP_COPY_DETAILS RC_PROXY_COPY_DETAILS RC_PROXY_ARCHIVELOG_DETAILS RC_BACKUP_DATAFILE_DETAILS RC_BACKUP_CONTROLFILE_DETAILS RC_BACKUP_ARCHIVELOG_DETAILS RC_BACKUP_SPFILE_DETAILS RC_BACKUP_SET_SUMMARY RC_BACKUP_DATAFILE_SUMMARY RC_BACKUP_CONTROLFILE_SUMMARY RC_BACKUP_ARCHIVELOG_SUMMARY RC_BACKUP_SPFILE_SUMMARY RC_BACKUP_COPY_SUMMARY RC_PROXY_COPY_SUMMARY RC_PROXY_ARCHIVELOG_SUMMARY RC_UNUSABLE_BACKUPFILE_DETAILS RC_RMAN_BACKUP_TYPE RC_DATABASE RC_DATABASE_INCARNATION RC_RESYNC RC_CHECKPOINT RC_TABLESPACE RC_DATAFILE RC_TEMPFILE RC_REDO_THREAD RC_REDO_LOG RC_LOG_HISTORY RC_ARCHIVED_LOG RC_BACKUP_SET RC_BACKUP_PIECE RC_BACKUP_DATAFILE RC_BACKUP_CONTROLFILE RC_BACKUP_SPFILE RC_DATAFILE_COPY RC_CONTROLFILE_COPY RC_BACKUP_REDOLOG RC_BACKUP_CORRUPTION RC_COPY_CORRUPTION RC_OFFLINE_RANGE RC_STORED_SCRIPT RC_STORED_SCRIPT_LINE RC_PROXY_DATAFILE RC_PROXY_CONTROLFILE RC_RMAN_CONFIGURATION RC_DATABASE_BLOCK_CORRUPTION RC_PROXY_ARCHIVEDLOG RC_RMAN_STATUS 53 rows selected. Compatibility: --------------- If you use an 8.1.6 RMAN executable to execute the "create catalog" command, then the recovery catalog is created as a release 8.1.6 recovery catalog. Compatibility=8.1.6 You cannot use the 8.1.6 catalog with a pre-8.1.6 release of the RMAN executable. If you use an 8.1.6 RMAN executable to execute the "upgrade catalog" command, then the recovery catalog is upgraded from a pre-8.1.6 release to a release 8.1.6 catalog. Compatibility=8.0.4 The 8.1.6 catalog is backwards compatible with older releases of the RMAN executable. To view compatibility: SQL> SELECT value FROM config WHERE name='compatible'; Use an older RMAN to create the catalog. Use the newer RMAN to upgrade the catalog. You can allwys do: RMAN> configure compatible = 8.1.5; *** EXTRA: different RMAN CATALOGS in 1 DATABASE *** Different versions in one database: ----------------------------------- In general, the rules of RMAN compatibility are as follows: - The RMAN catalog schema version (tables/views) should be greater than or equal to the catalog database version. - The RMAN catalog is backwards compatible with target databases from earlier releases. - The versions of the RMAN executable and the target database should be the same - RMAN cannot create release 8.1 or later catalog schemas in 8.0 catalog databases. Suppose you have 8.0.5 and 9i target databases. - create one 9i database rcdb - create 2 tablespaces: RCAT80 and RCAT9I - create corresponding rman users Create the 8.0.5 catalog in the 9.2.0 catalog database. # sql syntax for creating logical catalog 8.0.5 structure. create tablespace RCAT80 datafile '/export/home/dfreneuil/D817F/ DATAFILES/rcat80_01.dbf' size 20M ; Create the 9.2.0 catalog in the 9.2.0 catalog database. # sql syntax for creating logical catalog 8i structure. create tablespace RCAT9I datafile '/export/home/dfreneuil/D920F/ DATAFILES/rcat9i_01.dbf' size 20M ; # sql syntax for creating catalog 8.0.5 user owner. create user RMAN80 identified by rman80 default tablespace RCAT80 temporary tablespace temp quota unlimited on RCAT80 ; grant connect, resource,recovery_catalog_owner to rman80 ; # sql syntax for creating catalog 9i user owner. create user RMAN9I identified by rman9i default tablespace RCAT9I temporary tablespace temp quota unlimited on RCAT9I ; grant connect, resource,recovery_catalog_owner to rman9i ; - make tnsnames.ora OK - Create the 2 catalogs: 9.2.0 catalog views creation. $ rman catalog rman9i/rman9i -- to connect locally. or $ rman catalog rman8i/rman9i@alias to connect through NET8. RMAN> create catalog ; 8.0.5 catalog views creation. Since the catalogs database is an 8.1.7 database, connect to the 8.0.5 catalog via 8.0.5 SQL*Plus. $ sqlplus rman80/rman80@alias_to_rcat80 --> connect from the target machine to the 8.0.5 catalog. SQL> @?/rdbms/admin/catrman.sql Backup an 8.0.5 database with 8.0.5 RMAN into an 8.0.5 catalog in an 9.2.0 catalog database. $ rman rcvcat rman80/rman80@V817 8.0.5 db ----> 8.0.5 RMAN ----> 8.0.5 catalog in 9.2.0 db 9.2.0 db ----> 9.2.0 RMAN ----> 9.2.0 catalog in 9.2.0 db *** END EXTRA *** 24.8 Registering and un-registering the target database: -------------------------------------------------------- Register: --------- Now we must 'register' the target database. Suppose the target database is called 'airm'. Connect to the target and the catalog: $ rman target / catalog rman/rman@rcdb or $ rman system/passw@airm catalog rman/rman@rcdb RMAN-06005 connected to target database: AIRM RMAN-06008 connected to recovery catalog database RMAN>register database And the airm database will be registered in the catalog. If you connected to rcdb before the registering and give the following queries before and after registering airm: SQL> connect system/manager@rcdb Connected. before registering: SQL> select * from rman.db; no rows selected after registering: SQL> select * from rman.db; DB_KEY DB_ID CURR_DBINC_KEY ---------- ---------- -------------- 1 2092303715 2 Unregister: ----------- It's best to unregister the backups from the catalog first: RMAN> list backup of database; RMAN-03022: compiling command: list shows possible backupsets with their numbers fore example 989 RMAN> allocate channel for maintenance type disk; change backupset 989 delete; Next we un-register the target database. You will not use rman, but a special procedure. You must use this procedure with the DB_KEY and DB_ID parameters as values. In SQL*Plus: SQL>execute dbms_rcvcat.unregisterdatabase(1,2092303715) and the airm database will be unregistered. 24.9 Reset of the catalog: -------------------------- If you have opened the target database with the 'RESETLOGS' option, you have in fact created a new 'incarnation' of the database. This information must be 'told' to the recovery catalog via the 'reset database' command: $ rman target sys/passw catalog rman/rman@rcdb RMAN>reset database; -- VALIDATE: -- --------- You can use the VALIDATE option of the BACKUP command to verify that database files exist and are in the correct locations, and have no physical or logical corruptions that would prevent RMAN from creating backups of them. When performing a BACKUP... VALIDATE, RMAN reads the files to be backed up in their entirety, as it would during a real backup. It does not, however, actually produce any backup sets or image copies. If the backup validation discovers corrupt blocks, then RMAN updates the V$DATABASE_BLOCK_CORRUPTION view with rows describing the corruptions. You can repair corruptions using block media recovery, documented in Oracle Database Backup and Recovery Advanced User's Guide. After a corrupt block is repaired, the row identifying this block is deleted from the view. For example, you can validate that all database files and archived logs can be backed up by running a command as follows: BACKUP VALIDATE DATABASE ARCHIVELOG ALL; The RMAN client displays the same output that it would if it were really backing up the files. If RMAN cannot validate the backup of one or more of the files, then it issues an error message. For example, RMAN may show output similar to the following: RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03002: failure of backup command at 08/29/2002 14:33:47 ORA-19625: error identifying file /oracle/oradata/trgt/arch/archive1_6.dbf ORA-27037: unable to obtain file status SVR4 Error: 2: No such file or directory Additional information: 3 -- CONTROLFILE AUTOBACKUP -- ---------------------- Configuring Control File and Server Parameter File Autobackup RMAN can be configured to automatically back up the control file and server parameter file whenever the database structure metadata in the control file changes and whenever a backup record is added. The autobackup enables RMAN to recover the database even if the current control file, catalog, and server parameter file are lost. Because the filename for the autobackup uses a well-known format, RMAN can search for it without access to a repository, and then restore the server parameter file. After you have started the instance with the restored server parameter file, RMAN can restore the control file from an autobackup. After you mount the control file, the RMAN repository is available and RMAN can restore the datafiles and find the archived redo log. You can enable the autobackup feature by running this command: CONFIGURE CONTROLFILE AUTOBACKUP ON; You can disable the feature by running this command: CONFIGURE CONTROLFILE AUTOBACKUP OFF; Backing Up Control Files with RMAN You can back up the control file when the database is mounted or open. RMAN uses a snapshot control file to ensure a read-consistent version. If CONFIGURE CONTROLFILE AUTOBACKUP is ON (by default it is OFF), then RMAN automatically backs up the control file and server parameter file after every backup and after database structural changes. The control file autobackup contains metadata about the previous backup, which is crucial for disaster recovery. If the autobackup feature is not set, then you must manually back up the control file in one of the following ways: .Run BACKUP CURRENT CONTROLFILE .Include a backup of the control file within any backup by using the INCLUDE CURRENT CONTROLFILE option of the BACKUP command .Back up datafile 1, because RMAN automatically includes the control file and SPFILE in backups of datafile 1 Note: If the control file block size is not the same as the block size for datafile 1, then the control file cannot be written into the same backup set as the datafile. RMAN writes the control file into a backup set by itself if the block size is different. A manual backup of the control file is not the same as a control file autobackup. In manual backups, only RMAN repository data for backups within the current RMAN session is in the control file backup, and a manually backed-up control file cannot be automatically restored. 24.11 Create scripts: --------------------- If you are connected to the target and the catalog, you can create and store scripts in the catalog. Example: == XXX RMAN> create script complet_bac1 { 2> allocate channel c1 type disk; 3> allocate channel c2 type disk; 4> backup database; 5> sql 'ALTER SYSTEM ARCHIVE LOG ALL'; 6> backup archivelog all; 7> } RMAN-03022: compiling command: create script RMAN-03023: executing command: create script RMAN-08085: created script complet_bac1 To run such a script: $ rman target sys/passw@airm catalog rman/rman@rcdb RMAN>run { execute scipt complet_bac1; } You can also replace a script: RMAN>replace script b_whole_l0 { # back up whole database and archived logs allocate channel d1 type disk; allocate channel d2 type disk; allocate channel d3 type disk; backup incremental level 0 tag b_whole_l0 filesperset 6 format '/dev/backup/prod1/df/df_t%t_s%s_p%p' -- name of the backup piece (database); sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; backup filesperset 20 format '/dev/backup/prod1/al/al_t%t_s%s_p%p' (archivelog all delete input); } RMAN> SET CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO 'controlfile_%F'; RMAN> BACKUP AS COPY DATABASE; RMAN> RUN { SET CONTROLFILE AUTOBACKUP FORMAT FOR DEVICE TYPE DISK TO '/tmp/%F.bck'; BACKUP AS BACKUPSET DEVICE TYPE DISK DATABASE; } 24.12 Parallization: -------------------- RMAN executes commands serially; that is, it completes the current command before starting the next one. Parallelism is exploited only within the context of a single command. Consequently, if you want 5 datafile copies, issue a single copy command specifying all 5 copies rather than 5 separate copy commands. In the following example, you allocate 5 channels, and then you issued 5 separate copy commands. So, all copy commands are performed one after the other. run { allocate channel c1 type disk; allocate channel c2 type disk; allocate channel c3 type disk; allocate channel c4 type disk; allocate channel c5 type disk; copy datafile 22 to '/dev/prod/backup1/prod_tab5_1.dbf'; copy datafile 23 to '/dev/prod/backup1/prod_tab5_2.dbf'; copy datafile 24 to '/dev/prod/backup1/prod_tab5_3.dbf'; copy datafile 25 to '/dev/prod/backup1/prod_tab5_4.dbf'; copy datafile 26 to '/dev/prod/backup1/prod_tab6_1.dbf'; } To get the copy command run in parallel, use the following command: run { allocate channel c1 type disk; allocate channel c2 type disk; allocate channel c3 type disk; allocate channel c4 type disk; allocate channel c5 type disk; copy datafile 5 to '/dev/prod/backup1/prod_tab5_1.dbf', datafile 23 to '/dev/prod/backup1/prod_tab5_2.dbf', datafile 24 to '/dev/prod/backup1/prod_tab5_3.dbf', datafile 25 to '/dev/prod/backup1/prod_tab5_4.dbf', datafile 26 to '/dev/prod/backup1/prod_tab6_1.dbf'; } 24.13 Creating backups: ----------------------- 1. Image copy and Backup set: ----------------------------- - you can make 'image copies', which are actual complete copies of database files, controlfiles, or archived redologs, to disk. These are not stored in the special RMAN format, and can be used 'ouside' of rman if neccessary. - you can make for example backups of database files in a 'backup set' which are in the special rman format. You must use rman to process them. Examples: - image copy, using the copy command: RMAN>run { allocate channel c1 type disk; copy datafile 1 to '/staging/system01.dbf', datafile 2 to '/staging/data01.dbf', datafile 3 to '/staging/users01.dbf', current controlfile to '/staging/control1.ctl'; } RMAN> run { 2> allocate channel c1 type disk; 3> copy datafile 1 to 'df1.bak'; 4> } - backup set, using the backup command: RMAN> run { allocate channel c1 type disk; backup tablespace users including current controlfile; } RMAN> run { 2> allocate channel c1 type disk; 3> backup tablespace system; 4> } RMAN> This example backs up the tablespace to its default backup location, which is port-specific: on UNIX systems the location is $ORACLE_HOME/dbs. Because you do not specify the format parameter, RMAN automatically assigns the backup a unique filename. 2. Archive mode and No archive mode: ------------------------------------ If the database is in ARCHIVELOG mode, then the target database can be open or closed; you do not need to close the database cleanly (although Oracle recommends you do so that the backup is consistent). If the database is in NOARCHIVELOG mode, then you must close it cleanly prior to taking a backup. The following example shows that a tablespace backup does not work if the database is open and in no archive mode. RMAN> run { 2> allocate channel c1 type disk; 3> backup tablespace users; 4> } RMAN-03022: compiling command: allocate RMAN-03023: executing command: allocate RMAN-08030: allocated channel: c1 RMAN-08500: channel c1: sid=17 devtype=DISK RMAN-03022: compiling command: backup RMAN-03023: executing command: backup RMAN-08008: channel c1: starting full datafile backupset RMAN-08502: set_count=2 set_stamp=482962114 creation_time=10-JAN-03 RMAN-08010: channel c1: specifying datafile(s) in backupset RMAN-00571: =========================================================== RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS =============== RMAN-00571: =========================================================== RMAN-03007: retryable error occurred during execution of command: backup RMAN-07004: unhandled exception during command execution on channel c1 RMAN-10035: exception raised in RPC: ORA-19602: cannot backup or copy active file in NOARCHIVELOG mode RMAN-10031: ORA-19624 occurred during call to DBMS_BACKUP_RESTORE.BACKUPDATAFILE 3. Names and sizes: ------------------- Filenames for Backup Pieces: You can either let RMAN determine a unique name for the backup piece or use the format parameter to specify a name. If you do not specify a filename, RMAN uses the %U substitution variable to guarantee a unique name. The backup command provides substitution variables that allow you to generate unique filenames. Number and Size of Backup Set: Use the backupSpec clause to list what you want to back up as well as specify other useful options. The number and size of backup sets depends on: The number of backupSpec clauses that you specify. The number of input files specified or implied in each backupSpec clause. The number of channels that you allocate. The filesperset parameter, which limits the number of files for a backup set. The setsize parameter, which limits the overall size in bytes of a backup set. The most important rules in the algorithm for backup set creation are: Each allocated channel that performs work in the backup job--that is, that is not idle--generates at least one backup set. By default, this backup set contains one backup piece. RMAN always tries to divide the backup load so that all allocated channels have roughly the same amount of work to do. The maximum upper limit for the number of files per backup set is determined by the filesperset parameter of the backup command. The maximum upper limit for the size in bytes of a backup set is determined by the setsize parameter of the backup command. The filesperset parameter limits the number of files that can go in a backup set. The default value of this parameter is calculated by RMAN as follows: RMAN compares the value 64 to the rounded-up ratio of number of files / number of channels, nd sets filesperset to the lower value. For example, if you back up 70 files with one channel, RMAN divides 70/1, compares this value to 64, and sets filesperset to 64 because it is the lowest value. The number of backup sets produced by RMAN is the rounded-up ratio of number of datafiles / filesperset. For example, if you back up 70 datafiles and filesperset is 64, then RMAN produces 2 backup sets. setsize: Sets the maximum size in bytes of the backup set without specifying a limit to the number of files in the set. filesperset: Sets a limit to the number of files in the backup set without specifying a maximum size in bytes of the set. 4. Examples: ------------ - Backup and Recovery Database: ------------------------------- Other Examples: --------------- $ rman target / catalog rman/rman@rcat To write the output to a log file, specify the file at startup. For example, enter: $ rman target / catalog rman/rman@rcat log /oracle/log/mlog.f Allocate one or more channels of type disk or type 'sbt_tape'. This example backs up all the datafiles as well as the control file. It does not specify a format parameter, so RMAN gives each backup piece a unique name automatically and stores it in the port-specific default location ($ORACLE_HOME/dbs on UNIX). Whole database backups automatically include the current control file, but the current control file does not contain a record of the whole database backup. To obtain a control file backup with a record of the whole database backup, make a backup of the control file after executing the whole database backup. Include a backup of the control file within any backup by specifying the include current controlfile option. Optionally, use the set duplex command to create multiple identical backupsets. run { allocate channel ch1 type disk; backup database; sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; # archives current redo log as well as # all unarchived logs } Optionally, use the format parameter to specify a filename for the backup piece. For example, enter: run { allocate channel ch1 type disk; backup database format '/oracle/backup/%U'; # %U generates a unique filename } Optionally, use the tag parameter to specify a tag for the backup. For example, enter: run { allocate channel ch1 type 'sbt_tape'; backup database tag = 'weekly_backup'; # gives the backup a tag identifier } This script backs up the database and the archived redo logs: RMAN> run { allocate channel ch1 type disk; allocate channel ch2 type disk; backup database; sql 'ALTER SYSTEM ARCHIVE LOG ALL'; backup archivelog all; } RMAN> run { allocate channel ch1 type disk; allocate channel ch2 type disk; backup format 'i:\backup\full_db.bck' (database); sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; backup archivelog all; } - Backup tablespace: -------------------- run { allocate channel ch1 type disk; allocate channel ch2 type disk; allocate channel ch3 type disk; backup filesperset = 3 tablespace inventory, sales include current controlfile; } - Backup datafiles: ------------------- run { allocate channel ch1 type disk; backup (datafile 1,2,3,4,5,6 filesperset 3) datafilecopy '/oracle/copy/tbs_1_c.f'; } RMAN> run { allocate channel c1 type disk; copy datafile 6 to 'F:\oracle\backups\oem01.cpy'; release channel c1; } RMAN> run { allocate channel c1 type disk; backup format 'F:\oracle\backups\oem01.rbu' ( datafile 6 ); release channel c1; } RMAN> run { allocate channel ch1 type disk; allocate channel ch2 type disk; allocate channel ch3 type disk; backup (datafile 1,2,3 filesperset = 1 channel ch1) (datafilecopy '/oracle/copy/cf.f' filesperset = 2 channel ch2) (archivelog from logseq 100 until logseq 102 thread 1 filesperset = 3 channel ch3); } - Backup archived redologs: --------------------------- To back up archived logs, issue backup archivelog with the desired filtering options: run { allocate channel ch1 type 'sbt_tape'; backup archivelog all # Backs up all archived redo logs. delete input; # Optionally, delete the input logs } You can also specify a range of archived redo logs by time, SCN, or log sequence number. This example backs up all archived logs created more than 7 and less than 30 days ago: run { allocate channel ch1 type disk; backup archivelog from time 'SYSDATE-30' until time 'SYSDATE-7'; } - Incremental backups: ---------------------- This example makes a level 0 backup of the database: run { allocate channel ch1 type disk; backup incremental level = 0 database; } This example makes a level 1 backup of the database: run { allocate channel ch1 type disk; backup incremental level = 1 database; } Further examples: ------------------ Your database has to be in archive log mode for this script to work RMAN> run { 2> # backup the database to disk 3> allocate channel d1 type disk; 4> backup 5> full 6> tag full_db 7> format '/backups/db_%t_%s_p%p' 8> (database); 9> release channel d1; 10> } ---- This script will backup all archive logs. Your database has to be in archive log mode for this script to work. RMAN> run { 2> allocate channel d1 type disk; 3> backup 4> format '/backups/log_t%t_s%s_p%p' 5> (archivelog all); 6> release channel d1; 7> } ---- This script will backup all the datafiles. resync catalog; run { allocate channel c1 type disk; copy datafile 1 to 'C:\rman1.dbf'; copy datafile 2 to 'C:\rman2.dbf'; copy datafile 3 to 'C:\rman3.dbf'; copy datafile 4 to 'C:\rman4.dbf'; copy datafile 5 to 'C:\rman5.dbf'; } exit echo exiting after successful hot backup using RMAN ----- run { sql 'alter database close'; allocate channel d1 type disk; backup full tag full_offline_backup format 'c:\backup\db_t%t_s%s_p%p' (database); release channel d1; sql 'alter database open'; } 5. Complete Examples: --------------------- *************************************************************** L=0 BACKUP run { allocate channel d1 type disk; backup incremental level = 0 tag db_whole_l0 format 'i:\backup\l0_%d_t%t_s%s_p%p' (database); sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; backup format 'i:\backup\log_%d_t%t_s%s_p%p' (archivelog all); } or run { allocate channel d1 type disk; allocate channel d2 type disk; backup incremental level = 0 tag db_whole_l0 format 'i:\backup\l0_%d_t%t_s%s_p%p' (database channel d1); sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; backup format 'i:\backup\log_%d_t%t_s%s_p%p' (archivelog all channel d2); } L=1 BACKUP run { allocate channel d1 type disk; backup incremental level = 1 tag db_whole_l1 format 'i:\backup\l1_%d_t%t_s%s_p%p' (database); sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; backup format 'i:\backup\log_%d_t%t_s%s_p%p' (archivelog all); } ***************************************************************** RMAN>create script db_whole_l0 { # back up whole database and archived logs allocate channel d1 type disk; backup incremental level 0 tag db_whole_l0 filesperset 15 format 'i:\backup\l0_%d_t%t_s%s_p%p' -- name of the backup piece (database); sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; backup filesperset 20 format 'i:\backup\log_%d_t%t_s%s_p%p' (archivelog all delete input); } RMAN>create script db_whole_l1 { # back up whole database and archived logs allocate channel d1 type disk; backup incremental level 1 tag db_whole_l0 filesperset 15 format 'i:\backup\l1_%d_t%t_s%s_p%p' -- name of the backup piece (database); sql 'ALTER SYSTEM ARCHIVE LOG CURRENT'; backup filesperset 20 format 'i:\backup\log_%d_t%t_s%s_p%p' (archivelog all delete input); } On sunday : schedule RMAN>run { execute scipt db_whole_l0; } Other days: schedule RMAN>run { execute scipt db_whole_l1; } ********************************************** replace script backup_all_archives { execute script alloc_all_disks; backup filesperset 50 format '/bkup/SID/%d_al_t%t_s%s_p%p' (archivelog all delete input); execute script rel_all_disks; } # Incremental level 0 (whole) database backup # The control file is automatically included each time file 1 of the # system tablespace is backed up. # replace script backup_db_level_0_disk { # execute script alloc_all_disks; # set maxcorrupt for datafile 1 to 0; run { allocate channel c2 type disk; backup incremental level = 0 tag backup_db_level_0 # The skip inaccessible clause ensures the backup will continue # if any of the datafiles are inaccessible. skip inaccessible filesperset 9 format 'i:\backup\L0_%d.bck' (database); sql 'alter system archive log current'; execute script backup_all_archives; } ************************************************************* -- SUNDAY LEVEL 0 BACKUP run { allocate channel d1 type disk; setlimit channel d1 kbytes 2097150 maxopenfiles 32 readrate 200; set maxcorrupt for datafile 1,2,3,4,5,6 to 0; backup incremental level 0 cumulative skip inaccessible tag sunday_level_0 format 'c:\temp\df_t%t_s%s_p%p' database; copy current controlfile to 'c:\temp\sunday.ctl'; sql 'alter system archive log current'; backup format 'c:\temp\al_t%t_s%s_p%p' archivelog all delete input; release channel d1; } -- MONDAY LEVEL 2 BACKUP run { allocate channel d1 type disk; setlimit channel d1 kbytes 2097150 maxopenfiles 32 readrate 200; set maxcorrupt for datafile 1,2,3,4,5,6 to 0; backup incremental level 2 cumulative skip inaccessible tag monday_level_2 format 'c:\temp\df_t%t_s%s_p%p' database; copy current controlfile to 'c:\temp\monday.ctl'; sql 'alter system archive log current'; backup format 'c:\temp\al_t%t_s%s_p%p' archivelog all delete input; release channel d1; } -- TUESDAY LEVEL 2 BACKUP run { allocate channel d1 type disk; setlimit channel d1 kbytes 2097150 maxopenfiles 32 readrate 200; set maxcorrupt for datafile 1,2,3,4,5,6 to 0; backup incremental level 2 cumulative skip inaccessible tag tueday_level_2 format 'c:\temp\df_t%t_s%s_p%p' database; copy current controlfile to 'c:\temp\tuesday.ctl'; sql 'alter system archive log current'; backup format 'c:\temp\al_t%t_s%s_p%p' archivelog all delete input; release channel d1; } -- WEDNESDAY LEVEL 2 BACKUP run { allocate channel d1 type disk; setlimit channel d1 kbytes 2097150 maxopenfiles 32 readrate 200; set maxcorrupt for datafile 1,2,3,4,5,6 to 0; backup incremental level 2 cumulative skip inaccessible tag wednesday_level_2 format 'c:\temp\df_t%t_s%s_p%p' database; copy current controlfile to 'c:\temp\wednesday.ctl'; sql 'alter system archive log current'; backup format 'c:\temp\al_t%t_s%s_p%p' archivelog all delete input; release channel d1; } -- THURSDAY LEVEL 1 BACKUP run { allocate channel d1 type disk; setlimit channel d1 kbytes 2097150 maxopenfiles 32 readrate 200; set maxcorrupt for datafile 1,2,3,4,5,6 to 0; backup incremental level 1 cumulative skip inaccessible tag thursday_level_1 format 'c:\temp\df_t%t_s%s_p%p' database; copy current controlfile to 'c:\temp\thursday.ctl'; sql 'alter system archive log current'; backup format 'c:\temp\al_t%t_s%s_p%p' archivelog all delete input; release channel d1; } -- FRIDAY LEVEL 2 BACKUP run { allocate channel d1 type disk; setlimit channel d1 kbytes 2097150 maxopenfiles 32 readrate 200; set maxcorrupt for datafile 1,2,3,4,5,6 to 0; backup incremental level 2 cumulative skip inaccessible tag friday_level_2 format 'c:\temp\df_t%t_s%s_p%p' database; copy current controlfile to 'c:\temp\friday.ctl'; sql 'alter system archive log current'; backup format 'c:\temp\al_t%t_s%s_p%p' archivelog all delete input; release channel d1; } -- SATURDAY LEVEL 2 BACKUP run { allocate channel d1 type disk; setlimit channel d1 kbytes 2097150 maxopenfiles 32 readrate 200; set maxcorrupt for datafile 1,2,3,4,5,6 to 0; backup incremental level 2 cumulative skip inaccessible tag saturday_level_2 format 'c:\temp\df_t%t_s%s_p%p' database; copy current controlfile to 'c:\temp\saturday.ctl'; sql 'alter system archive log current'; backup format 'c:\temp\al_t%t_s%s_p%p' archivelog all delete input; release channel d1; } 6. Third Party: --------------- You can use rman in combination with third party storage managers. In this case, rman is used with a MML library and possibly some API that uses it's own configuration files, for example: backup.scr script: run { allocate channel t1 type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=c:\RMAN\scripts\tdpo.opt)'; allocate channel t2 type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=c:\RMAN\scripts\tdpo.opt)'; backup filesperset 5 format 'df_%t_%s_%p' (database); release channel t1; release channel t2; } run { allocate channel d1 type 'sbt_tape' connect 'internal/manager@scdb2' parms 'ENV=(TDPO_OPTFILE=/usr/tivoli/tsm/client/oracle/bin64/tdpo.opt)'; allocate channel d2 type 'sbt_tape' connect 'internal/manager@scdb1' parms 'ENV=(TDPO_OPTFILE=/usr/tivoli/tsm/client/oracle/bin64/tdpo.opt)'; backup format 'ctl_t%t_s%s_p%p' tag cf (current controlfile); backup full filesperset 8 format 'db_t%t_s%s_p%p' tag fulldb (database); release channel d1; release channel d2; } The PARMS parameter sends instructions to the media manager. For example, the following vendor-specific PARMS setting instructs the media manager to back up to a volume pool called oracle_tapes: PARMS='ENV=(NSR_DATA_VOLUME_POOL=oracle_tapes)' parms='ENV=(DSMO_FS=oracle)' Another example: RUN { ALLOCATE CHANNEL c1 DEVICE TYPE sbt PARMS='ENV=(NSR_SERVER=tape_srv,NSR_GROUP=oracle_tapes)'; } If you do not receive an error message, then Oracle successfully l oaded the shared library. However, channel allocation can fail with the ORA-27211 error: To delete an old backup: run { allocate channel for delete type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=c:\RMAN\scripts\tdpo.opt)'; change backupset primary_key delete; } To schedule scripts: -------------------- orcschedppim.cmd rem ================================================== rem orcsched.cmd rem ================================================== rem ================================================== rem set rman executable rem ================================================== set ora_exe=d:\oracle\ora81\bin\rman rem ================================================== rem set script and log directory rem ================================================== rem set ora_script_dir=d:\oracle\scripts\ set ora_script_dir=c:\progra~1\tivoli\tsm\agentoba\ rem ================================================== rem run the backup script rem ================================================== %ora_exe% target system/manager@ppim rcvcat rman_db1/rman_db1@orcl cmdfile %ora_script_dir%bkdbppim.scr msglog %ora_script_dir%bkdbppim.log bkdbppim.scr run { allocate channel t1 type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=C:\Progra~1\Tivoli\TSM\AgentOBA\tdpoppim.opt)'; allocate channel t2 type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=C:\Progra~1\Tivoli\TSM\AgentOBA\tdpoppim.opt)'; backup filesperset 5 format 'df_%t_%s_%p' (database); release channel t1; release channel t2; } ------------------------------------ Remarks: -------- The following is what needs to be changed. - Old Way allocate channel for maintenance type 'sbt_tape' parms 'ENV=(DSMO_NODE=tora, DSMI_ORC_CONFIG=/opt/tivoli/tsm/client/oracle/bin/dsm.opt)' allocate channel t1 type 'sbt_tape' parms > 'ENV=(DSMO_NODE=rx_r50, > DSMI_CONFIG=/usr/tivoli/tsm/client/ba/bin/dsm.opt, > DSMO_PSWDPATH=/usr/tivoli/tsm/client/oracle/bin, > DSMI_DIR=/usr/tivoli/tsm/client/ba/bin, > DSMO_AVG_SIZE00)'; > - New Way allocate channel for maintenance type 'sbt_tape' parms 'ENV=(TDPO_OPTFILE=/opt/tivoli/tsm/client/oracle/bin/tdpo.opt)' Contents of tdpo.opt DSMI_ORC_CONFIG /opt/tivoli/tsm/client/oracle/bin/dsm.opt DSMI_LOG /opt/tivoli/tsm/client/oracle/bin/tdpoerror.log TDPO_FS rman_fs TDPO_NODE tora *TDPO_OWNER TDPO_PSWDPATH /opt/tivoli/tsm/client/oracle/bin *TDPO_DATE_FMT 1 *TDPO_NUM_FMT 1 *TDPO_TIME_FMT 1 *TDPO_MGMT_CLASS2 mgmtclass2 *TDPO_MGMT_CLASS3 mgmtclass3 *TDPO_MGMT_CLASS4 mgmtclass4 It is recomended TDP_NUM_BUFFERS be set to a value of 1 only. 7. Recovery: ------------ A restore can be as easy as: RMAN> RESTORE DATABASE; RMAN> RECOVER DATABASE; Or a single tablespace: Restore the tablespace or datafile with the RESTORE command, and recover it with the RECOVER command. (Use configured channels, or if desired, use a RUN block and allocate channels to improve performance of the RESTORE and RECOVER commands.) RMAN> RESTORE TABLESPACE users; RMAN> RECOVER TABLESPACE users; If RMAN reported no errors during the recovery, then bring the tablespace back online: RMAN> SQL 'ALTER TABLESPACE users ONLINE'; Use the RMAN restore command to restore datafiles, control files, or archived redo logs from backup sets or image copies. RMAN restores backups from disk or tape, but image copies only from disk. Restore files to either: - The default location, which overwrites the files with the same name. - A new location specified by the set newname command. Restoring the Database to its Default Location ---------------------------------------------- If you do not specify set newname commands for the datafiles during a restore job, the database must be closed or the datafiles must be offline. RMAN> run { allocate channel c1 type 'sbt_tape'; restore database; recover database; } run { set until logseq 5 thread 1; allocate auxiliary channel dupdb1 type disk; duplicate target database to dupdb; } Examples Restoring the Database to a point in time (same incarnation) --------------------------------------------------------------------- Example 1: ---------- RMAN> run 2 { 3 set until time '23-DEC-2006 13:45:00'; 4 restore database; 5 recover database; 6 } Example 2: ---------- To recover the database until a specified time, SCN, or log sequence number: After connecting to the target database and, optionally, the recovery catalog database, ensure that the database is mounted. If the database is open, shut it down and then mount it: SHUTDOWN IMMEDIATE; STARTUP MOUNT; Determine the time, SCN, or log sequence that should end recovery. For example, if you discover that a user accidentally dropped a tablespace at 9:02 a.m., then you can recover to 9 a.m. --just before the drop occurred. You will lose all changes to the database made after that time. You can also examine the alert.log to find the SCN of an event and recover to a prior SCN. Alternatively, you can determine the log sequence number that contains the recovery termination SCN, and then recover through that log. For example, query V$LOG_HISTORY to view the logs that you have archived. RECID STAMP THREAD# SEQUENCE# FIRST_CHAN FIRST_TIM NEXT_CHANG ---------- ---------- ---------- ---------- ---------- --------- ---------- 1 344890611 1 1 20037 24-SEP-02 20043 2 344890615 1 2 20043 24-SEP-02 20045 3 344890618 1 3 20045 24-SEP-02 20046 Perform the following operations within a RUN command: Set the end recovery time, SCN, or log sequence. If specifying a time, then use the date format specified in the NLS_LANG and NLS_DATE_FORMAT environment variables. If automatic channels are not configured, then manually allocate one or more channels. Restore and recover the database. The following example performs an incomplete recovery until November 15 at 9 a.m. RUN { SET UNTIL TIME 'Nov 15 2002 09:00:00'; # SET UNTIL SCN 1000; # alternatively, specify SCN # SET UNTIL SEQUENCE 9923; # alternatively, specify log sequence number RESTORE DATABASE; RECOVER DATABASE; } If recovery was successful, then open the database and reset the online logs: ALTER DATABASE OPEN RESETLOGS; Moving the Target Database to a New Host with the Same File System ------------------------------------------------------------------ A media failure may force you to move a database by restoring a backup from one host to another. You can perform this procedure so long as you have a valid backup and a recovery catalog or control file. Because your restored database will not have the online redo logs of your production database, you will need to perform incomplete recovery up to the lowest SCN of the most recently archived redo log in each thread and then open the database with the RESETLOGS option. To restore the database from HOST_A to HOST_B with a recovery catalog: Copy the initialization parameter file for HOST_A to HOST_B using an operating system utility. Connect to the HOST_B target instance and HOST_A recovery catalog. For example, enter: % rman target sys/change_on_install@host_b catalog rman/rman@rcat Start the instance without mounting it: startup nomount Restore and mount the control file. Execute a run command with the following sub-commands: Allocate at least one channel. Restore the control file. Mount the control file. run { allocate channel ch1 type disk; restore controlfile; alter database mount; } Because there may be multiple threads of redo, use change-based recovery. Obtain the SCN for recovery termination by finding the lowest SCN among the most recent archived redo logs for each thread. Start SQL*Plus and use the following query to determine the necessary SCN: SELECT min(scn) FROM (SELECT max(next_change#) scn FROM v$archived_log GROUP BY thread#); Execute a run command with the following sub-commands: Set the SCN for recovery termination using the value obtained from the previous step. Allocate at least one channel. Restore the database. Recover the database. Open the database with the RESETLOGS option. run { set until scn = 500; # use appropriate SCN for incomplete recovery allocate channel ch1 type 'sbt_tape'; restore database; recover database; alter database open resetlogs; } Moving the Target Database to a New Host with a different File System --------------------------------------------------------------------- Follow the procedure as above, but now use the 'set newname' command. run { set until scn 500; # use appropriate SCN for incomplete recovery allocate channel ch1 type disk; set newname for datafile 1 to '/disk1/%U'; # rename each datafile manually set newname for datafile 2 to '/disk1/%U'; set newname for datafile 3 to '/disk1/%U'; set newname for datafile 4 to '/disk1/%U'; set newname for datafile 5 to '/disk1/%U'; set newname for datafile 6 to '/disk2/%U'; set newname for datafile 7 to '/disk2/%U'; set newname for datafile 8 to '/disk2/%U'; set newname for datafile 9 to '/disk2/%U'; set newname for datafile 10 to '/disk2/%U'; alter database mount; restore database; switch datafile all; # points the control file to the renamed datafiles recover database; alter database open resetlogs; } Warning: restore with use catalog: If you issue switch commands, RMAN considers the restored database as the target database, and the recovery catalog becomes corrupted. If you do not issue switch commands, RMAN considers the restored datafiles as image copies that are candidates for future restore operations. restore with no catalog: If you issue switch commands, RMAN considers the restored database as the target database. If you do not issue switch commands, the restore operation has no effect on the repository. Restoring a tablespace: ----------------------- Suppose tablespace DATA_BIG has become unusable. run { allocate channel ch1 type disk; restore tablespace data_big; } run { allocate channel ch1 type disk; recover tablespace data_big; } This script will perform datafile recovery RMAN> run { 2> allocate channel d1 type disk; 3> sql "alter tablespace users offline immediate"; 4> restore datafile 5; 5> recover datafile 5; 6> sql "alter tablespace users online"; 7> release channel d1; 8> } RMAN> run { allocate channel ch1 type disk; restore database; recover database; alter database open resetlogs; } Duplexing the Target Database to a New Host: --------------------------------------------- - create instance on second host - create init.ora, password file etc.. - create similar directories on second host - make sure net8 works from target en rman to second host - startup nomount - neccesary archived redologs are present on second host $ rman target sys/target_pwd@target_str catalog rman/cat_pwd@cat_str auxiliary sys/aux_pwd@aux_str run { allocate auxiliary channel ch1 type 'sbt_tape'; duplicate target database to dupdb nofilenamecheck; } run { # allocate at least one auxiliary channel of type disk or tape allocate auxiliary channel dupdb1 type 'sbt_tape'; . . . # set new filenames for the datafiles set newname for datafile 1 TO '$ORACLE_HOME/dbs/dupdb_data_01.f'; set newname for datafile 2 TO '$ORACLE_HOME/dbs/dupdb_data_02.f'; . . . # issue the duplicate command duplicate target database to dupdb # create at least two online redo log groups logfile group 1 ('$ORACLE_HOME/dbs/dupdb_log_1_1.f', '$ORACLE_HOME/dbs/dupdb_log_1_2.f') size 200K, group 2 ('$ORACLE_HOME/dbs/dupdb_log_2_1.f', '$ORACLE_HOME/dbs/dupdb_log_2_2.f') size 200K; } 24.14 Common RMAN errors: ------------------------- What are the common RMAN errors (with solutions)? Some of the common RMAN errors are: PROBLEM 1. ---------- RMAN-20242: Specification does not match any archivelog in the recovery catalog. Add to RMAN script: sql 'alter system archive log current'; PROBLEM 2. ---------- RMAN-06089: archived log xyz not found or out of sync with catalog Execute from RMAN: change archivelog all validate; PROBLEM 3. ---------- fact: Oracle Server - Enterprise Edition 8 fact: Oracle Server - Enterprise Edition 9 fact: Recovery Manager (RMAN) symptom: RMAN backup fails symptom: RMAN-10035: exception raised in RPC symptom: ORA-19505: failed to identify file symptom: ORA-27037: unable to obtain file status symptom: SVR4 error:2:no such file or directory cause: Datafile existed in previous backup set, but has been subsequently removed or renamed. fix: Resync the RMAN Catalog $ rman target sys/@target catalog rman/@catalog RMAN> resync catalog; Or Validate the backup pieces. $ rman target sys/@target catalog rman/@catalog RMAN> allocate channel for maintenance type disk; RMAN> crosscheck backup; RMAN> resync catalog; PROBLEM 4. ---------- RMAN> connect target sys/change_on_install@TARGETDB RMAN-00569: ================error message stack follows RMAN-04005: error from target database: ORA-01017: invalid username/password; logon denied Problem Explanation: Recovery Manager automatically requests a connection to the target database as SYSDBA. Solution Description: Recovery Manager automatically requests a connection to the target database as SYSDBA. In order to connect to the target database as SYSDBA, you must either: 1. Be part of the operating system DBA group with respect to the target database. This means that you have the ability to CONNECT INTERNAL to the trget database without a password. - or - 2. Have a password file setup. This requires the use of the "orapwd" command and the initialization parameter "remote_login_passwordfile". See Chapter 1 of the Oracle8(TM) Server Administrator's Guide, Release 8.0 for details. Note that changes to the password file will not take affect until after the database is shutdown and restarted. For Unix, also ensure TWO_TASK is _not_ set. e.g. % env | grep -i two If set, unset it. % unsetenv TWO_TASK PROBLEM 5. --------- RMAN cannot connect to the target database through a multi-threaded server (MTS) dispatcher: it requires a dedicated server process Create a net service name in the tnsnames.ora file that connects to the non-shared SID. For example, enter: inst1_ded = (description= (address=(protocol=tcp)(host=inst1_host)(port1521)) (connect_data=(service_name=inst1)(server=dedicated)) ) $ rman target sys/oracle@inst1_ded catalog rman/rman@rcat PROBLEM 6. --------- No MML libary found. RMAN will: 1. Attempts to load the library indicated by the SBT_LIBRARY parameter in the ALLOCATE CHANNEL or CONFIGURE CHANNEL command. If the SBT_LIBRARY parameter is not specified, then Oracle proceeds to the next step. 2. Attempts to load the default media management library. The filename of the default library is operating system specific. On UNIX, the library filename is $ORACLE_HOME/lib/libobk.so, with the extension name varying according to platform: .so, .sl, .a, and so forth. On Windows NT the library is named %ORACLE_HOME%\bin\orasbt.dll. If Oracle is unable to locate the MML library,then RMAN issues an ORA-27211 error and exits. Whenever channel allocation fails, Oracle writes a trace file to the USER_DUMP_DEST directory. The following shows sample output: SKGFQ OSD: Error in function sbtinit on line 2278 SKGFQ OSD: Look for SBT Trace messages in file /oracle/rdbms/log/sbtio.log SBT Initialize failed for /oracle/lib/libobk.so 24.15 RMAN 10g Notes: --------------------- ========================== 25. UPGRADE AND MIGRATION: ========================== 25.1 Version and release numbers: --------------------------------- Oracle 7 -> 8,8i,9i Oracle 8 -> 8i Oracle 8.1.x -> 8.1.y Oracle 8,8i -> 9i Upgrade: move upwarde from one release in the same version to a higher release within the same base version, for example 8.1.6 -> 8.1.7 Migration: move to a different version, for example 7.4.3 -> 8.1.5 Patches : bugfixes Patchset : smaller patches combined to latest patchset Example version: 8.1.6.2 -> 8=version,1=release number,6=maintenance release number,2=patch number Exp Imp matrix: --------------- 1. Migration to Oracle9i release 1 - 9.0.1.x : ------------------------------------------- Direct migration with a full database export and full database import is only supported if the source database is: - Oracle7 : 7.3.4 - Oracle8 : 8.0.6 - Oracle8i: 8.1.5 or 8.1.6 or 8.1.7 Migration to Oracle9i release 2 - 9.2.0.x : ------------------------------------------- Direct migration with a full database export and full database import is only supported if the source database is: - Oracle7 : 7.3.4 - Oracle8 : 8.0.6 - Oracle8i: 8.1.7 - Oracle9i: 9.0.1 Tools that can be used to migrate from one version to another: -------------------------------------------------------------- - exp/imp - MIG Migration Utility - ODMA Oracle Data Migration Assistant There also exists the "Migration Workbench" for migrating Access, SQL Server etc.. to Oracle. 25.2 Migration From 7 to 8,8i: ------------------------------ Take into account the following: - Changed standard directories of init, alert, dump - Changed and obsolete init.ora parameters - Changed and obsolete sqlnet.ora, tnsnames.ora and listener.ora parameters - Rowid values have changed from "restricted" to "extended" format Obsolete init.ora parameters: init_sql_files lm_domains lm_non_fault_tolerant parallel_default_max_scans parallel_default_scansize sequence_cache_hash_buckets serializable session_cached_cursors v733_plans_enabled Change init.ora parameters: compatible snapshot_refresh_interval -> job_queue_interval snapshot_refresh_process -> job_queue_processes db_writers -> dbwr_io_slaves user_dump_dest, background_dump_dest, ifile Three main tools: - exp/imp OWNER= or FULL exp/imp In case of a full exp/imp you must run catalog.sql of new database - Migration utility This is a command line utility. From 7 to 8 or higher: the Rowid will not be changed automatically. Migration utility will create a "conversion file" instance_name.dbf Move this file to the /dbs directory of Oracle 8,9. Startup svrmgrl or sqlplus alter database convert; alter database open resetlogs; - ODMA This tool uses a GUI. 25.2 Example Upgrade of 8.1.6 to 9 using ODMA: ---------------------------------------------- 1. Install the Oracle 9i software in it's own ORACLE_HOME. 2. Prepare the original init.ora DB_DOMAIN=correct domain JOB_QUEUE_PROCESS=0 AQ_TM_PROCESSES=0 REMOTE_LOGIN_PASSWORDFILE=NONE 3. Resize the SYSTEM tablespace to have more than 100M free 4. Prepare the system rollbacksegment to be big enough alter rollback segment system storage(maxextents 505 optimal null next 1M); 5. Verify that SYSTEM is the default tablespace for SYS and SYSTEM 6. Make sure there is no user MIGRATE. ODMA will use a user called MIGRATE. 7. Shutdown the database cleanly. 8. Make a backup 9. Setup the environment variables for the 9i software. Also, ODMA uses the java GUI, just like the OUI 10. Start ODMA $ cd $ORACLE_HOME/bin $ odma 11. Basically, follow the instructions. ODMA will ask you the instance that must be upgraded. On unix, this is read from the oratab file. Then it will ask you to confirm both the old and new ORACLE_HOME. It will also ask for the location of the init.ora file. Then it will proceed in the upgrade. The upgrade is primarily about the datadictionary. 12. When ODMA is ready, do the following: check the alert log and other logs Also check oratab, optionally run utlrp.sql to automatically rebuild any invalid objects. Check for invalid objects and check indexes. Analyze all tables plus indexes. 25.3 Example Upgrade of 8.1.6 to 8.1.7: --------------------------------------- 1. Install the new Oracle software in a different $ORACLE_HOME For example $ cd $ORACLE_BASE $ cd product $ ls 8.1.6 8.1.7 Backup and shutdown the 8.1.6 database, and stop the listener 2. Set the correct env variables for 8.1.7 3. Create a softlink in the new $ORACLE_HOME/dbs to the init.ora in the $ORACLE_BASE/admin/sid/pfile directory Startup the database with new Oracle release 4. Startup the database using the new Oracle software sqlplus internal (or via svrmgrl) startup restrict; 5. Run the upgrade script $ORACLE_HOME/rdbms/admin/u0801060.sql This will also rebuild the datadictionary (catalog, catproc) 6. You optionally run utlrp.sql to automatically rebuild any invalid objects 7. Change on unix oratab for new $ORACLE_HOME 8. Change listener.ora for $ORACLE_HOME value 9. Set COMPATIBLE in init.ora 10. Checks: check the alert log and other logs Also check oratab, optionally run utlrp.sql to automatically rebuild any invalid objects. Check for invalid objects and check indexes. Analyze all tables plus indexes. ===================== 26. Some info on Rdb: ===================== Rdb is most often seen on Digtal unix, or OpenVMS VAX, or OpenVMS alpha, but there exists a port to NT / 2000 as well. Samples directory: ------------------ - digital unix: /usr/lib/dbs/vnn/examples - OpenVMS: SQL$EXAMPLE In digital unix, to create a sample database: $/usr/lib/dbs/sql/vnn/examples/personnel : S, M, MSDB : enter a directory where you want the database to be created. $/usr/lib/dbs/sql/vnn/examples/personnel m /tmp/ Invoking SQL: -------------- - In OpenVMS. Create a symbol $ SQL:==$SQL$ $ SQL SQL> - In digital unix: $ SQL SQL> Attach to database: ------------------- SQL>ATTACH 'FILENAME mf_personnel'; SQL>ATTACH 'FILENAME DISK$1:[GERALDO.DB]SUPPLIES MULTISCHEMA IS OFF' Detach from database: --------------------- SQL>exit $ or SQL>DISCONNECT DEFAULT; SQL> Editing a SQL Statement: ------------------------ SQL>EDIT ... EXIT OpenVMS: Defining a Logical name for a database: ------------------------------------------------ $ DEFINE SQL$DATABASE DISK01:[FIELDMAN.DBS]mf_personnel You do not need to attach to the database anymore. Digital unix: Defining a configuration parameter: ------------------------------------------------- $ SQL_DATABASE /usr/fieldman/dbs/mf_personnel SHOW Statements: ---------------- SQL> SHOW TABLES -- shows all tables SQL> SHOW TABLE * SQL> SHOW ALL TABLES SQL> SHOW TABLE WORK_STATUS -- displays info about table WORK_STATUS SQL> SHOW VIEWS -- shows all views SQL> SHOW VIEW CURRENT_SALARY -- shows info about this view only SQL> SHOW DOMAINS -- display all domains SQL> SHOW DOMAIN DATE_DOM SQL> SHOW INDEXES SQL> SHOW INDEXES ON SALARY_HISTORY SQL> SHOW INDEX DEG_EMP_ID SQL> SHOW DATABASE -- returns the database name SQL> SHOW STORAGE AREAS Single file or multifile database: ---------------------------------- A database that stores tables in one file (file type .rdb) is a single file database. Alternately, you can have a database in which system information is stored in a database root file (.rdb) and the data and metadata are stored in one or more storage area files (type .rda). Single file: - a database root file which contains all user data and information about the status of all database operations. - a snapshot file (.snp file) which contains copies of rows (before images) that are beiing modified by users updating the database. Multifile: - a database root file which contains information about the status of all database operations. - a storage area file, .rda file, for the system tables (RDB$SYSTEM) - one or more .rda files for user data. - snapshot files for each .rda file and for the database root file. Create multifile database example: ---------------------------------- $ SQL SQL> CREATE DATABASE FILENAME mf_personnel_test cont> ALIAS MF_PERS cont> RESERVE 6 JOURNALS cont> RESERVE 15 STORAGE AREAS cont> DEFAULT STORAGE AREA default_area cont> SYSTEM INDEX COMPRESSION IS ENABLED cont> CREATE STORAGE AREA default_area FILENAME default_area cont> CREATE STORAGE AREA RDB$SYSTEM FILENAME pers_system_area; Datatypes: ---------- Rdb Oracle -------------------------------------------------- CHAR CHAR, NCHAR VARCHAR VARCHAR2, NVARCHAR2 SMALLINT (16 bits) NUMBER(L,P) INTEGER (32 bits) can be used with NUMBER(L,P) a scale factor INTEGER(2) BIGINT (64 bits) RAW, LONG, LONG RAW, VARYING DATE ANSI (year, month, day) TIME INTERVAL TIMESTAMP (year, month, day, DATE hours, min, sec) DATE VMS ODBC for RDB: ------------- --------------- The current driver version is 3.00.02.05 which doesnt work, and the older driver version (which does work) is 2.10.17.00 (DriverConf1 outputs attached). --------------- I am trying to run a DTS job to import data from an Oracle 7.3 RDB (DEC) platform into SLQ Server 2000. I have an odbc connection set up and I am using it in MS Access 2000 to view the table that I want to import. When I create the job in SQL Server, I can preview the data and everything looks fine, as in the Access table, but when I try and run the job I get an: [ORACLE][ODBC]Function Sequence Error error message. Any experience with these type of errors and RDB. Thanks, John Campbell This can - I understand - occur where the version of the ODBC drivers on the NT box with SQL Server running is incompatible with the services running on the VMS box. I can't remember the various numbers I'm afraid (or even where I found the stuff - it was some time ago). We're running VMS 7.2-1 and Oracle 7.3 and found that this produced a similar error with the most recent version of the Oracle ODBC Drivers for RdB - but we have no problems running the v2.10 drivers (v2.10.17 to be exact). HTH --------------- ODBC driver for RDB uses SQSAPI32.ini JInitiator: ----------- Oracle heeft deze standaard aangepast, specifiek gericht op het uitvoeren van Webforms. Deze aanpassingen houden verband met stabiliteit (bugfixes) en performance verbetering, zoals JAR file caching, incremental JAR file loading en applet caching. Met behulp van JInitiator kunnen Oracle Forms in een browser (Webforms) worden uitgevoerd. JInitiator is géén JVM, maar een extensie op de JVM standaard, waarmee Oracle Webforms op een stabiele én ondersteunde wijze in een browser kunnen worden uitgevoerd. JInitiator is alleen beschikbaar voor het Windows platform. Op dit moment is het niet mogelijk om Webforms uit te voeren in de standaard Microsoft JVM. Jinitiator zal in de volgende release niet meer terugkeren. Webforms wordt gecertificeerd op de standaard Java Plugin. De Microsoft JVM conformeert zich ook aan deze standaard (géén certificatie), waardoor Webforms op termijn in een standaard Microsoft Internet Explorer browser uitgevoerd zal kunnen worden. Dit kan echter pas met zekerheid gesteld worden na grondig testen. Installatie JInitiator JInitiator wordt bij het eerste gebruik automatisch gedownload vanaf de Application Server. Overigens kan de JInitiator ook handmatig worden geïnstalleerd op de client machines. ============================ 28. Some info on IFS ============================ First some remarks about IFS in versions 9.0.2 and 9.0.3: 9.0.2 ===== In version 9.0.2, IFS (Internet File System) is a separate product. 9.0.3 ===== In version 9.0.3, CM SDK runs in conjunction with Oracle9i Application Server and an Oracle9i database. The Oracle Content Management SDK (Oracle CM SDK) is the new name for the product formerly known as the Oracle Internet File System (Oracle 9iFS). This new naming is official as of version 9.0.3. Oracle CM SDK runs in conjunction with Oracle9i Application Server and an Oracle9i database. Written entirely in Java, Oracle CM SDK is an extensible content management system with file server convenience. 27.1 IFS 9.0.2 -------------- -------------- We first will turn our attention to iFS 9.0.2: ---------------------------------------------- The Oracle 9i database stores all content that comprises the filesystem, from the files themselves to metadata like owners and group information. On most occasions, 9iFS stores the files contents as LOB's in the database. Tools: ------ - Oracle 9iFS Configuration Assistant. Allows you to create a new 9iFS Domain, and add nodes etc.. - Oracle 9IFS Credential Manager Configuration Assistant. To change the default credential manager to be applied to each user. - OEM for 9iAS website (9iAS Home Page) You can manage 9iFS from the 9iAS OEM website. - OEM console (Oracle Enterprise Manager) You can manage 9iFS from the OEM console. - Oracle 9iFS Manager Graphical java based interface on iFS. - Webinterface iFS manager - Command line utilities ifsshell etc.. - Import/Export utility The Import/Export utility exports Oracle 9iFS objects (content and users) into an export file. Domain: ------- 9iFS is organized in a Domain concept, with an administrative Domain controller and possibly other nodes as members in the Domain. Repository: ----------- All data managed by 9iFS resides in an 9i database schema, called the 9iFS repository. You specify the database instance and schemaname during installation of 9iFS. Commands: --------- Stop IFS: Oracle Internet File System 1.1.x ORACLE_HOME\ifs1.1\bin\ifsstop.bat Oracle 9iFS 9.0.1 (and higher) ORACLE_HOME\9ifs\bin\ifsstopdomain.bat start iFS OC4J instance Windows NT or 2K: > ifsstartoc4j.bat start up ifs domain controller process Windows NT or 2K > ifslaunchdc.bat Start ifs node processes Windows NT or 2K > ifslaunchnode.bat Activate the iFS domain controller and Nodes Windows NT or 2K > ifsstartdomain.bat Here is a script example to run on windows NT or 2K: StartIfs902.bat =============== D:\ora902\9ifs\bin\ifsstartoc4j.bat start D:\ora902\9ifs\bin\ifslaunchdc.bat start D:\ora902\9ifs\bin\ifslaunchdomain.bat D:\ora902\9ifs\bin\ifsstartdomain -s myifshost:53140 ifssys echo "iFS 902 started" - Home: Oracle CM SDK must be installed in the Oracle9i Application Server, Release 2 home. Make sure to select the file location carefully; once installed, the Oracle CM SDK software cannot be moved without deinstalling and reinstalling. Oracle 9iFS requires an Oracle 9.0.2 home, which means you must install and configure Oracle9i Application Server, Release 2 in an Oracle home separate from that of the database. The Oracle home can be on the same machine (resources allowing), or on a different machine. - Install with Oracle Universal Installer. Installation and configuration of Oracle 9iFS starts from the Oracle Universal Installer, the graphical user interface wizard that copies all necessary software to the Oracle home on the target machine. The Oracle 9iFS Configuration tool launches automatically at the end of the Oracle Universal Installer process and guides you through the process of identifying the Oracle database to be used for the Oracle Internet File System schema; selecting the type of authentication to use (native Oracle 9iFS credential manager or Oracle Internet Directory for credential management); and various other configuration tasks. The specific configuration tasks vary, depending on the type of deployment (new Oracle 9iFS domain vs. additional Oracle 9iFS nodes, for example) - Starting install wizard again: ORACLE_HOME\ifs\cmsdk\bin\ifsca.bat - connect to database: The Oracle CM SDK Configuration Assistant attempts to make a connection as SYS AS SYSDBA using a database string, and therefore needs the database to be configured with a password file. - Directory service: Select either CMSDK Directory Service or Oracle Internet Directory Service for user authentication. The default Oracle Internet Directory super user name/password is cn=orcladmin/welcome1. The default Oracle Internet Directory root Oracle context is set to cn=OracleContext. - Launch Internet File System Manager from a Web browser: http://hostname.mycompany.com:7778/cmsdk/admin Access paths and directory structure: ------------------------------------- - Oracle FileSync Client Software: In addition to using the networking protocols or client applications native to the Windows operating system, Windows users can install and use Oracle FileSync to keep local directories on a desktop machine and folders in Oracle CM SDK synchronized. Double-click Setup.exe to run the installation program, or run O:\ifs\clients\filesync\setup.exe from the Windows Start...Run Menu. - CUP (Command-line Utilities Protocol) Client The Oracle Command-line Utilities Protocol server enables administrators and developers to perform a variety of tasks quickly and easily from a Windows command-line or a UNIX shell. copy /ifs/clients/cmdline/win32 to a local directory. ============================ 28. Some info on 9iAS rel. 2 ============================ 28.1 General Information: ========================= Oracle9i Application Server (Oracle9iAS) is a part of the Oracle9i platform, a complete and integrated e-business platform. Oracle9i platform consists of: - Oracle9i Developer Suite for developing applications - Oracle9i Application Server for deploying Internet applications - Oracle9i Database Server for storing content 9iAS is not just a webserver. A webserver is only part of the 9iAS system. 9iAS offers OC4J (Oracle Containers for J2EE), portals, webserver and webcache, and BusinessIntelligence and other components. OC4J: ----- The "core" of the AS (thus the application part), is the OC4J architecture. The OC4J infrastructure supports EJB, JSP and Servlet applications. Developers can write J2EE applications, like EJB, Servlet and JSP applications, that will run on 9iAS. OC4J itself is written in Java and runs on a Java virtual machine. BusinessIntelligence: --------------------- A set of services and client applications that make reports and all types of analysis possible. For example, the 'Oracle Reports service' , an application in the middle tier, uses a queue for submitted client requests. These request might create reports of a Datawarehouse in a Customer database etc... 28.1.1 Components: ------------------ There are 3 install types: -J2EE and Web Cache -Portal and Wireless -BusinessIntelligence and Forms Note: The Oracle 9iAS 9.0.2 Concepts and the 9iAS Install guides mentions 3 install types, but the Admin guide Rel. 9.0.2 mentions 4 install types. The fourth additional one is "Unified Messaging". This Enables you to integrate different types of messages into a single framework. It includes all of the components available in the Business Intelligence and Forms install type. Component J2EE and Web Cache Portal and Wireless BusinessInt. and Forms Oracle9iAS Web Cache YES YES YES Oracle HTTP Server YES YES YES Oracle9iAS Container for J2EE YES YES YES Oracle EM Web site YES YES YES Oracle9iAS Portal no YES YES Oracle9iAS Wireless no YES YES Oracle9iAS Discoverer no no YES Oracle9iAS Reports Services no no YES Oracle9iAS Clickstream Int. no no YES Oracle9iAS Forms Services no no YES Oracle9iAS Personalization no no YES 28.1.2. Need of Oracle9iAS Infrastructure: ------------------------------------------ Prior to installing an instance of the "Portal and Wireless" or "Business Intelligence and Forms" install type, you must install and configure the Oracle9iAS Infrastructure somewhere in your network, optimally on a separate computer. The J2EE and Web Cache install type does not require Oracle9iAS Infrastructure. You can install single or multiple instances of Oracle9iAS install types, J2EE and Web Cache, Portal and Wireless, and Business Intelligence and Forms, on the same host, which is not a very realistic scenarion. Multiple instances of different Oracle9iAS install types, can use one instance of Oracle9iAS Infrastructure, and this could be a realistic scenario. 28.1.3. Metadata Repository in the Infrastructure: -------------------------------------------------- The Oracle9iAS Infrastructure installation consists of: - Oracle9iAS Metadata Repository: Pre-seeded database containing metadata needed to run Oracle9iAS instances. - Oracle Internet Directory OID: Directory service that enables sharing information about dispersed users and network resources. Oracle Internet Directory implements LDAP v3. - Oracle9iAS Single Sign-On SSO: Creates an enterprise-wide user authentication to access multiple accounts and Oracle9iAS applications. - Oracle Management Server OMS: Processes system management tasks and administers the distribution of these tasks across the network using the Oracle Enterprise Manager Console. The Console and its three-tier architecture can be used with the Oracle Enterprise Manager Web site to manage not only Oracle9iAS, but your entire Oracle environment. - J2EE and Web Cache: For internal use with Oracle9iAS Infrastructure. Not used for component application deployment. Application server installations and their components use an infrastructure in the following ways: -- Components and applications use the Single Sign-on service provided by Oracle9iAS Single Sign-On. -- Application server installations and components store configuration information and user and group privileges in Oracle Internet Directory. -- Components use schemas that reside in the metadata repository. SSO is required for "Portal and Wireless" and "Business Intelligence and Forms" install types. Also required for application server clustering with J2EE and Web Cache install type. 28.1.4. Customer database: -------------------------- This could be any database on any Host, containing business data. But, The following components require a customer database: Oracle9iAS Discoverer Oracle9iAS Personalization Oracle9iAS Unified Messaging If you configure any of these components during installation, their setup and configuration will not be complete at the end of installation. You need to take additional steps to install and tune a customer database, load schemas into the database, and finish configuring the component to use the customer database. 28.1.5. Oracle Home: -------------------- Oracle home is the directory in which Oracle software is installed. Different Oracle versions always get their own Oracle Homes. Multiple instances of Oracle9iAS install types (J2EE and Web Cache, Business Intelligence and Forms, and Portal and Wireless) must be installed in separate Oracle homes on the same computer. You must install Oracle9iAS Infrastructure in its own Oracle home directory, preferably on a separate host. The Oracle9iAS installation cannot exist in the same Oracle home as the Oracle9iAS Infrastructure installation. 28.1.6. Oracle9iAS Infrastructure Port Usage: --------------------------------------------- !! Oracle9iAS Infrastructure requires exclusive use of port 1521 Installation of Oracle9iAS Infrastructure requires exclusive use of port 1521 on your computer. If one of your current system applications uses this port, then complete one of the following actions before installing Oracle9iAS Infrastructure: If you have an existing application using port 1521, then reconfigure the existing application to use another port. If you have an existing Oracle Net listener and an Oracle9i database, then proceed with the installation of Oracle9iAS Infrastructure. Your Oracle9iAS Infrastructure will use the existing Oracle Net listener. If you have an existing Net8 listener in use by an Oracle8i database, then you must upgrade to the Oracle9i Net listener version by installing Oracle9iAS Infrastructure. 28.1.6. Using the Oracle Enterprise Manager Console: ---------------------------------------------------- The Oracle Enterprise Manager console provides a wider view of your Oracle environment, beyond Oracle9iAS. Use the Console to automatically discover and manage databases, application servers, and Oracle applications across your entire network. The Console and its related components are installed with the Oracle Management Server as part of the Oracle9iAS Infrastructure installation option. The Console is part of the Oracle Management Server component of the Oracle9iAS Infrastructure. The Management Server, the Console, and Oracle Agent are installed on the Oracle9iAS Infrastructure host, along with the other infrastructure components. 28.1.7. Starting and Stopping the Oracle Management Server on Windows: ---------------------------------------------------------------------- On Windows systems, use the Services control panel to start and stop the management server. The name of the service is in the following format: OracleORACLE_HOMEManagementServer For example: OracleOraHome902ManagementServer 28.1.8. OEM Website: -------------------- You can verify the Enterprise Manager Web site is started by pointing your browser to the Web site URL. For example: http://hostname:1810 get console http://hostname:1810 http://127.0.0.1:1810 get welcome http://hostname:7777 https://192.168.1.106:5500/em/ To start or stop the Enterprise Manager Web site on Windows, use the Services control panel. The name of the service is in the following format: OracleORACLE_HOMEEMwebsite Or Start the Enterprise Manager Web site (UNIX) ORACLE_HOME/bin/emctl start (Windows) ORACLE_HOME\bin\emctl start Stop the Enterprise Manager Web site emctl stop Example Services: Oracleias902Discoverer Oracleias902ProcessManager Oracleias902WebCache Oracleias902WebCacheAdmin Oracleinfra902Agent = Agent for Management Server Oracleinfra902EMWebsite = Enterprise Manager Web site Oracleinfra902InternetDirectory_iasdb Oracleinfra902ManagementServer = OEM Management Server Oracleinfra902ProcessManager OracleOraHome901TNSListener = just the Listener OracleServiceIASDB = infra structure db OracleServiceO901 = regular customer db Note for Oracle 10g RDBMS EM DB console: ======================================== Sites: ------ Enterprise Manager Database Control URL - (dbname) :http://hostname:1158/em http://127.0.0.1:1810 http://127.0.0.1:1158 The iSQL*Plus URL is:http://localhost:5561/isqlplus The iSQL*Plus DBA URL is:http://localhost:5561/isqlplus/dba emctl prompt tool: ------------------ C:\ora10g\product\10.2.0\db_1\NETWORK\ADMIN>emctl status dbconsole Oracle Enterprise Manager 10g Database Control Release 10.2.0.1.0 Copyright (c) 1996, 2005 Oracle Corporation. All rights reserved. http://xpwsora:1158/em/console/aboutApplication Oracle Enterprise Manager 10g is running. Logs are generated in directory C:\ora10g\product\10.2.0\db_1/xpwsora_SPLCONF/sysman/logServices: --------- C:\ora10g\product\10.2.0\db_1\NETWORK\ADMIN>net start | find "Ora" OracleDBConsolesplconf OracleOraDb10g_home1iSQL*Plus OracleOraDb10g_home1TNSListener OracleServiceSPLCONF C:\ora10g\product\10.2.0\db_1\NETWORK\ADMIN> 28.1.9. emctl tool : for controlling EM website: ------------------------------------------------- Enterprise manager homepage http://hostname:1810 can only be accessed if EM webste is running. Usage:: emctl start|stop|status emctl reload | upload emctl set credentials [[:]] emctl gencertrequest emctl installcert [-ca|-cert] emctl set ssl test|on|off|password [ ] emctl set password emctl authenticate emctl switch home [-silent ] emctl config emctl start : Start the Enterprise Manager Web site. emctl stop : Stop the Enterprise Manager Web site (requires ias_admin password). emctl status : Verify the status of the Enterprise Manager Web site. emctl set password new_password : Reset the ias_admin password. emctl authenticate password : Verify that the supplied password is the ias_admin password. emctl config options can be listed by typing "emctl config" emctl status C:\temp>emctl status EMD is up and running : 200 OK Note for 11g: ------------- Starting and Stopping the Database Control Console Process To access Oracle Enterprise Manager Database Control (Database Control) from a client browser, the Database Control console process dbconsole needs to be running on the database host computer. The dbconsole process is automatically started after installation. However, in the event that the system needs to be restarted, you can start it manually at the command line or start it as a service in Windows. You can also run commands to stop the dbconsole process or view its status. If the dbconsole process is stopped, it must be manually restarted before you use the Database Control console. To start or stop, or to view the status of the dbconsole process from the command line: Open a operating system command window and go to your Oracle_home/bin directory. Do one of the following: To start the dbconsole process, run the following command: ./emctl start dbconsole To stop the dbconsole process, run the following command: ./emctl stop dbconsole To view the status of the dbconsole process, run the following command: ./emctl status dbconsole 28.1.10. OEMCTL tool: for controlling Management Server: -------------------------------------------------------- EM control D:\temp>oemctl "Syntax: OEMCTL START OMS " " OEMCTL STOP OMS /" " OEMCTL STATUS OMS /[@]" " OEMCTL PING OMS " " OEMCTL START PAGING [BootHost Name] " " OEMCTL STOP PAGING [BootHost Name] " " OEMCTL ENABLE EVENTHANDLER" " OEMCTL DISABLE EVENTHANDLER" " OEMCTL EXPORT EVENTHANDLER " " OEMCTL IMPORT EVENTHANDLER " " OEMCTL DUMP EVENTHANDLER" " OEMCTL IMPORT REGISTRY /@" " OEMCTL EXPORT REGISTRY /@" " OEMCTL CONFIGURE RWS" 28.1.11. The Intelligent Agent: ------------------------------- The Oracle Intelligent Agent is installed whenever you install Oracle9iAS on a host computer. For example, if you select the J2EE and Web Cache installation type, the Oracle Universal Installer installs Oracle Enterprise Manager Web site and the Oracle Intelligent Agent, along with the J2EE and Web Cache software. This means the Intelligent Agent software is always available if you decide to use the Console and the Management Server to manage your Oracle9iAS environment. The Console and Management Server are installed as part of the Oracle9iAS Infrastructure. In most cases, you install the Infrastructure on a dedicated host that can be used to centrally manage multiple application server instances. The Infrastructure includes Oracle Internet Directory, Single Sign-On, the metadata repository, the Intelligent Agent, and Oracle Management Server. You only need to run the Intelligent Agent if you are using Oracle Management Server in your enterprise. In order for Oracle Management Server to detect application server installations on a host, you must make sure the Intelligent Agent is started. Note that one Intelligent Agent is started per host and must be started after every system boot. 28.1.12. AGENTCTL: for controlling the Intelligent Agent: --------------------------------------------------------- (UNIX) You can run the following commands in the Oracle home of the primary installation (the first installation on the host) to get status and start the Intelligent Agent: ORACLE_HOME/bin/agentctl status agent ORACLE_HOME/bin/agentctl start agent (Windows) You can check the status and start the Intelligent Agent using the Services control panel. The name of the service is in the following format: OracleORACLE_HOMEAgent (the executable is agntsrvc.exe) start the Intelligent Agent in the Oracle home of the primary installation: ORACLE_HOME/bin/agentctl start agent 28.1.13. Backup and Restore: ---------------------------- To ensure that you can make a full recovery from media failures, you should perform regular backups of the following: - Application Server and Infrastructure Oracle Homes - Oracle Internet Directory - Metadata Repository - Customer Databases You should perform regular backups of all files in the Oracle home of each application server and infrastructure installation in your enterprise using your preferred method of filesystem backup. Oracle Internet Directory offers command-line tools for backing up and restoring the Oracle Internet Directory schema and subtree. The metadata repository is an Oracle9i Enterprise Edition Database that you can back up and restore using several different tools and operating system commands. The customer databases can be backupped using any standard method, the same way you would do for any other 9iEE database. Applications: ============= 28.2 Report services: --------------------- Client contacts the Report Server - Web,through url - Nonweb, rwclient -requests goes to a jobqueue -users with webbrowser: http Server must be running, and you use or reports servlet, a JSP, or CGI components on 9iAS The reports server must be running. - default it is an inprocess server httpd -> mod_oc4j {reports servlet} -> Reports Server - CGI httpd -> CGI -> Reports Server - starting from URL: http://machine:port/reports/rwservlet commandline: rwserver server=machinename - The servlet is part of the OC4J instance: OC4J_BI_FORMS - its possible to make it a service of its own: rwserver -install autostart=yes/no - verify the Reports Servlet and Server Are Running: http://missrv/rwservlet/help (show help page with rwservlet command line arguments) http://machine:port/reports/rwservlet/showjobs?server=server_name (show a listing of the jobqueue) IP:7778/reports/rwservlet/showenv http://:/reports/rwservlet/getserverinfo? http://:/reports/rwservlet/getserverinfo?authid=orcladmin/ http://machinename/servlet/RWServlet/showmap?server=Rep60_servername - stopping Reports Server: commandline: rwserver server=machinename shutdown=normal/immediate authid=admin/password Enterprise Manager: stop Reports Server The reports servlet uses the PORT parameter configured in the httpd.conf reports_user/welcome1 ias_admin/welcome1 orcladmin /welcome1 Reports Servlet url : http://missrv:7778/reports/rwservlet em username : reports_user em password : welcome1 reports store : d:\reports (change in registry, key is REPORTS_PATH) - Reports Server configuration files: ORACLE_HOME\reports\conf\server_name.conf ORACLE_HOME\reports\dtd\rwserverconf.dtd ORACLE_HOME\reports\conf\rwbuilder.conf ORACLE_HOME\reports\conf\rwservlet.properties - Check miskm.propery files: $9ias_home\j2ee\OC4J_iFS_cmsdk\applications\brugpaneel\FrontOffice\WEB-INF\classes. Het gaat om de volgende bestanden: misIfs.properties : parameters van iFs interface/front office. miskm.properties : parameters van MIS Front Office applications XSQLConfig.xml : XSQL Parameters, moet wijzen naar mis_owner schema. Er wordt ook gebruik gemaakt van JDBC. De instellingen van deze connectie staan in het bestand: $9ias_home\j2ee\OC4J_iFS_cmsdk\applications\brugpaneel\META-INF\data-sources.xml miskm.properties: ----------------- # miskm.reports parameters are used in order to display reports that are built # using Oracle Reports. # The action of the hidden form. #miskm.reports.action=http://dgas40.mindef.nl/reports/rwservlet miskm.reports.action=http://missrv.miskm.mindef.nl:7778/reports/rwservlet # The schemaname/schemapassword@tns_names entry where the data is stored. #miskm.reports.connectstring=mis_owner/mis_owner@miskm_demo miskm.reports.connectstring=mis_owner/mis_owner@miskm_dev # The name of the Reports Server (after default installation: rep_missrv) #miskm.reports.repserver=rep_dgas40 miskm.reports.repserver=rep_missrv # The location where the output is placed on the server. miskm.reports.destype=cache # The output of the the generated report (e.g html, pdf, etc.) #miskm.reports.desformat=pdf miskm.reports.desformat=rtf&mimetype=application/msword # The reports server is a partner application, therefore a sso username/password # is required. miskm.reports.ssoauthid=reports_user/welcome1 - Reports Server configuration files: ORACLE_HOME\reports\conf\server_name.conf ORACLE_HOME\reports\dtd\rwserverconf.dtd ORACLE_HOME\reports\conf\rwbuilder.conf ORACLE_HOME\reports\conf\rwservlet.properties (inprocess or standallone) reports_server_name.conf cgicmd.dat jdbcpds.conf proxyinfo.xml rwbuilder.conf rwserver.template rwservlet.properties textpds.conf xmlpds.conf in ORACLE_HOME/reports/conf Reports Servlet 9i Rapportages worden gemaakt met behulp van de Reports Builder en moeten worden opgeslagen in een directory op de applicatieserver (standaard is dit d:\reports). Om de Reports Servlet te laten weten waar allemaal reports zijn opgeslagen dient de registersleutel van Windows REPORTS_PATH te worden uitgebreid met de directory waar de rapportages zijn opgeslagen. De servlet is onderdeel van de OC4J instance: OC4J_BI_FORMS, dus om hier gebruik van te maken, moet deze instance opgestart zijn. De servlet maakt gebruik van Oracle SSO en daarom dient een er een SSO gebruiker aangemaakt te worden die in staat is om gebruik te maken van de servlet: 1. Ga naar http://missrv.miskm.mindef.nl:7777/oiddas 2. Log in als de portal gebruiker (standaard portal/welcome1) 3. Maak een nieuwe gebruiker aan, bijvoorbeeld: reports_user. 4. Sta deze gebruiker de privilege: ‘Allow resource management for Oracle Reports and Forms’ toe. 5. Controleer of deze gebruiker overeenstemt met de sleutel: miskm.reports.ssoauthid in het bestand miskm.properties 28.3 Internet Directory and Single Sign-On: ------------------------------------------- Oracle Internet Directory, an LDAP directory, provides a single repository and administration for user accounts. Oracle9iAS Single Sign-On enables users to login to Oracle9iAS and gain access to those applications for which they are authorized, without requiring them to re-enter a user name and password for each application. It is fully integrated with Oracle Internet Directory, which stores user information. It supports LDAP-based user and password management through OID. Oracle Internet Directory is installed as part of the Oracle9iAS Infrastructure installation. Oracle9iAS Single Sign-On is installed as part of the Oracle9iAS Infrastructure installation. SSO is Portal's authentication engine. In 9iAS all applications may use SSO. Without a functioning SSO, users will not be able to logon and use SSO. The first test following a failure to authenticate is to login directly using SSO: http://servername:port/pls/orasso Examples: Single Sign-On Server : oasdocs.us.oracle.com:7777 Internet Directory : oasdocs.us.oracle.com:389 Infrastructure database : iasdb.oasdocs.us.oracle.com missrv.miskm.mindef.nl:1521:iasdb In a start script, you may find commands like the following to start the OID server: %INFRA_BIN%\oidmon start %INFRA_BIN%\oidctl server=oidldapd instance=1 start In a stop script, you may notice the following commands to stop the OID server: %INFRA_BIN%\oidctl server=oidldapd instance=1 stop %INFRA_BIN%\oidmon stop When oidctl is executed, it connects to the database as user ODSCOMMON and simply inserts/updates rows into a table ODS.ODS_PROCESS depending on the options used in the command. A row is inserted if the START option is used, and updated if the STOP or RESTART option is used. So there are no processes started at this point, and LDAP server is not started. Both the listener/dispatcher process and server process are called oidldapd on unix, and oidldapd.exe on NT. Oidmon is also a process (called oidmon on unix, oidmon.exe/oidservice.exe on windows). To control the processes (servers) we need to have OID Monitor (oidmon) running. This monitor is often called daemon or guardian process as well. When oidmon is running, it periodically connects to the database and reads the ODS.ODS_PROCESS table in order to start/stop/restart related processes. NOTE: Because the only task oidctl has is to insert / update table ODS.ODS_PROCESS in the database, it's obvious that the database and listener have to be fully accessible when oidctl is used. Also, oidmon connects periodically to the database. So the database and listener must be accessible for oidmon to connect. 28.4 Example and default values: -------------------------------- Information Example Values Your Information Oracle home location D:\ora9ias Instance Name instance1 ias_admin Password welcome1 Single Sign-On Server HostName/server oasdocs.us.oracle.com Single Sign-On Port Number 7777 Internet Directory Hostname/server oasdocs.us.oracle.com Internet Directory Port Number 389 / 4032 Internet Directory Username orcladmin, cn=orcladmin (the Oracle Internet Directory administrator) Internet Directory Password welcome1 9iAS Metadata Repository oasdocs.us.oracle.com 9iAS Reports Services Outgoing Mail Server oasdocs.us.oracle.com http Server oasdocs.us.oracle.com:7777 Metadata database connection string oasdocs.us.oracle.com:1521:iasdb:iasdb.oasdocs.us.oracle.com Oracle Universal Installer creates a file showing the port assignments during installation of Oracle9iAS components. This file is ORACLE_HOME\install\portlist.ini It contains entries like the following default values: Oracle HTTP Server port = 7777 Oracle HTTP Server SSL port = 4443 Oracle HTTP Server listen port = 7778 Oracle HTTP Server SSL listen port = 4444 Oracle HTTP Server Jserv port = 8007 Enterprise Manager Servlet port = 1810 The ID username and password are defined in Oracle Internet Directory as either the: - orcladmin (root user) - a user who is member of the IASAdmins group in Oracle Internet Directory The SSO schema is now 'ORASSO' and the ORASSO user is registered with OID after an infra install. THe default user is 'orcladmin' with a login of your ias_admin password. EM Website: http://: (port 1810 assigned by default) You will login using the 'ias_admin' username and the password you entered during the Infrastructure installation. SSO Login Page: http://:/pls/orasso You will login using the 'orcladmin' username and the password for the 'ias_admin'. The port will be the HTTP Server port of your Infrastructure, (port 7777 by default) http://missrv.miskm.mindef.nl:7777/pls/orasso OID_DAS Page: http://:/oiddas You will login using the 'orcladmin' username and the password for the 'ias_admin'. The port will be the HTTP Server port of your Infrastructure, (port 7777 by default). The OC4J_DAS component must be UP for this test to succeed. 28.5 Management tools: ---------------------- 28.5.1. OEM Website: ------------------- You can access the Welcome Page by pointing your browser to the HTTP Server URL for your installation. For example, the default HTTP Server URL is: http://hostname:7777 This page offer many options to explore features of 9iAS. You can also go directly to the Oracle Enterprise Manager Web site using the following instructions: http://hostname:1810 http:// Enterprise manager homepage http://hostname:1810 can only be accessed if EM webste is running. This corresponds to the service like "Oracleinfra902EMWebsite". The username for the administrator user is ias_admin. The password is defined during the installation of Oracle9iAS. The default password is welcome1. Depending upon the options you have installed, the Administration section of the Oracle9iAS Instance Home Page provides additional features that allow you to perform the following tasks: -Associate the current instance with an existing Oracle9iAS Infrastructure. -Configure additional Oracle9iAS components that have been installed, but not configured -Change the password or default schema for a component Start or stop on NT/W2K: To start or stop the Enterprise Manager Web site on Windows, use the Services control panel. The name of the service is in the following format: OracleORACLE_HOMEEMwebsite For example, if the name of the Oracle Home is OraHome902, the service name is: OracleOraHome902EMWebsite You can also use net start OracleOraHome902EMWebsite net stop OracleOraHome902EMWebsite Start or stop on UNIX: Start the Enterprise Manager Web site: emctl start Stop the Enterprise Manager Web site: emctl stop Or use the kill command if it does not respond Changing the ias_admin Password: 1. Using Oracle Enterprise Manager Web Site: Navigate to the Instance Home Page. Select Preferences in the top right corner. This displays the Change Password Page. Enter the new password and new password confirmation. Click OK. This resets the ias_admin password for all application server installations on the host. Restart the Oracle Enterprise Manager Web site. 2. Using the emctl Command-Line Tool: To change the ias_admin user password using a command-line tool: Enter the following command in the Oracle home of the primary installation (the first installation on the host): (UNIX) ORACLE_HOME/bin/emctl set password new_password (Windows) ORACLE_HOME\bin\emctl set password new_password For example: (UNIX) ORACLE_HOME/bin/emctl set password m5b8r5 (Windows) ORACLE_HOME\bin\emctl set password m5b8r5 Restart the Enterprise Manager Web site. The Enterprise Manager Web site relies on various technologies to discover, monitor, and administer the Oracle9iAS environment. These technologies include: - Oracle Dynamic Monitoring Service (DMS) The Enterprise Manager Web site uses DMS to gather performance data about your Oracle9iAS components. - Oracle HTTP Server and Oracle Containers for J2EE (OC4J) the Enterprise Manager Web site also uses HTTP Server and OC4J to deploy its management components. - Oracle Process Management Notification (OPMN) OPMN manages Oracle HTTP Server and OC4J processes within an application server instance. It channels all events from different component instances to all components interested in receiving them. - Distributed Configuration Management (DCM) This will be used with clusters or farms. DCM manages configurations among application server instances that are associated with a common Infrastructure (members of an Oracle9iAS farm). It enables Oracle9iAS cluster-wide deployment so you can deploy an application to an entire cluster, or make a single host or instance configuration change applicable across all instances in a cluster. 28.5.2 OEM Console: ------------------- The console is a non Web, Java tool, and part of the 3-tier OMS architecture. See also section 28.1. The Oracle Enterprise Manager console provides a wider view of your Oracle environment, beyond Oracle9iAS. Use the Console to automatically discover and manage databases, application servers, and Oracle applications across your entire network. The Console and its related components are installed with the Oracle Management Server as part of the Oracle9iAS Infrastructure installation option. The Console is part of the Oracle Management Server component of the Oracle9iAS Infrastructure. The Management Server, the Console, and Oracle Agent are installed on the Oracle9iAS Infrastructure host, along with the other infrastructure components. The Console offers advanced management features, such as an Event system to notify administrators of changes in your environment and a Job system to automate standard and repetitive tasks, such as executing a SQL script or executing an operating system command. The Console and Management Server are installed as part of the Oracle9iAS Infrastructure. Use the OEMCTL commandline tool for controlling OMS. See section 28.1.10. 29. Starting and stopping 9iAS and components: ============================================== 29.1 Starting a simple Webcache/J2EE installation: -------------------------------------------------- Start the Enterprise Manager Web site. Even though you are not using the Web site, this ensures that the processes to support the dcmctl command-line tool are started. To start the Web site, execute the following command in the Oracle home of the primary installation on your host: (UNIX) ORACLE_HOME/bin/emctl start (Windows) ORACLE_HOME\bin\emctl start Start Oracle HTTP Server and OC4J (the rest of the commands in this section should be executed in the Oracle home of the J2EE and Web Cache instance): (UNIX) ORACLE_HOME/dcm/bin/dcmctl start (Windows) ORACLE_HOME\dcm\bin\dcmctl start If Web Cache is configured, start Web Cache: (UNIX) ORACLE_HOME/bin/webcachectl start (Windows) ORACLE_HOME\bin\webcachectl start 29.2 Startin and stopping Advanced 9iAS installations ----------------------------------------------------- Start/Stop Enterprise: ---------------------- Starting an Application Server Enterprise: The order in which to start the pieces of an application server enterprise is as follows: 1. Start the infrastructure. If your enterprise contains more than one infrastructure, start the primary infrastructure first. 2. Start customer databases. If your enterprise contains customer databases, you can start them using several methods, including SQL*Plus and Oracle Enterprise Manager Console. Remember that iFS could also be installed into the customer database. 3. Start application server instances. You can start application server instances in any order. If instances are part of a cluster, start them as part of starting the cluster. The order in which to stop the pieces of an application server enterprise is as follows: 1. Stop application server instances. You can stop application server instances in any order. If instances are part of a cluster, stop them as part of stopping the cluster. 2. Stop customer databases. If your enterprise contains customer databases, you can stop them using several methods, including SQL*Plus and Oracle Enterprise Manager Console. 3. Stop the infrastructure. If your enterprise contains more than one infrastructure, stop the primary infrastructure last. Start/Stop Instance: -------------------- Start: First you have started the infrastructure instance, and customer database instance. 1. Preliminary: - Enterprise Manager Web Site (Required): The first step before starting an application server instance is to ensure that the Enterprise Manager Web site is running on the host. The Web site provides underlying processes required to run an application server instance and must be running even if you intend to use command-line tools to start your instance. There is one Enterprise Manager Web site per host. It resides in the primary installation (or first installation) on that host. The primary installation can be an application server installation or an infrastructure. This Web site usually listens on port 1810 and provides services to all application server instances and infrastructures on that host. To verify the status of the Enterprise Manager Web site, run the following command in the Oracle home of the primary installation: (UNIX) ORACLE_HOME/bin/emctl status (Windows) ORACLE_HOME\bin\emctl status To start the Enterprise Manager Web site, run the following command in the Oracle home of the primary installation: (UNIX) ORACLE_HOME/bin/emctl start (Windows) ORACLE_HOME\bin\emctl start Or on NT/W2K: net start OracleORACLE_HOMEEMwebsite - Intelligent Agent (Optional) You only need to run the Intelligent Agent if you are using Oracle Management Server in your enterprise. In order for Oracle Management Server to detect application server installations on a host, you must make sure the Intelligent Agent is started. Note that one Intelligent Agent is started per host and must be started after every system boot. (UNIX) You can run the following commands in the Oracle home of the primary installation (the first installation on the host) to get status and start the Intelligent Agent: ORACLE_HOME/bin/agentctl status agent ORACLE_HOME/bin/agentctl start agent (Windows) You can check the status and start the Intelligent Agent using the Services control panel. The name of the service is in the following format: OracleORACLE_HOMEAgent 2. Start the instance using OEM Website: You can start, stop, and restart all types of application server instances using the Instance Home Page on the Enterprise Manager Web site. Or... 3. Start the 'J2EE and Web Cache' instance using commands: Start OEM Website: ORACLE_HOME\bin\emctl start or net start OracleORACLE_HOMEEMwebsite Start Oracle HTTP Server and OC4J: ORACLE_HOME\dcm\bin\dcmctl start If Web Cache is configured, start Web Cache: ORACLE_HOME\bin\webcachectl start 4. Stop the 'J2EE and Web Cache' instance using commands: ORACLE_HOME\bin\webcachectl stop ORACLE_HOME\dcm\bin\dcmctl stop Start/Stop components: ---------------------- You can start, stop, and restart individual components using the Instance Home Page or the component home page on the Enterprise Manager Web site. You can also start and stop some components using command-line tools. Oracle HTTP Server Start: ORACLE_HOME\dcm\bin\dcmctl start -ct ohs Stop : ORACLE_HOME\dcm\bin\dcmctl stop -ct ohs Individual OC4J Instances Start: ORACLE_HOME\dcm\bin\dcmctl start -co instance_name Stop : ORACLE_HOME\dcm\bin\dcmctl stop -co instance_name All OC4J Instances Start: ORACLE_HOME\dcm\bin\dcmctl start -ct oc4j Stop : ORACLE_HOME\dcm\bin\dcmctl stop -ct oc4j Web Cache Start: ORACLE_HOME\bin\webcachectl start Stop : ORACLE_HOME\bin\webcachectl stop Reports Start: ORACLE_HOME\bin\rwserver server=name Stop : ORACLE_HOME\bin\rwserver server=name shutdown=yes You cannot start or stop some components. The radio buttons in the Select column on the Instance Home Page are disabled for these components, and their component home pages do not have Start, Stop, or Restart buttons. Start/Stop the Infrastructure: ------------------------------ No matter which procedure you use, starting an infrastructure involves performing the following steps in order: Start the Metadata Repository = infrastructure database Start OID, Oracle Internet Directory Start the Enterprise Manager Web site. Start OHS, Oracle HTTP Server Start the OC4J_DAS instance Start Web Cache (optional) Start Oracle Management Server and Intelligent Agent (optional) No matter which procedure you use, stopping an infrastructure involves performing the following steps in order: Stop all middle-tier application server instances that use the infrastructure. Stop Oracle Management Server and Intelligent Agent (optional) Stop Web Cache (optional) Stop OC4J instances Stop Oracle HTTP Server Stop Oracle Internet Directory Stop the Metadata Repository The next section describes how to start an infrastructure using command-line tools on Windows. Except where noted, all commands should be run in the Oracle home of the infrastructure. -- --------------------------------------------------------------------- -Start the metadata repository listener: ORACLE_HOME\bin\lsnrctl start -Set the ORACLE_SID environment variable to the metadata repository system identifier (default is iasdb). You can set the ORACLE_SID system variable using the System Properties control panel. -Start the metadata repository instance using SQL*Plus: ORACLE_HOME\bin\sqlplus /nolog sql> connect sys/password_for_sys as sysdba sql> startup sql> quit -- --------------------------------------------------------------------- - Start Oracle Internet Directory. Make sure the ORACLE_SID is set to the metadata repository system identifier (refer to previous step). Start the Oracle Internet Directory monitor: ORACLE_HOME\bin\oidmon start -Start the Oracle Internet Directory server: ORACLE_HOME\bin\oidctl server=oidldapd configset=0 instance=n start where n is any instance number (1, 2, 3...) that is not in use. For example: ORACLE_HOME\bin\oidctl server=oidldapd configset=0 instance=1 start -- --------------------------------------------------------------------- - Start the Enterprise Manager Web site. Even though you are using command-line, the Web site is required because it provides underlying support for the command-line tools. The Web site must be started after every system boot. You can check the status and start the Enterprise Manager Web site using the Services control panel. The name of the service is in the following format: OracleORACLE_HOMEEMwebsite You can also start the service using the following command line: net start WEB_SITE_SERVICE_NAME -- --------------------------------------------------------------------- -Start Oracle HTTP Server. ORACLE_HOME\dcm\bin\dcmctl start -ct ohs Note that starting Oracle HTTP Server also makes Oracle9iAS Single Sign-On available. -- --------------------------------------------------------------------- - Start the OC4J_DAS instance. ORACLE_HOME\dcm\bin\dcmctl start -co OC4J_DAS Note that the infrastructure instance contains other OC4J instances, such as OC4J_home and OC4J_Demos, but these do not need to be started; their services are not required and incur unnecessary overhead. -- --------------------------------------------------------------------- -Start Web Cache (optional). Web Cache is not configured in the infrastructure by default, but if you have configured it, start it as follows: ORACLE_HOME\bin\webcachectl start -- --------------------------------------------------------------------- - Start Oracle Management Server and Intelligent Agent (optional). Perform these steps only if you have configured Oracle Management Server. Start Oracle Management Server: ORACLE_HOME\bin\oemctl start oms -- --------------------------------------------------------------------- Start the Intelligent Agent. In order for Oracle Management Server to detect the infrastructure and any other application server installations on this host, you must make sure the Intelligent Agent is started. Note that one Intelligent Agent is started per host and must be started after every reboot. You can check the status and start the Intelligent Agent using the Services control panel. The name of the service is in the following format: OracleORACLE_HOMEAgent 30. Creating a Database Access Descriptor (DAD) for mod_plsql: --------------------------------------------------------------- Oracle HTTP Server contains the mod_plsql module, which provide support for building PL/SQL-based applications on the Web. PL/SQL stored procedures retrieve data from a database and generate HTTP responses containing data and code to display in a Web browser. In order to use mod_plsql you must install the PL/SQL Web Toolkit into a database and create a Database Access Descriptor (DAD) which provides mod_plsql with connection information for the database. 31. Configuring HTTP Server, OC4J, and Web Cache: -------------------------------------------------- You can use the OEM website in order to configure components as HTTP Server, OC4J, and Web Cache, or you can manually edit configuration files. If you edit Oracle HTTP Server or OC4J configuration files manually, instead of using the Enterprise Manager Web site, you must use the DCM command-line utility dcmctl to notify the DCM repository of the changes. Otherwise, your changes will not go into effect and will not be reflected in the Enterprise Manager Web site. Note that the dcmctl tool is located in: UNIX) ORACLE_HOME/dcm/bin/dcmctl (Windows) ORACLE_HOME\dcm/bin\dcmctl To notify DCM of changes made to: Use this command: Oracle HTTP Server configuration files: dcmctl updateConfig -ct ohs OC4J configuration files : dcmctl updateConfig -ct oc4j All configuration files : dcmctl updateConfig - HTTP Server: You can configure Oracle HTTP Server using the Oracle HTTP Server Home Page on the Oracle Enterprise Manager Web site. You can perform tasks such as modifying directives, changing log properties, specifying a port for a listener, modifying the document root directory, managing client requests, and editing server configuration files. You can access the Oracle HTTP Server Home Page in the Name column of the System Components table on the Instance Home Page. - OC4J: You can configure Oracle9iAS Containers for J2EE (OC4J) using the Enterprise Manager Web site. You can use the Instance Home Page to create and delete OC4J instances, each of which has its own OC4J Home Page. You can use each individual OC4J Home Page to configure the corresponding OC4J instance and its deployed applications. Creating an OC4J Instance. Every application server instance has a default OC4J instance named OC4J_home. You can create additional instances, each with a unique name, within an application server instance. To create a new OC4J instance: - Navigate to the Instance Home Page on the Oracle Enterprise Manager Web site. Scroll to the System Components section. - Click Create OC4J Instance. This opens the Create OC4J Instance Page. - In the Create OC4J Instance Page, type a unique instance name in the OC4J instance name field. Click Create. - A new OC4J instance is created with the name you provided. - This OC4J instance shows up on the Instance Home Page in the System Components section. - The instance is initially in a stopped state and can be started any time after creation. Each OC4J instance has its own OC4J Home Page which allows you to configure global services and deploy applications to that instance. 32. 9iAS CONFIG FILES: ----------------------- --------------------------------------------- 32.1 9iAS Rel. 2 most obvious config files: --------------------------------------------- Oracle HTTP Server: ------------------- httpd.conf oracle_apache.conf access.conf magic mime.types mod_oc4j.conf srm.conf in ORACLE_HOME/Apache/Apache/conf JServ: ------ jserv.conf jserv.properties zone.properties in ORACLE_HOME/Apache/Jserv/etc mod_oradav: ----------- moddav.conf in ORACLE_HOME/Apache/oradav/conf mod_plsql: ---------- cache.conf dads.conf in ORACLE_HOME/Apache/modplsql/conf Oracle9iAS Web Cache: --------------------- internal.xml internal_admin.xml webcache.xml in ORACLE_HOME/webcache Oracle9iAS Reports Services: ---------------------------- reports_server_name.conf cgicmd.dat jdbcpds.conf proxyinfo.xml rwbuilder.conf rwserver.template rwservlet.properties textpds.conf xmlpds.conf in ORACLE_HOME/reports/conf Oracle9iAS Discoverer: ---------------------- configuration.xml in ORACLE_HOME/j2ee/OC4J_BI_Forms/applications/discoverer/web/WEB-INF/lib viewer_config.xml in ORACLE_HOME/j2ee/OC4J_BI_Forms/applications/discoverer/web/viewer_files plus_config.xml in ORACLE_HOME/j2ee/OC4J_BI_Forms/applications/discoverer/web/plus_files portal_config.xml in ORACLE_HOME/j2ee/OC4J_BI_Forms/applications/discoverer/web/portal pref.txt in ORACLE_HOME/discoverer902/util .reg_key.dc in ORACLE_HOME/discoverer902/bin/.reg --------------------------------------------- 32.2 9iAS Rel. 2 list of all .conf files: --------------------------------------------- Now as an example, follows a listing of all .conf configuration files of a real 9iAS Server. -- ------------------------------------------------------------------- -- BEGIN LISTING FROM AN REAL LIFE 9iAS rel. 9.0.2 Server: -- ------------------------------------------------------------------- Directory of D:\ORACLE\ias902\Apache\Apache\conf 06/25/2002 10:55p 293 access.conf 12/01/2003 02:07p 46,178 httpd.conf 12/01/2003 02:07p 3,342 mod_oc4j.conf 12/01/2003 02:07p 517 mod_osso.conf 12/01/2003 02:07p 811 oracle_apache.conf 06/25/2002 10:55p 305 srm.conf 12/01/2003 02:07p 551 wireless_sso.conf 7 File(s) 51,997 bytes Directory of D:\ORACLE\ias902\Apache\Apache\conf\osso 04/23/2003 08:41p 433 osso.conf 1 File(s) 433 bytes Directory of D:\ORACLE\ias902\Apache\Jserv\conf 04/23/2003 08:38p 10,745 jserv.conf 1 File(s) 10,745 bytes Directory of D:\ORACLE\ias902\Apache\jsp\conf 12/01/2003 02:07p 594 ojsp.conf 1 File(s) 594 bytes Directory of D:\ORACLE\ias902\Apache\modplsql\conf 12/01/2003 02:07p 840 cache.conf 12/01/2003 02:07p 2,122 dads.conf 12/01/2003 02:07p 1,598 plsql.conf 3 File(s) 4,560 bytes Directory of D:\ORACLE\ias902\Apache\oradav\conf 12/01/2003 02:07p 785 moddav.conf 12/01/2003 02:07p 396 oradav.conf 2 File(s) 1,181 bytes Directory of D:\ORACLE\ias902\click\conf 12/01/2003 02:07p 427 click-apache.conf 1 File(s) 427 bytes Directory of D:\ORACLE\ias902\click\conf\templates 01/14/2002 11:21p 445 click-apache.conf 1 File(s) 445 bytes Directory of D:\ORACLE\ias902\dcm\config 02/17/2004 01:31p 186 dcm.conf 1 File(s) 186 bytes Directory of D:\ORACLE\ias902\dcm\config\plugins\apache 06/27/2002 11:01p 43,623 httpd.conf 1 File(s) 43,623 bytes Directory of D:\ORACLE\ias902\dcm\repository.install\dcm\config 04/23/2003 08:57p 185 dcm.conf 1 File(s) 185 bytes Directory of D:\ORACLE\ias902\forms90\server 12/01/2003 02:07p 2,997 forms90.conf 1 File(s) 2,997 bytes Directory of D:\ORACLE\ias902\ldap\das 12/01/2003 02:07p 165 oiddas.conf 1 File(s) 165 bytes Directory of D:\ORACLE\ias902\opmn\conf 02/17/2004 01:31p 45 ons.conf 1 File(s) 45 bytes Directory of D:\ORACLE\ias902\portal\conf 12/01/2003 02:07p 1,407 portal.conf 1 File(s) 1,407 bytes Directory of D:\ORACLE\ias902\RDBMS\demo 12/01/2003 02:07p 482 aqxml.conf 1 File(s) 482 bytes Directory of D:\ORACLE\ias902\reports\conf 04/28/2003 02:59p 3,386 Copy (2) of rep_vbas99.conf 05/17/2002 08:45p 7,421 jdbcpds.conf 04/28/2003 02:59p 3,386 rep_vbas99.conf 05/17/2002 08:45p 6,381 textpds.conf 05/17/2002 08:45p 454 xmlpds.conf 5 File(s) 21,028 bytes Directory of D:\ORACLE\ias902\ultrasearch\webapp\config 12/01/2003 02:07p 320 ultrasearch.conf 1 File(s) 320 bytes Directory of D:\ORACLE\ias902\xdk\admin 12/01/2003 02:07p 294 xml.conf 1 File(s) 294 bytes Directory of D:\ORACLE\infra902\Apache\Apache\conf 06/25/2002 10:55p 293 access.conf 04/23/2003 08:23p 46,224 httpd.conf 04/23/2003 08:23p 1,500 mod_oc4j.conf 04/23/2003 08:23p 519 mod_osso.conf 04/23/2003 08:23p 747 oracle_apache.conf 06/25/2002 10:55p 305 srm.conf 6 File(s) 49,588 bytes Directory of D:\ORACLE\infra902\Apache\Apache\conf\osso 04/23/2003 08:20p 433 osso.conf 1 File(s) 433 bytes Directory of D:\ORACLE\infra902\Apache\Jserv\conf 04/23/2003 08:04p 10,763 jserv.conf 1 File(s) 10,763 bytes Directory of D:\ORACLE\infra902\Apache\jsp\conf 04/23/2003 08:23p 598 ojsp.conf 1 File(s) 598 bytes Directory of D:\ORACLE\infra902\Apache\modplsql\conf 04/23/2003 08:23p 842 cache.conf 04/23/2003 08:23p 1,485 dads.conf 04/23/2003 08:23p 1,606 plsql.conf 3 File(s) 3,933 bytes Directory of D:\ORACLE\infra902\Apache\oradav\conf 04/23/2003 08:23p 789 moddav.conf 04/23/2003 08:23p 2 oradav.conf 2 File(s) 791 bytes Directory of D:\ORACLE\infra902\dcm\config 02/17/2004 01:31p 188 dcm.conf 1 File(s) 188 bytes Directory of D:\ORACLE\infra902\dcm\config\plugins\apache 06/27/2002 11:01p 43,623 httpd.conf 1 File(s) 43,623 bytes Directory of D:\ORACLE\infra902\dcm\repository.install\dcm\config 04/23/2003 08:24p 187 dcm.conf 1 File(s) 187 bytes Directory of D:\ORACLE\infra902\ldap\das 04/23/2003 08:23p 165 oiddas.conf 1 File(s) 165 bytes Directory of D:\ORACLE\infra902\oem_webstage 04/23/2003 08:23p 943 oem.conf 1 File(s) 943 bytes Directory of D:\ORACLE\infra902\opmn\conf 02/17/2004 01:31p 45 ons.conf 1 File(s) 45 bytes Directory of D:\ORACLE\infra902\RDBMS\demo 04/23/2003 08:23p 477 aqxml.conf 1 File(s) 477 bytes Directory of D:\ORACLE\infra902\sqlplus\admin 04/23/2003 08:23p 1,454 isqlplus.conf 1 File(s) 1,454 bytes Directory of D:\ORACLE\infra902\sso\conf 04/23/2003 08:23p 154 sso_apache.conf 1 File(s) 154 bytes Directory of D:\ORACLE\infra902\ultrasearch\webapp\config 04/23/2003 08:23p 324 ultrasearch.conf 1 File(s) 324 bytes Directory of D:\ORACLE\infra902\xdk\admin 04/23/2003 08:23p 291 xml.conf 1 File(s) 291 bytes Directory of D:\ORACLE\ora901\Apache\Apache\conf 08/20/2001 11:00a 285 access.conf 04/23/2003 07:26p 43,205 httpd.conf 04/23/2003 07:33p 472 oracle_apache.conf 08/20/2001 11:00a 297 srm.conf 4 File(s) 44,259 bytes Directory of D:\ORACLE\ora901\Apache\Jserv\conf 04/23/2003 07:26p 6,710 jserv.conf 1 File(s) 6,710 bytes Directory of D:\ORACLE\ora901\Apache\jsp\conf 04/23/2003 07:33p 511 ojsp.conf 1 File(s) 511 bytes Directory of D:\ORACLE\ora901\Apache\modose\conf 04/23/2003 07:27p 637 ose.conf 1 File(s) 637 bytes Directory of D:\ORACLE\ora901\Apache\modplsql\cfg 04/23/2003 07:29p 318 plsql.conf 1 File(s) 318 bytes Directory of D:\ORACLE\ora901\BC4J 04/23/2003 07:33p 121 bc4j.conf 1 File(s) 121 bytes Directory of D:\ORACLE\ora901\oem_webstage 04/23/2003 07:33p 682 oem.conf 1 File(s) 682 bytes Directory of D:\ORACLE\ora901\rdbms\demo 04/23/2003 07:26p 326 aqxml.conf 1 File(s) 326 bytes Directory of D:\ORACLE\ora901\sqlplus\admin 04/23/2003 07:33p 1,476 isqlplus.conf 1 File(s) 1,476 bytes Directory of D:\ORACLE\ora901\ultrasearch\jsp\admin\config 05/02/2001 08:26p 10,681 mod__ose.conf 1 File(s) 10,681 bytes Directory of D:\ORACLE\ora901\xdk\admin 04/23/2003 07:33p 253 xml.conf 1 File(s) 253 bytes Total Files Listed: 71 File(s) 321,045 bytes 33. Deploying J2EE Applications: ---------------------------------- You can deploy J2EE applications using the OC4J Home Page on the Enterprise Manager Web site. To navigate to an OC4J Home Page, do the following: -Navigate to the Instance Home Page where the OC4J instance resides. Scroll to the System Components section. -Select the OC4J instance in the Name column. This opens the OC4J Home Page for that OC4J instance. -Scroll to the Deployed Applications section on the OC4J Home Page. Clicking Deploy EAR File or Deploy WAR File starts the deployment wizard, which deploys the application to the OC4J instance and binds any Web application to a URL context. Your J2EE application can contain the following modules: -- Web applications The Web applications module (WAR files) includes servlets and JSP pages. -- EJB applications The EJB applications module (EJB JAR files) includes Enterprise JavaBeans (EJBs). -- Client application contained within a JAR file Now archive the JAR and WAR files that belong to an enterprise Java application into an EAR file for deployment to OC4J. The J2EE specifications define the layout for an EAR file. The internal layout of an EAR file should be as follows: - |--META_INF | | | -----application.xml | |--EJB JAR file | |--WEB WAR file | |--Client JAR file | When you deploy an application within a WAR file, the application.xml file is created for the Web application. When you deploy an application within an EAR file, you must create the application.xml file within the EAR file. Thus, deploying a WAR file is an easier method for deploying a Web application. ------------- 34. Errors: ------------- -- TROUBLESHOOTING 9iAS Rel. 2 -- Version 2.0 -- 4 juli 2004 -- Albert van der Sel With an 9iAS Release 2 Full install (Business Inteligence install), a tremendous amount of errors might be encountered. Here you will find my own experiences, as well as some threads from metalink. OPMN = Oracle Process Manager and Notification Server JAZN / JAAS = Oracle Application Server Java Authentication and Authorization Service DCM = Distributed Configuration Management OPMN stands for 'oracle process management notification' and is Oracle's 'high availability' system. OPMN monitors processes and brings them up again automatically if they go down. It is started when you start enterprise manager website with emctl start from the prompt in the infrastructure oracle home, and doing this starts 2 opmn processes for each oracle home. OPMN consists of two components - Oracle Process Manager and Oracle Notification System. DCM stands for 'distributed component management' and is the framework by which all IAS R2 components hang together. DCM is a layer that ensures that if something is changed in one components, others like Enterprise Manager are made aware as well. It is not a process as such, but rather a generic term for a framework and utilities. It is controlled directly with the dcmctl command. DMS Dynamic Monitoring Services . These processes are started when you start ohs. DMS basically gathers information on components. Jserv Jserv works in much the same way as R1 except oracle components no longer use this servlet architecture, but use oc4j instead. mod_plsql works the same way as R1. mod_oradav oradav allows web folders to be shared with clients e.g. PC's and accessed as if they were NT folders. OC4J_DAS is used by Portal for the management of users and groups. You access this via http://machine:port/oiddas ============================ PART 1: GENERAL 9iAS ERRORS: ============================ 1. troubleshooting the targets.xml: =================================== If you change the HOSTNAME for the repository (infrastructure) database, then you need to update the ssoServerMachineName property for the oracle SSO target in INFRA_ORACLE_HOME/sysman/emd/targets.xml The $ORACLE_HOME/sysman/emd/targets.xml file is created during installation of 9iAS and includes descriptions of all currently known targets. This file is used as the source of targets for the EM Website. sample targets.xml: - - - + - .. .. - - - - The above file stores amongst other things, the encrypted passwords that EM uses for access to components. Search for oracle_portal, oracle_repserv etc. Although encrypted, you can change these to be a password in Englidh as long as you flag it ENCRYPTED=FALSE. This should only be done for specific bug problems as recommended by oracle support. Do not change these passwords for any other reason!! The following is a list of things to check when there appears to be a problem with targets.xml. 1. Check the permissions on the active targets.xml file and restart all the infrastructure components (database, listener, oid, emctl in that order). The targets.xml file should be owned by the user who installed 9iAS and who starts emctl. Accidentally starting emctl as root recreates the targets.xml under root ownership. Fix this by changing ownership on targets.xml and restarting emctl. 2. Check which targets are listed, to ensure there is information on each expected target. 3. Check whether the hosts file and targets.xml have matching hostnames, and whether both have fully qualified hostnames. 4. What should be done if targets.xml is empty, or missing targets? a. Restore targets.xml from backup b. Copy $ORACLE_HOME/sysman/emd/discoveredTargets.xml to $ORACLE_HOME/sysman/emd/targets.xml, although it may not be complete if additional targets were installed following installation. See EM Website has no Entries for the 9iAS Instances 226226.1 and EM Web Site Fails to Display Application Servers 210552.1 and Login as ias_admin to 9iAS R2 Enterprise Manager, A Blank List is Displayed for Targets 209540.1 c. Check the amount of disk space available. See Bug 2508930 - TARGETS.XML IS EMPTY IF WE HAVE NO DISK SPACE. d. Reinstall. See De-Installing 9iAS Release 2 (9.0.2) From Unix Platforms 218277.1 5. Is there an Infrastructure and Mid-Tier install on the system? When installing both the infrastructure and a mid-tier on the same server (in different homes), the installation of the infrastructure creates the emtab file pointing to its own home. During installation of the mid-tier, the mid-tier installation routine uses the emtab file pointing to the infrastructure home so it knows where to write configuration information required for the infrastructure EM Website, so it can see not only information concerning itself but also information related to the mid-tier. If the emtab file is removed/renamed after installation of of the infrastructure but before installation of the mid-tier, a new emtab file is created pointing to the mid-tier home. The configuration file routines of the mid-tier installation therefore do not know about the existence of the infrastructure and write the new configuration information into files in its own home and not into the files in the infrastructure home. In addition to entries in the targets.xml in the infrastructure home, other files such as the ias.properties file in the infrastructure home are also updated with information concerning the mid-tier. Merging the targets.xml file from both homes may solve some of the display problems, though they may not solve control of component issues due to incomplete configuration files in the infrastructure home. References to renaming the emtab file should be disregarded when performing infrastructure/mid-tier installs on the same server, and may have in fact been specific to certain platforms and specific for certain circumstances. The EM Web Site is launched as a J2EE application. The configuration files consist of many XML files and properties files. Here are some of those files: targets.xml emd.properties logging.properties iasadmin.properties 2. Cleanly Restarting OID After A 9iAS 9.0.2 Crash: =================================================== A problem that often seems to happen when Oracle 9iAS 9.0.2 crashes is that you can't seem to restart OID using OIDCTL. For example, a situation might arise when a server is bounced without 9iAS being shut down cleanly. When you reboot the PC, and use DCMCTL to check the status of the OC4J instances prior to starting them, you get the following error message: C:\ocs_onebox\infra\dcm\bin>dcmctl getState -V ADMN-202026 A problem has occurred accessing the Oracle9iAS infrastructure database. Base Exception: oracle.ias.repository.schema.SchemaException:Unable to connect to Directory Server :javax.naming.CommunicationException: markr.plusconsultancy.co. uk:4032 [Root exception is java.net.ConnectException: Connection refused: connect] Please, refer to the base exception for resolution, or call Oracle support. Or, when you watch an ias start script, at the point oid get started, you will see C:\ocs_onebox\infra\bin>oidctl server=oidldapd configset=0 instance=1 start which should startup an OID instance. However, sometimes this fails to work and you get the error message: C:\ocs_onebox\infra\bin>oidctl server=oidldapd configset=0 instance=1 start *** Instance Number already in use. *** *** Please try a different Instance number. *** oidmon is the 'monitor' process. It pools the database ( table ODS.ODS_PROCESS ) for new ldap server launch requests, and if it finds one, (also placed there by oidctl as user ODSCOMMON ) , then it starts a 'dispatcher/listener process.' As such, oidctl does not actually start the ldap processes. Oidmon then spawns 'dispatcher' and 'server' oidldapd processes. What actually happens behind the scenes is that a row is inserted or updated in the ODS.ODS_PROCESS table that contains the instance name (which must be unique), the process ID, and a flag called 'state', which has three values - 0,1,2 and 3 which stand for stop, start, running and restart. A second process, OIDMON, polls the ODS.ODS_PROCESS table and when it finds a row with state=0, it reads the pid and stops the process. When it finds a state=1, oidmon starts a new process and updates pid with a new process id. With state=2, oidmon reads the pid, and checks that the process with the same pid is running. If it's not, oidmon starts a new process and updates the pid. Lastly, with state=3, oidmon reads the pid, stops the process, starts a new one and updates the pid accordingly. If oidmon can't start the server for some reason, it retries 10 times, and if still unsuccessful, it deletes the row from the ODS.ODS_PROCESS table. Therefore, OIDCTL only inserts or updates state information, and OIDMON reads rows from ODS.ODS_PROCESS, and performs specified tasks based on the value of the state column. This all works fine except when 9iAS crashes; when this happens, OIDMON exits but the OIDLDAPD processes are not killed, and in addition, stray rows are often left in the ODS.ODS_PROCESS table that are detected when you try to restart the oidldapd instance after a reboot. The way to properly deal with this is to take two steps. 1. Kill any stray OIDLDAPD processes still running (if you haven't rebooted the server since the crash) 2. Delete any rows in the ODS.ODS_PROCESS table connect to the IASDB database as the ODS user, or as SYSTEM select * from ODS.ODS_PROCESS; (there should be at least one row) delete form ODS.ODS_PROCESS; commit; 3. Restart the OID instance again, using C:\ocs_onebox\infra\bin>oidctl server=oidldapd configset=0 instance=1 start OID uses the configfile: $INFRA_ORACLE_HOME/network/admin/ldap.ora Sample: # LDAP.ORA Network Configuration File: d:\oracle\infra902\network\admin\ldap.ora # Generated by Oracle configuration tools. DEFAULT_ADMIN_CONTEXT = "" DIRECTORY_SERVERS= (missrv.miskm.mindef.nl:4032:4031) DIRECTORY_SERVER_TYPE = OID 3. Deobfuscate Errors After Reboot, Crash, or Network Change. ============================================================= This can occur occur under these scenarios: * A reboot has just occurred for the first time after 9iAS was installed. (And, you had to change the /etc/hosts file during installation) OR * A system crash occurred, and trying to recover. The 9iAS installion is placed on a machine with the same hostname and IP address as before the crash occurred. OR * Hardware changes have occurred to the machine. (ie, CPU, NIC) AND * Everything was working under the current 9iAS configuration. (A 9iAS configuration change causing this can be a different problem) There are different times when this error can occur. But, it basically occurs when a change to the system has been done. This can be after a reboot or a crash, but there is a difference on the machine before and after the occurance. It is usually a network configuration change that has caused the problem. When you try to start the Oracle HTTP Server, the following error might appear in the opmn logs: "Syntax error on line 6 of OH/Apache/Apache/conf/mod_osso.conf: Unable to deobfuscate the SSO server config file, OH/Apache/Apache/conf/osso/osso.conf, error Bad padding pattern detected in the last block." Most of the Mid-Tier components will fail to connect to the Infrastructure, and will give the following error: "oracle.ias.repository.schema.SchemaException:Password could not be retrieved" Possible solution: 1. Start Infrastructure DB 2. Start the Infrastructure OID 3. Include $ORACLE_HOME/lib in the LD_LIBRARY_PATH, SHLIB_PATH, or LIBPATH environment variable, depending on your platform. -For AIX LIBPATH=$ORACLE_HOME/lib:$ORACLE_HOME/lib64:$LIBPATH; export LIBPATH -For HPUX SHLIB_PATH=$ORACLE_HOME/lib32:$ORACLE_HOME/lib:$SHLIB_PATH; export SHLIB_PATH -For Solaris, Linux and Tru64 LD_LIBRARY_PATH=$ORACLE_HOME/lib:$LD_LIBRARY_PATH; export LD_LIBRARY_PATH 4. Run the command to reset the iAS password. Please use the SAME password, as we are not attempting to change the password you enter when signing onto EM. That is done with the emctl utility. This command changes it internally, and we want to re-instate the current obfuscated password: resetiASpasswd.sh "cn=orcladmin" <$ORACLE_HOME> Note: There is a resetiASpasswd.bat on Windows, to be used the same way, just in case these steps are followed on Windows. The above stated problem is specific to UNIX, but there may be occasions to run through the same steps. 5. Use the showPassword utility to obtain the password for the orasso user. Then, re-register the listener, being sure to add this information to the ossoreg command in Step 6: -schema orasso -pass ReplaceWithPassword 6. Run the command to re-register mod_osso. * Make sure there are no spaces after the trailing '\'s If on Windows, use all one line, withouth the "\" * Replace the uppercase with proper items * The following assumes the to-be registered http server is on the mid-tier * If on Windows, use "SYSTEM", instead of "root" for -u java -jar $ORACLE_HOME/sso/lib/ossoreg.jar \ -host $INFRA_HOST \ -port 1521 \ -sid iasdb \ -site_name MID_HOST:MID_PORT\ -oracle_home_path $ORACLE_HOME \ -success_url http://MID_HOST:MID_PORT/osso_login_success \ -logout_url http://MID_HOST:MID_PORT/osso_logout_success \ -cancel_url http://MID_HOST:MID_PORT/ \ -home_url http://MID_HOST:MID_PORT/ \ -config_mod_osso TRUE \ -u root \ -sso_server_version v1.2 \ -schema orasso \ -pass NOTE: The following command will not work on 9iAS 9.0.2.0.x, unless a patched dcm.jar has previously been applied with a patch (or 9.0.2.1). Since this cannot be run on previous versions, just proceed to step 8. 7. Run following commands on the machine where the change occurred, (not the associated Mid-Tiers): a. Solaris i. $ORACLE_HOME/dcm/bin/dcmctl resetHostInformation ii. $ORACLE_HOME/bin/emctl set password b. NT i. Make sure the Oracle9iAS is stopped ii. Edit %ORACLE_HOME%\sysman\j2ee\config\jazn-data.xml iii.Search for ïias_adminï iv. Replace obfuscated text between and with "!" where "" is the password. Example: !welcome1 v. Save the file. 8. Continue starting 9iAS, as in Note 200475.1. The next step is: % dcmctl start -ct ohs This is what was originally failing. After successfully starting OHS, You may want to take a backup of the deobfuscated information as described in Note215955.1. 3. Not able to access the Middle Tier from EM Website. ====================================================== 3.1 --- Thread Status: Active From: Ishaq Baig 19-Nov-03 10:47 Subject: Enable to Access the Middle Tier Instance from EM Website RDBMS Version: 8.1.7 Operating System and Version: WIN2K Service Pack3 Product (i.e., OAS, IAS, etc): IAS Product Version: 9.0.2 JDK Version: 1.3.1.9 Error number: Enable to Access the Middle Tier Instance from EM Website Hi, We have an 9IAS (9.0.2) Infrastructure and Middle Tier instance running on ONE Box (Win2k),thing we fine until while trying to implement the Single Sigon after making the changes as instructed in Note:199072.1 we stopped the HTTP Server so that change could take effect,but every since we have stopped the HTTP Server we couldn't gain access to the Middle Instance from the EM WEB SITE the page just hangs......on the other hand the INFRASTRUCTURE instance is working fien.We even tried starting the HTTP server through the DCM UTILITY the following was the message Content-Type: text/html Response: 0 of 1 processes started. Check opmn log files such as ipm.log and ons.log for detailed.". Resolve the indicated problem at the Oracle9iAS instance where it occurred thenresync the instance Remote Execute Exception 806212 oracle.ias.sysmgmt.exception.ProcessMgmtException: OPMN operation failure at oracle.ias.sysmgmt.clustermanagement.OpmnAgent.validateOperation(Unknown Source) at oracle.ias.sysmgmt.clustermanagement.OpmnAgent.startOHS(Unknown Source) at oracle.ias.sysmgmt.clustermanagement.StartInput.execute(Unknown Source) at oracle.ias.sysmgmt.clustermanagement.ClusterManager.execute(Unknown Source) at oracle.ias.sysmgmt.task.ClusterManagementAdapter.execute(Unknown Source) at oracle.ias.sysmgmt.task.TaskMaster.execute(Unknown Source) at oracle.ias.sysmgmt.task.TaskMasterReceiver.process(Unknown Source) at oracle.ias.sysmgmt.task.TaskMasterReceiver.handle(Unknown Source) at oracle.ias.sysmgmt.task.TaskMasterReceiver.run(Unknown Source) is Any Inputs highly appreciated,we need to get it up as soon as possible. Regards Ishaq Baig From: Oracle, Rhoderick Butial 19-Nov-03 14:36 Subject: Re : Enable to Access the Middle Tier Instance from EM Website Hello, What type of changes did you alter? Did you try restarting all of the other components on the mid tier? There should be some errors generated in the error_log file, please post these errors in your next reply. You may want to review the following notes: Note.236112.1 Wrong user supplied to ossoreg causing ADMN-906025 exception, 806212 Note.223586.1 Starting Oracle HTTP Server gives ADMN-906025 error Note.222051.1 Starting Oracle HTTP Server gives ADMN-906025 Error Also, I noticed that you have listed your 9iAS version as 9.0.2, did you apply the latest patchsets before implementing the changes? If not, you will need to apply the patchsets first before making the changes. Please review.. Note.215882.1 9iAS Release 2 Patching Recommendations Within the Version Lifecycle Thank you, Rod Oracle Technical Support 3.2 --- Displayed below are the messages of the selected thread. Thread Status: Closed From: Ron Miller 28-Oct-03 16:13 Subject: EM Website extremely slow for 9iAS RDBMS Version:: 9.0.1.3.0 Operating System and Version:: AIX 4.3.3 Product (i.e. Trace, DB Diff, Expert, etc):: Oracle9i Application Server Product Version:: 9.0.2.2.0 OEM Console Operating System and Version:: Windows 2000 EM Website extremely slow for 9iAS When I use the EM website to access the components of my 9i App server, the response time is very slow. It takes 2 or 3 minutes to go from screen to another. I have found information on this forum that others are experiencing the same problem. The response from Oracle support has been that this is a known problem and there is a bug, 2756262, which is to be fixed in 9.0.4. However, I cannot find any information on when this release will be available. It seems to keep getting pushed back. Does anyone know a release date? Has anyone requested a backport of this fix to an earlier release? Thanks for any response. From: Oracle, Kathy Ting 29-Oct-03 05:41 Subject: Re : EM Website extremely slow for 9iAS The base architecture is being redesign. Due to the redesign, backports are not being accepted. Look for a much better improved EM website in future releases. Thank you for using the MetaLink Forum, Kathy Oracle Support. From: Ron Miller 29-Oct-03 14:52 Subject: Re : Re : EM Website extremely slow for 9iAS Thanks for the reply Kathy. I will look forward to the redesign since the current product is pretty much useless. From: Oracle, Kathy Ting 29-Oct-03 22:04 Subject: Re : Re : Re : EM Website extremely slow for 9iAS As do we. Thank you for using the MetaLink Forum, Kathy Oracle Support. 4. Explanation of IAS_ADMIN and ORCLADMIN Accounts ================================================== Note:244161.1 Subject: Explanation of IAS_ADMIN and ORCLADMIN Accounts Type: BULLETIN Status: PUBLISHED PURPOSE ------- To provide an explanation for the IAS_ADMIN and ORCLADMIN accounts that are established with Oracle9i Application Server (9iAS) Release 2 (9.0.2.x). SCOPE & APPLICATION ------------------- Website Administrators installing and maintaining 9iAS Explanation of IAS_ADMIN and ORCLADMIN Accounts ------------------------------------------------ There are two users that can create some confusion: ias_admin and orcladmin. However, the interaction is more or less internally managed. You log into the EM Website with ias_admin, but use the orcladmin password after initially installing 9iAS. So when changing the orcladmin password, you may not get the results intended with the ias_admin login. But, if the obfuscation gets skewed, we found you sometimes need to reinstate the password obfuscation between the two with the resetiASpasswd script. This assumes the same password is used, and no resulting changes are noted. The *change* occurred internally. These changes, and methods, can cause some confusion. You can actually change the EM Website login separately with the emctl utility. Or, change the orcladmin username separately, depending on how your want to manage this. IAS_ADMIN Account ----------------- In EM 9.0.2 and 9.0.3, you will need to use the IAS_ADMIN account to access the EM Website Home Page. This account is not known within the database or to the Oracle Management Server. Instead, it is a new account used only for access to the 9iAS Administration (EM) Web Site. The following note can be used to supplement the Documentation and Release Notes dealing with modifying this password: [NOTE:204182.1] How to Change the IAS_ADMIN password for Enterprise Manager NOTE: If you change the IAS_ADMIN password (as described in Note:204182.1), the ORCLADMIN password does not change. ORCLADMIN Account ----------------- ORCLADMIN is used as a superuser account for administering 9iAS. During the initial installation of 9iAS, the installer prompts you to create the IAS_ADMIN password. This password is then also assigned to the ORCLADMIN account. To reset (not change) the ORCLADMIN password, you must run the script, ResetiASpasswd.sh. $ORACLE_HOME/bin/resetiASpasswd.sh "cn=orcladmin" <$ORACLE_HOME> Note: There is a resetiASpasswd.bat on Windows, to be used the same way. If you suspect that the encryption is skewed, use the SAME password, to *reset* this. If you desire to change the password you enter when signing onto EM, use the emctl utility, (as described in Note:204182.1). If you actually want to change the ORCLADMIN password, you should use the Oracle Directory Manager, to modify this super user. - Start the Directory Manager from $ORACLE_HOME/bin/oidadmin - In the navigator pane, expand Oracle Internet Directory Servers. - Select a server. The group of tab pages for that server appear in the right pane. - Select the System Passwords tab. This page displays the current user names and passwords for each type of user. Note that passwords are not displayed in the password fields. SUMMARY ------- Is the goal to reset the internally encrypted ias_admin password, change the actual orcladmin password, or just change the password when logging onto EM? Thats the main question to ask. 1. To reset the internally encrypted ias_admin password, use the resetiASpasswd script, and use the same password as previously given. 2. To change the orcladmin password, it is best to use the Oracle Directory Manager. Please see the Oracle Internet Directory Administrator's Guide for more information. 3. Change the EM website or emctl password: Within the EM Web Site...Preferences link...top right-hand side of the screen. Or, on command line, using emctl. RELATED DOCUMENTS ----------------- [NOTE:234712.1] Managing Schemas of the 9iAS Release 2 Metadata Repository [NOTE:253149.1] Resetting the Single Sign-On password for ORCLADMIN . 5. Password for ORASSO Database Schema ====================================== Password for ORASSO Database Schema goal: What is the password for the ORASSO database schema? fact: Oracle9i Application Server Enterprise Edition 9.0.2 fact: Oracle9iAS Single Sign-On 9.0.2 fix: During installation a random password is generated for the ORASSO database schema. You need to look up this password in the Oracle Internet Directory. The following text is taken from the Interoperability Patch Readme (a patch that was mandatory for 9.0.2.0.0 but is no longer needed for 9.0.2.0.1): If you do not know the password for the orasso schema, you can use the following procedure to determine the password: Note: Do not use the "alter user" SQL command to change the orasso password. If you need to change the orasso password, use Enterprise Manager so that it can propagate the password to all components that need to access orasso. Start up the Oracle Internet Directory administration tool from infrastructure machine. prompt> $ORACLE_HOME/bin/oidadmin Log into the oidadmin tool using the OID administrator account (cn=orcladmin) for the Infrastructure installation. Username: cn=orcladmin Password: administrator_password Server : host running Oracle Internet Directory and port number where Oracle Internet Directory is listening The administrator password is the same as the ias_admin password. The default port for Oracle Internet Directory is 389 (without SSL). Navigate the Single Sign-On schema (orasso) entry using the administration tool. > cn=orcladmin@OID_hostname:OID_port (for example: cn=orcladmin@infra.acme. com:389) > Entry Management > cn=OracleContext > cn=Products > cn=IAS > cn=Infrastructure Databases > orclReferenceName=Single Sign-On database SID:Single Sign-On Server hostname (for example: orclReferenceName=iasdb:infra.acme.com) > orclResourceName=ORASSO Click the above entry and look for the orclpasswordattribute attribute value on the right panel. This value is the password for the orasso schema. NOTE: If you have multiple Infrastructures installed using one Oracle Internet Directory, ensure that you are looking at the correct Single Sign-On database entry since all the infrastructure instances would have an ORASSO schema entry, but only one of them is actually being used. 6. Windows Script to Determine orasso Password in 9iAS Release 2 (9.0.2) ======================================================================== Note:205984.1 Subject: Windows Script to Determine orasso Password in 9iAS Release 2 (9.0.2) Type: BULLETIN Status: PUBLISHED PURPOSE ------- The showPassword utility was developed to avoid having to use the oidadmin tool to look up various OID passwords, by using ldapsearch with Oracle9i Application Server (9iAS) Release 2 (9.0.2). As a script, varying on different environments, it is not supported by Oracle Support Services. It is intended as an example, to aid in the understanding of the product. SCOPE & APPLICATION ------------------- 9iAS Administrators and Windows Administrators Windows Script to Determine orasso Password in 9iAS Release 2 (9.0.2) --------------------------------------------------------------------- 1. Paste the following script in a file named showPassword.bat and copy it in a directory. Please also ensure that ldapserach is there in PATH on your widnows machine. 8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8< set OIDHOST=bldel18.in.oracle.com set OIDPORT=4032 if "%1"== "" goto cont if "%2"== "" goto cont ldapsearch -h %OIDHOST% -p %OIDPORT% -D "cn=orcladmin" -w "%1" -b "cn=IAS Infrastructure Databases,cn=IAS,cn=Products,cn=OracleContext" -s sub "orclResourceName=%2" orclpasswordattribute goto :end :cont echo Correct Syntax is echo showpassword.bat orcladminpassword username :end 8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8<8< Note that the "ldapsearch...orclpasswordattribute" commands should be put on one line. 2. Edit the script and update with your own hostname and OID port OIDHOST=bldel18.in.oracle.com OIDPORT=4032 3. Ensure that you have ldapsearch from the correct ORACLE_HOME in the PATH 4. Check that OID is up and running before proceeding. 5. Run the script, and enter the schema name as: orasso, and the password value is shown. For example: (all ONE line...may be easier to copy/paste from Notepad) C:\> showPassword.bat oracle1 orasso OrclResourceName=ORASSO,orclReferenceName=iasdb.bldel18.in.oracle.com,cn=IAS Inf rastructure Databases,cn=IAS,cn=Products,cn=OracleContext orclpasswordattribute=Gbn3Fd24 The orasso password in this example is Gbn3Fd24. 6. STARTING AND STOPPING 9iAS WITH SCRIPTS. =========================================== ---------------------------------------------------------------- 5.1 From metalink: a) StartInfrastructure.bat: REM #################################################### REM #################################################### REM ## Script to start Infrastructure ## REM ## ## REM #################################################### REM #################################################### REM ## REM ## Set environment variables for Infrastructure REM #################################################### set ORACLE_HOME=D:\IAS90201I set ORACLE_SID=IASDB set PATH=%ORACLE_HOME%\bin;%ORACLE_HOME%\dcm\bin;%ORACLE_HOME%\opmn\bin;%PATH%; REM ##################################################### REM ## Start Oracle Internet Directory processes REM ##################################################### echo .....Starting %ORACLE_HOME% Internet Directory ...... oidmon start oidctl server=oidldapd instance=1 start timeout 20 REM ##################################################### REM ## Start Oracle HTTP Server and OC4J processes REM ##################################################### echo .....Starting OHS and OC4J processes....... call dcmctl start -ct ohs call dcmctl start -ct oc4j REM ##################################################### REM ## Check OHS and OC4J processes are running REM ##################################################### echo .....Checking OHS and OC4J status..... call dcmctl getstate -v pause REM #################################################### b) StartMidTier.bat: REM #################################################### REM #################################################### REM ## Script to start MidTier ## REM ## ## REM #################################################### REM #################################################### REM ## REM ## Set environment variables for Midtier REM #################################################### set ORACLE_HOME=D:\IAS90201J set PATH=%ORACLE_HOME%\bin;%ORACLE_HOME%\dcm\bin;%ORACLE_HOME%\opmn\bin;%PATH%; REM ##################################################### REM ## Start Oracle HTTP Server and OC4J processes REM ##################################################### echo .....Starting OHS and OC4J processes....... call dcmctl start -ct ohs call dcmctl start -ct oc4j REM ##################################################### REM ## Check OHS and OC4J processes are running REM ##################################################### echo .....Checking OHS and OC4J status..... call dcmctl getstate -v REM #################################################### REM ## Start Webcache REM #################################################### echo .....Starting Webcache.......... webcachectl start REM #################################################### REM ## Start Enterprise Manager Website REM #################################################### echo .....Starting EM Website..... net start Oracleias90201iEMWebsite echo ....Done pause REM #################################################### c) StopMidTier.bat: REM #################################################### REM #################################################### REM ## Script to stop Midtier ## REM ## ## REM #################################################### REM #################################################### REM ## REM ## Set environment variables for Midtier REM #################################################### set ORACLE_HOME=D:\IAS90201J set PATH=%ORACLE_HOME%\bin;%ORACLE_HOME%\dcm\bin;%ORACLE_HOME%\opmn\bin;%PATH%; REM #################################################### REM ## Stop Enterprise Manager Website REM #################################################### echo .....Stopping EM Website..... net stop Oracleias90201iEMWebsite REM #################################################### REM ## Stop Webcache REM #################################################### echo .....Stopping %ORACLE_HOME% Webcache.......... webcachectl stop REM #################################################### REM ## Stop Oracle HTTP Server and OC4J processes REM #################################################### echo .....Stopping %ORACLE_HOME% OHS and OC4J........ dcmctl shutdown echo ....Done pause REM #################################################### d)StopInfrastructure.bat: REM #################################################### REM #################################################### REM ## Script to stop Infrastructure ## REM ## ## REM #################################################### REM #################################################### REM ## REM ## Set environment variables for Infrastructure REM #################################################### set ORACLE_HOME=D:\IAS90201I set ORACLE_SID=IASDB set PATH=%ORACLE_HOME%\bin;%ORACLE_HOME%\dcm\bin;%ORACLE_HOME%\opmn\bin;%PATH%; set EM_ADMIN_PWD= REM #################################################### REM ## Stop Enterprise Manager Website REM #################################################### echo .....Stopping EM Website..... call emctl stop REM #################################################### REM ## Stop Oracle HTTP Server and OC4J processes REM #################################################### echo .....Stopping %ORACLE_HOME% OHS and OC4J........ call dcmctl shutdown REM ##################################################### REM ## Stop Oracle Internet Directory processes REM ##################################################### echo .....Stopping %ORACLE_HOME% Internet Directory ...... oidctl server=oidldapd configset=0 instance=1 stop timeout 20 oidmon stop echo ....Done pause REM ##################################################### ---------------------------------------------------------------- 5.2 Our scripts: Starting: ========= @ECHO OFF TITLE Startup all REM ********************************************************** REM Adjust the following values set ORACLE_BASE=D:\oracle set IAS_HOME=%ORACLE_BASE%\ias902 set IAS_BIN=%IAS_HOME%\bin set INFRA_HOME=%ORACLE_BASE%\infra902 set INFRA_BIN=%INFRA_HOME%\bin REM ********************************************************** echo ********************************************************** echo Parameters used are: echo ORACLE_BASE = %ORACLE_BASE% echo IAS_HOME = %IAS_HOME% echo IAS_BIN = %IAS_BIN% echo INFRA_HOME = %INFRA_HOME% echo INFRA_BIN = %INFRA_BIN% echo ********************************************************** echo ********************************************************** echo "Starting up infra" echo ********************************************************** echo "Starting iasdb instance" echo connect sys/change_on_install as sysdba > $$tmp$$ echo startup >> $$tmp$$ echo exit >> $$tmp$$ %INFRA_BIN%\sqlplus /nolog < $$tmp$$ del $$tmp$$ echo "Starting Oracle Internet Directory..." %INFRA_BIN%\oidmon start %INFRA_BIN%\oidctl server=oidldapd instance=1 start timeout 10 echo "Starting Enterprise manager Services..." net start Oracleinfra902EMWebsite echo "Starting OEM ..." net start Oracleinfra902ManagementServer rem net start Oracleinfra902TNSListener net start Oracleinfra902Agent echo "Starting up infra services..." %INFRA_HOME%\opmn\bin\opmnctl startall echo ********************************************************** echo "Done kickin' up infra!" echo ********************************************************** echo. echo ********************************************************** echo "Starting all mid tier services..." echo ********************************************************** %IAS_HOME%\opmn\bin\opmnctl startall echo "Starting webcache..." %IAS_BIN%\webcachectl start echo "Starting all services..." net start Oracleias902Discoverer rem net start Oracleias902ProcessManager rem net start Oracleias902WebCacheAdmin rem net start Oracleias902WebCache echo ********************************************************** echo "Done starting up mid tier!" echo ********************************************************** pause Stopping: ========= @ECHO OFF TITLE Shutdown all REM ********************************************************** REM Adjust the following values set ORACLE_BASE=D:\oracle set IAS_HOME=%ORACLE_BASE%\ias902 set IAS_BIN=%IAS_HOME%\bin set INFRA_HOME=%ORACLE_BASE%\infra902 set INFRA_BIN=%INFRA_HOME%\bin REM ********************************************************** echo ********************************************************** echo Parameters used are: echo ORACLE_BASE = %ORACLE_BASE% echo IAS_HOME = %IAS_HOME% echo IAS_BIN = %IAS_BIN% echo INFRA_HOME = %INFRA_HOME% echo INFRA_BIN = %INFRA_BIN% echo ********************************************************** echo ********************************************************** echo "Shutting down mid tier..." echo ********************************************************** echo "Stopping all mid tier services..." %IAS_HOME%\opmn\bin\opmnctl stopall echo "Stopping webcache..." %IAS_BIN%\webcachectl stop echo "Stopping Discoverer service..." net stop Oracleias902Discoverer echo "Sanity stops for WebCache" net stop Oracleias902WebCache net stop Oracleias902WebCacheAdmin echo ********************************************************** echo "Done shutting down mid tier!" echo ********************************************************** echo. echo ********************************************************** echo "Shutting down Infrastructure..." echo ********************************************************** echo "Stopping Enterprise Manager Website" call %INFRA_BIN%\emctl stop welcome1 echo "Stopping Enterprise Manager Management Console..." call %INFRA_BIN%\oemctl stop oms sysman/sysman echo "Stopping Infra Services..." %INFRA_HOME%\opmn\bin\opmnctl stopall echo "Stopping Oracle Internet Directory..." %INFRA_BIN%\oidctl server=oidldapd instance=1 stop timeout 10 %INFRA_BIN%\oidmon stop echo "Stopping infra database..." echo connect sys/change_on_install as sysdba > $$tmp$$ echo shutdown immediate >> $$tmp$$ echo exit >> $$tmp$$ %INFRA_BIN%\sqlplus /nolog < $$tmp$$ del $$tmp$$ echo "Stopping all Remaining NT Services..." rem net stop Oracleinfra902TNSListener net stop Oracleinfra902Agent echo ********************************************************** echo "Done shutting down infra!" echo ********************************************************** pause Starting BI: ============ @echo off title Starting Oracle Reports rem ******************************************************************** set IAS_HOME=d:\oracle\ias902 set IAS_BIN=%IAS_HOME%\bin rem ******************************************************************** echo ******************************************************************** echo Parameters used: echo. echo IAS_HOME = %IAS_HOME% echo IAS_BIN = %IAS_BIN% echo ******************************************************************** echo. echo ******************************************************************** echo Bringing up OC4J_BI_Forms (Business Intelligence/Forms) echo ******************************************************************** call %IAS_HOME%\dcm\bin\dcmctl start -co OC4J_BI_Forms -v timeout 5 echo Check to see if the instance really started up: echo. call %IAS_HOME%\dcm\bin\dcmctl getReturnStatus echo Done starting up OC4J_BI_FORMS... pause Starting CMSDK: =============== @echo off title Starting Oracle CM SDK 9.0.3.1. rem ******************************************************************** set IAS_HOME=d:\oracle\ias902 set IAS_BIN=%IAS_HOME%\bin rem ******************************************************************** echo ******************************************************************** echo Parameters used: echo. echo IAS_HOME = %IAS_HOME% echo IAS_BIN = %IAS_BIN% echo ******************************************************************** echo. echo ******************************************************************** echo Bringing up Domain Controller, note default password is: ifsdp echo ******************************************************************** call %IAS_HOME%\ifs\cmsdk\bin\ifsctl start echo Done bringing up Domain Controller echo. echo ******************************************************************** echo Bringing up OC4J Instance... echo ******************************************************************** call %IAS_HOME%\dcm\bin\dcmctl start -co OC4J_iFS_cmsdk -v timeout 5 echo Check to see if the instance really started up: echo. call %IAS_HOME%\dcm\bin\dcmctl getReturnStatus echo Done starting up OC4J Instance. echo Done starting up CM SDK. pause 8. Warning: Stop EMD Before Using DCMCTL Utility. ================================================= Note:207208.1 Subject: Warning: Stop EMD Before Using DCMCTL Utility Type: BULLETIN Status: PUBLISHED PURPOSE ------- Issue a warning for the use of the dcmctl utility when administering the Oracle9i Application Server (9iAS) Release 2 (9.0.2.0.x). There is now a Patch available which resolves the issue of running DCM and EM at the same time. SCOPE & APPLICATION ------------------- This article is intended for 9iAS Administrators. It gives a general description of a problem that can occur when dcmctl is used without precautions. DCMCTL RESTRICTIONS ------------------- 1. Do not use dcmctl while EMD (Enterprise Manager Console/Website) is running. The dcmctl utility is issuing DCM commands to control the state of components in 9iAS. The same is done from the EMD, which is generally reachable at the following URLs: http://yourserver:1810/emd/console http://yourserver:1810/ When the dcmctl utility is used while EMD is running, this may cause out-of-sync problems with your 9iAS instance. This is caused by only one DCM daemon being available to 'listen' to requests. How to Avoid Problems --------------------- Stop EMD: $ emctl stop Issue your command with dcmctl When you are done, restart EMD: $ emctl start 2. If an Infrastructure and Mid-Tier(s) are installed on same server, EM must be stopped when issuing dcmctl from either the Infrastructure or a Mid-tier directories. This is because EM is common to all 9iAS instances on the server. Stopping multiple instances of EM across multiple servers is not neccessary. The DCM/EM concurrency conflict will only come into play with instances on a given machine. 3. Do not issue multiple DCM commands at once, and do not issue a DCM command while one might still be running. 4. If you start a component with DCM, it is recommended to also stop it with DCM. If you start a component with the EM Website, it is recommended stop it with the EM Website. SOLUTION -------- If out-of-sync errors occur because of EM being up while using dcmctl, then a reinstall may be neccessary. Please apply the following patches in order to prevent this concurrency problem from happening inadvertently: Patch 2542920 : 9iAS 9.0.2.1 Core Patchset Patch 2591631 : DCM/EM Concurrency Fix * The 9.0.2.1 Patchset is a pre-requisite of the DCM Patch. * Both patches should be applied to all associated 9iAS Tiers. * Please refer to the readme for important information. * Future releases (9.0.2.2+) will have this fix included. 9. MISCELLANEOUS: ================= 9.1 Change of hostname: ----------------------- If you change the HOSTNAME for the repository (infrastructure) database, then you need to update the ssoServerMachineName property for the oracle SSO target in INFRA_ORACLE_HOME/sysman/emd/targets.xml If you change the PORT for the repository database, discoverer is affected - update the port for discodemo in tnsnames.ora. 9.2 Files with IP in the name: ------------------------------ 9.3 ldapcheck and ldapsearch examples: -------------------------------------- List users and or passwords: use ldapcheck and ldapsearch Example 1: ---------- ldapsearch -h uks799 -p 4032 -D "cn=orcladmin" -w your_ias_or_oid_password -b "cn=Users,dc=uk,dc=oracle,dc=com" -s sub -v "objectclass=*" set OIDHOST=bldel18.in.oracle.com set OIDPORT=4032 if "%1"== "" goto cont if "%2"== "" goto cont ldapsearch -h %OIDHOST% -p %OIDPORT% -D "cn=orcladmin" -w "%1" -b "cn=IAS Infrastructure Databases,cn=IAS,cn=Products,cn=OracleContext" -s sub "orclResourceName=%2" orclpasswordattribute goto :end :cont echo Correct Syntax is echo showpassword.bat orcladminpassword username :end C:\> showPassword.bat oracle1 orasso OrclResourceName=ORASSO,orclReferenceName=iasdb.bldel18.in.oracle.com,cn=IAS Inf rastructure Databases,cn=IAS,cn=Products,cn=OracleContext orclpasswordattribute=Gbn3Fd24 The orasso password in this example is Gbn3Fd24. Example 2: ---------- 9.4 dcmctl commands: -------------------- On a simple 9iAS webcache/j2ee installation, you might try the following command: F:\oracle\ias902\dcm\bin>dcmctl getstate -V Current State for Instance:ias902dev.localhost Component Type Up Status In Sync Status =========================================================================== 1 home oc4j Up True 2 HTTP Server ohs Up True 3 OC4J_Demos oc4j Up True 4 OC4J_iFS_cmsdk oc4j Up True dcmctl getstate -ct ohs - show status of ohs of the current instance ONLY. dcmctl updateConfig Atempt to update DCM's view of the world after a manual configuration change. dcmctl getstate -v determines which component aren't starting. dcmctl resyncInstance -force force resync of the instance. 9.5 Fault tolerance: ==================== 217368.1 from Metalink - "Advanced Configurations and Topologies for Enterprise Deployments of E-Business" Hot site Oracle disaster recovery configuration Oracle failover with Oracle standby database Oracle failover with Oracle9i Dataguard Oracle failover with Oracle9i TAF (Transparent Application Failover) Oracle failover with Oracle9i Real Application Clusters (RAC) |----------------------------------| |Machine A | | | | |-----------------------------| | | |Instance A | | | | - Cluster manager | | | | - Distributed Lock Manager | | | | - OS Shared Disk Driver | |-------------- | ----------------------------- | | |----------------------------------| | | | | interconnect ------------ | | Shared | |----------------------------------| | Disk | |Machine B | | Subsystem| | | ------------ | |-----------------------------| | | | |Instance B | | | | | - Cluster manager | | | | | - Distributed Lock Manager | | | | | - OS Shared Disk Driver | |--------------- | ----------------------------- | |----------------------------------| Note 1: ------- Local Clustering Definition Local cluster is defined as two or more physical machines (nodes) that share common disk storage and logical IP address. Clustered nodes exchange cluster information over heartbeat link(s). Cluster software collects information and checks the situation on both nodes. On error condition, software will execute a predefined script and switch the clustered services over to a secondary machine. Oracle instance, as one of clustered services, will be switched off together with listener process, and restarted on the secondary (surviving) node. HA Oracle Agent HA Oracle Agent software controls Oracle database activity on Sun Cluster nodes. The agent performs fault checking using two processes on the local node and two process on the remote node by querying V$SYSSTAT table for active sessions. If the database has no active sessions, HA Agent will open a test transaction (connect and execute in serial create, insert, update, drop table commands). Return error codes from HA Agent have been validated against a special action file on location. /etc/opt/SUNWscor/haoracle_config_V1: # Action file for HA-DBMS Oracle fault monitor # State DBMS_er proc_di log_msg timeout int_err new_sta action message --- co * * * * 1 * stop Internal HA-DBMS Oracle error connecting to db on 28 * * * * di none Session killed by DBA, will reconnect * 50 * * * * di takeover O/S error occurred while obtaining an enqueue co 0 * * 1 0 * restart A timeout has occured during connect -- Takeover - cluster software will switch to another node. Stop - cluster will stop DBMS None - no action taken Restart - database restarted locally on the same node HA Oracle Agent requires Oracle configuration files (listener.ora, oratab and tnsnames.ora) on unique predefined location /var/opt/oracle Note 2: ------- You Asked (Jump to Tom's latest followup) If I want to use Oracle Fail Safe and Dataguard do the servers have to be clustered? Right now I have a primary database on one server and a separate server for the logical standby database. I want automatic failover, but it looks like Oracle Fail Safe requires clustered servers. The DATAGUARD manual mentions that you can use ORACLE FAIL SAFE on the windows platform, but the ORACLE FAIL SAFE documentation doesn't say squat about DATAGUARD or how to configure for it. Is there any documentation of this subject that you can refer me to? and we said... Fail Safe is a clustering solution. The two (data guard & failsafe) are complimentary but somewhat orthogonal here. Failsafe is designed to keep the single database up and available -- in a single data center. As long as that room exists -- failsafe keeps the database up. data guard is a disaster recovery solution. It is for when the room the data center is in "goes away" for whatever reason. Data guard wants the machines to be independent (no clusters) of eachother and separated by some geographic distance. Failsafe, like 9i RAC, wants the machines to be tethered together - sitting right next to eachother in a cluster. Failsafe is HA (high availability) Data guard is DR (disaster recovery) Failsafe will give you automated failover. As long as the data center exists, that database is up. With data guard -- you do not WANT automated failover (many *think* they do but you don't). Do you really want your DR solution to kick in due to a WAN failure? No, not really. For DR to take over, you want a human to say "yup, data center burnt to the ground, lets head for the mountains". You do not want the DR site to kick in because "it thinks the primary site is gone" -- you need to tell it "the primary site is gone". In a cluster -- the machines are very aware of eachother and automated failover is "safe" So, data guards reference to failsafe is incidental. That failsafe doesn't talk about data guard is of no real consequence. They are independent feature/functions. Note 3: terms: -------------- Note 4: ------- FAQ RAC: Real Application Clusters General RAC Is it supported to install CRS and RAC as different users. (09-SEP-04) I have changed my spfile with alter system set =.... scope=spfile. The spfile is on ASM storage and the database will not start. (18-APR-04) Is it difficult to transition from Single Instance to RAC? (18-JUL-05) What are the dependencies between OCFS and ASM in Oracle10g ? (05-MAY-05) What software is necessary for RAC? Does it have a separate installation CD to order? (05-MAY-05) Do we have to have Oracle RDBMS on all nodes? (02-APR-04) What kind of HW components do you recommend for the interconnect? (02-APR-04) Is rcp and/or rsh required for normal RAC operation ? (06-NOV-03) Are there any suggested roadmaps for implementing a new RAC installation? (26-NOV-02) What is Cache Fusion and how does this affect applications? (26-NOV-02) Can I use iSCSI storage with my RAC cluster? (13-JUL-05) Can I use RAC in a distributed transaction processing environment? (16-JUN-05) Is it a good idea to add anti-virus software to my RAC cluster? (31-JAN-05) When configuring the NIC cards and switch for a GigE Interconnect should it be set to FULL or Half duplex in RAC? (05-NOV-04) What would you recomend to customer, Oracle clusterware or Vendor Clusterware (I.E. MC Service Guard, HACMP, Sun Cluster, Veritas etc.) with Oracle Database 10g Real Application Clusters? (21-OCT-04) What is Standard Edition RAC? (01-SEP-04) High Availability If I use Services with Oracle Database 10g, do I still need to set up Load Balancing ? (16-JUN-05) Why do we have a Virtual IP (VIP) in 10g? Why does it just return a dead connection when its primary node fails? (12-MAR-04) I am receiving an ORA-29740 error. What should I do? (02-DEC-02) Can RMAN backup Real Application Cluster databases? (26-NOV-02) What is Server-side Transparent Application Failover (TAF) and how do I use it? (07-JUL-05) What is CLB_GOAL and how should I set it? (16-JUN-05) Can I use TAF and FAN/FCF? (16-JUN-05) What clients provide integration with FAN and FCF? (28-APR-05) What are my options for load balancing with RAC? Why do I get an uneven number of connections on my instances? (15-MAR-05) Can our 10g VIP fail over from NIC to NIC as well as from node to node ? (10-DEC-04) Can I use ASM as mechanism to mirror the data in an Extended RAC cluster? (18-OCT-04) What does the Virtual IP service do? I understand it is for failover but do we need a separate network card? Can we use the existing private/public cards? What would happen if we used the public ip? (15-MAR-04) What do the VIP resources do once they detect a node has failed/gone down? Are the VIPs automatically acquired, and published, or is manual intervention required? Are VIPs mandatory? (15-MAR-04) Scalability I am seeing the wait events 'ges remote message', 'gcs remote message', and/or 'gcs for action'. What should I do about these? (02-APR-04) What are the changes in memory requirements from moving from single instance to RAC? (02-DEC-02) What is the Load Balancing Advisory? (16-JUN-05) What is Runtime Connection Load Balancing? (16-JUN-05) How do I enable the load balancing advisory? (16-JUN-05) Manageability How do I stop the GSD? (22-MAR-04) How should I deal with space management? Do I need to set free lists and free list groups? (16-JUN-03) I was installing RAC and my Oracle files did not get copied to the remote node(s). What went wrong? (26-NOV-02) What is the Cluster Verification Utiltiy (cluvfy)? (16-JUN-05) What versions of the database can I use the cluster verification utility (cluvfy) with? (16-JUN-05) What are the implications of using srvctl disable for an instance in my RAC cluster? I want to have it available to start if I need it but at this time to not want to run this extra instance for this database. (31-MAR-05) Platform Specific How many nodes can be had in an HP/Sun/IBM/Compaq/NT/Linux cluster? (21-OCT-04) Is crossover cable supported as an interconnect with 9iRAC/10gRAC on any platform ? (21-FEB-05) Is it possible to run RAC on logical partitions (i.e. LPARs) or virtual separate servers. (18-MAY-04) Can the Oracle Database Configuration Assistant (DBCA) be used to create a database with Veritas DBE / AC 3.5? (10-JAN-03) How do I check RAC certification? (26-NOV-02) Where I can find information about how to setup / install RAC on different platforms ? (08-AUG-02) Is Veritas Storage Foundation 4.0 supported with RAC? (05-OCT-04) Platform Specific -- Linux Is 3rd Party Clusterware supported on Linux such as Veritas or Redhat? (11-MAY-05) Can you have multiple RAC $ORACLE_HOME's on Linux? (19-JUL-05) After installing patchset 9013 and patch_2313680 on Linux, the startup was very slow (20-DEC-04) Is CFS Available for Linux? (20-DEC-04) Where can I find more information about hangcheck-timer module on Linux ? And how do we configure hangcheck-timer module ? (20-DEC-04) Can RAC 10g and 9i RAC be installed and run on the same physical Linux cluster? (20-DEC-04) Is the hangcheck timer still needed with Oracle Database 10g RAC? (20-DEC-04) How to configure bonding on Suse SLES8. (29-NOV-04) How to configure bonding on Suse SLES9. (29-NOV-04) Platform Specific -- Solaris Does RAC run faster with Sun-cluster or Veritas cluster-ware? (these being alternatives with Sun hardware) Is there some clusterware that would make RAC run faster? (20-DEC-04) Platform Specific -- HP-UX Is HMP supported with 10g on all HP platforms ? (20-DEC-04) Platform Specific -- Windows Does the Oracle Cluster File System (OCFS) support network access through NFS or Windows Network Shares? (27-JAN-05) Can I run my 9i RAC and RAC 10g on the same Windows cluster? (01-JUL-05) My customer wants to understand what type of disk caching they can use with their Windows RAC Cluster, the install guide tells them to disable disk caching? (31-MAR-05) Platform Specific -- IBM AIX Do I need HACMP/GPFS to store my OCR/Voting file on a shared device. (20-DEC-04) Platform Specific -- IBM-z/OS (Mainframe) Can I run Oracle RAC 10g on my IBM Mainframe Sysplex environment (z/OS)? (07-JUL-05) Diagnosibility What are the cdmp directories in the background_dump_dest used for? (11-AUG-03) EBusiness Suite with RAC What is the optimal migration path to be used while migrating the E-Business suite to RAC? (08-JUL-05) Is the Oracle E-Business Suite (Oracle Applications) certified against RAC? (04-JUN-03) Can I use TAF with e-Business in a RAC environment? (02-APR-03) How to configure concurrent manager in a RAC environment? (20-SEP-02) Should functional partitioning be used with Oracle Applications? (20-SEP-02) Which e-Business version is prefereable? (20-SEP-02) Can I use Automatic Undo Management with Oracle Applications? (20-SEP-02) Clustered File Systems Can I use OCFS with SE RAC? (01-SEP-04) What are the maximum number of nodes under OCFS on Linux ? (06-NOV-03) Where can I find documentation on OCFS ? (06-NOV-03) What files can I put on Linux OCFS? (14-AUG-03) Is Sun QFS supported with RAC? What about Sun GFS? (19-JAN-05) Is Red Hat GFS(Global File System) is certified by Oracle for use with Real Application Clusters? (22-NOV-04) Oracle Clusterware (CRS) Is it possible to use ASM for the OCR and voting disk? (19-JUL-05) Is it supported to rerun root.sh from the Oracle Clusterware installation ? (05-MAY-05) Is it supported to allow 3rd Party Clusterware to manage Oracle resources (instances, listeners, etc) and turn off Oracle Clusterware management of these? (05-MAY-05) What is the High Availability API? (05-MAY-05) How to move the OCR location ? (24-MAR-04) Does Oracle Clusterware support application vips? (11-JUL-05) Why is the home for Oracle Clusterware not recommended to be subdirectory of the Oracle base directory? (11-JUL-05) Can I use Oracle Clusterware to provide cold failover of my 9i or 10g single instance Oracle Databases? (01-JUL-05) How do I put my application under the control of Oracle Clusterware to achieve higher availability? (16-JUN-05) How do I protect the OCR and Voting in case of media failure? (05-MAY-05) How do I use multiple network interfaces to provide High Availability for my interconnect with Oracle Clusterware? (06-APR-05) How to Restore a Lost Voting Disk used by Oracle Clusterware 10g (02-DEC-04) With Oracle Clusterware 10g, how do you backup the OCR? (02-DEC-04) Does the hostname have to match the public name or can it be anything else? (05-NOV-04) Is it a requirement to have the public interface linked to ETH0 or does it only need to be on a ETH lower than the private interface?: - public on ETH1 - private on ETH2 (05-NOV-04) How do I restore OCR from a backup? On Windows, can I use ocopy? (27-OCT-04) What should the permissions be set to for the voting disk and ocr when doing a RAC Install? (22-OCT-04) Which processes access to OCR ? (22-OCT-04) Can I change the name of my cluster after I have created it when I am using Oracle Database 10g Clusterware? (05-OCT-04) Can I change the public hostname in my Oracle Database 10g Cluster using Oracle Clusterware? (05-OCT-04) During CRS installation, I am asked to define a private node name, and then on the next screen asked to define which interfaces should be used as private and public interfaces. What information is required to answer these questions? (24-MAR-04) Answers I have changed my spfile with alter system set =.... scope=spfile. The spfile is on ASM storage and the database will not start. How to recover: In $ORACLE_HOME/dbs . oraenv sqlplus "/ as sysdba" startup nomount create pfile='recoversp' from spfile / shutdown immediate quit Now edit the newly created pfile to change the parameter to something sensible. Then: sqlplus "/ as sysdba" startup pfile='recoversp' (or whatever you called it in step one). create spfile='+DATA/GASM/spfileGASM.ora' from pfile='recoversp' / N.B.The name of the spfile is in your original init.ora so adjust to suit shutdown immediate startup quit Modified: 18-APR-04 Ref #: ID-5068 -------------------------------------------------------------------------------- Is it supported to install CRS and RAC as different users. Yes, CRS and RAC can be installed as different users. The CRS user and the RAC user must both have "oinstall" as their primary group, and the RAC user should be a member of the OSDBA group. Modified: 09-SEP-04 Ref #: ID-5769 -------------------------------------------------------------------------------- Do we have to have Oracle RDBMS on all nodes? Each node of a cluster will typically have the RDBMS and RAC software loaded on it, but not actual datafiles (these need to be available via shared disk). For example, if you wish to run RAC on 2 nodes of a 4-node cluster, you would need to install it on all nodes, but it would only need to be licensed on the two nodes running the RAC database. Note that using a clustered file system, or NAS storage can provide a configuration that does not necessarily require the Oracle binaries to be installed on all nodes. Modified: 02-APR-04 Ref #: ID-4024 -------------------------------------------------------------------------------- What kind of HW components do you recommend for the interconnect? The general recommendation for the interconnect is to provide the highest bandwith interconnect, together with the lowest latency protocol that is available for a given platform. In practice, Gigabit Ethernet with UDP has proven sufficient in every case it has been implemented, and tends to be the lowest common denominator across platforms. Modified: 02-APR-04 Ref #: ID-4049 -------------------------------------------------------------------------------- Are there any suggested roadmaps for implementing a new RAC installation? Yes, Oracle Support recommends the following best practices roadmap to successfully implement RAC: A Smooth Transition to Real Application Clusters The Purpose of this document is to provide a best practices road map to successfully implement Real Application Clusters. Modified: 26-NOV-02 Ref #: ID-4062 -------------------------------------------------------------------------------- What is Cache Fusion and how does this affect applications? Cache Fusion is a new parallel database architecture for exploiting clustered computers to achieve scalability of all types of applications. Cache Fusion is a shared cache architecture that uses high speed low latency interconnects available today on clustered systems to maintain database cache coherency. Database blocks are shipped across the interconnect to the node where access to the data is needed. This is accomplished transparently to the application and users of the system. Cache Fusion scales to clusters with a large numbers of nodes. For more information about cache fusion see the following links: Additional Information can be found at: Understanding 9i Real Application Clusters Cache Fusion There is also a whitepaper ""Cache Fusion Delivers Scalability"" available at http://otn.oracle.com/products/oracle9i/content.html Cache Fusion in the Oracle Documentation Modified: 26-NOV-02 Ref #: ID-4065 -------------------------------------------------------------------------------- Is it difficult to transition from Single Instance to RAC? If the cluster and the cluster software are not present, these components must be installed and configured. The RAC option must be added using the Oracle Universal Installer, which necessitates the existing DB instance must be shut down. There are no changes necessary on the user data within the database. However, a shortage of freelists and freelist groups can cause contention with header blocks of tables and indexes as multiple instances vie for the same block. This may cause a performance problem and require data partitioning. However, the need for these changes should be rare. Recommendation: apply automatic space segment management to perform these changes automatically. The free space management will replace the freelists and freelist groups and is better. The database requires one Redo thread and one Undo tablespace for each instance, which are easily added with SQL commands or with Enterprise Manager tools. Datafiles will need to be moved to either a clustered file system (CFS) or raw devices so that all nodes can access it. Also, the MAXINSTANCES parameter in the control file must be greater than or equal to number of instances you will start in the cluster. For more detailed information, please see Migrating from single-instance to RAC in the Oracle Documentation With Oracle Database 10g Release 2, $ORACLE_HOME/bin/rconfig tool can be used to convert Single instance database to RAC. This tool takes in a xml input file and convert the Single Instance database whose information is provided in the xml. You can run this tool in "verify only" mode prior to performing actual conversion. This is documented in the RAC admin book and a sample xml can be found $ORACLE_HOME/assistants/rconfig/sampleXMLs/ConvertToRAC.xml. Grid Control 10g Release 2 provides a easy to use wizard to perform this function. Note: Please be aware that you may hit bug 4456047 (shutdown immediate hangs) as you convert the database. The bug is updated with workaround and the w/a should is release noted as well. Modified: 18-JUL-05 Ref #: ID-4101 -------------------------------------------------------------------------------- What are the dependencies between OCFS and ASM in Oracle10g ? In an Oracle Database 10g RAC environment, there is no dependency between Automatic Storage Management (ASM) and Oracle Cluster File System (OCFS). OCFS is not required if you are using Automatic Storage Management (ASM) for database files. You can use OCFS on Windows( Version 2 on Linux ) for files that ASM does not handle - binaries (shared oracle home), trace files, etc. Alternatively, you could place these files on local file systems even though it's not as convenient given the multiple locations. If you do not want to use ASM for your database files, you can still use OCFS for database files in Oracle Database 10g. Please refer to ASM and OCFS Positioning Modified: 05-MAY-05 Ref #: ID-4116 -------------------------------------------------------------------------------- Is rcp and/or rsh required for normal RAC operation ? rcp"" and ""rsh"" are not required for normal RAC operation. However ""rsh"" and ""rcp"" should to be enabled for RAC and patchset installation. In future releases, ssh will be used for these operations. Modified: 06-NOV-03 Ref #: ID-4117 -------------------------------------------------------------------------------- What software is necessary for RAC? Does it have a separate installation CD to order? Real Application Clusters is an option of Oracle Database and therefore part of the Oracle Database CD. With Oracle 9i, RAC is part of Oracle9i Enterprise Edition. If you install 9i EE onto a cluster, and the Oracle Universal Installer (OUI) recognizes the cluster, you will be provided the option of installing RAC. Most UNIX platforms require an OSD installation for the necessary clusterware. For Intel platforms (Linux and Windows), Oracle provides the OSD software within the Oracle9i Enterprise Edition release. With Oracle Database 10g, RAC is an option of EE and available as part of SE. Oracle provides Oracle Clusterware on its own CD included in the database CD pack. Please check the certification matrix (Note 184875.1) or with the appropriate platform vendor for more information. @ Sent by Karin Brandauer Modified: 05-MAY-05 Ref #: ID-4132 -------------------------------------------------------------------------------- What is Standard Edition RAC? With Oracle Database 10g, a customer who has purchased Standard Edition is allowed to use the RAC option within the limitations of Standard Edition(SE). For licensing restrictions you should read the Oracle Database 10g License Doc. At a high level this means that you can have a max of 4 cpus in the cluster, you must use ASM for all database files. Oracle Cluster File System (OCFS) is not supported for use with SE RAC. Modified: 01-SEP-04 Ref #: ID-5750 -------------------------------------------------------------------------------- Can I use iSCSI storage with my RAC cluster? For iSCSI, Oracle has made the statement that, as a block protocol, this technology does not require validation for single instance database. There are many early adopter customers of iSCSI running Oracle9i and Oracle Database 10g. As for RAC, Oracle has chosen to validate the iSCSI technology (not each vendor's targets) for the 10g platforms - this has been completed for Linux, Unix and Windows. For Windows we have tested up to 4 nodes - Any Windows iSCSI products that are supported by the host and storage device are supported by Oracle. No vendor-specific information will be posted on Certify. Modified: 13-JUL-05 Ref #: ID-5788 -------------------------------------------------------------------------------- What would you recomend to customer, Oracle clusterware or Vendor Clusterware (I.E. MC Service Guard, HACMP, Sun Cluster, Veritas etc.) with Oracle Database 10g Real Application Clusters? You will be installing and using Oracle Clusterware whether or not you use the Vendor Clusterware. The question you need to ask is whether the Vendor Clusterware gives you something that Oracle Clusterware does not. Is the RAC database on the same server as the application server? Are there any other processes on the same server as the database that you require Vendor Clusterware to fail over to another server in the cluster if the server it is running on fails? IF this is the case, you may want the vendor clusterware, if not, why spend the extra money when Oracle Clusterware supplies everything you need to for the clustered database included with your RAC license. Modified: 21-OCT-04 Ref #: ID-5968 -------------------------------------------------------------------------------- When configuring the NIC cards and switch for a GigE Interconnect should it be set to FULL or Half duplex in RAC? You've got to use Full Duplex, regardless of RAC or not, but for all network communication. Half Duplex means you can only either send OR receive at the same time. Modified: 05-NOV-04 Ref #: ID-6048 -------------------------------------------------------------------------------- Is it a good idea to add anti-virus software to my RAC cluster? For customers who choose to run anti-virus (AV) software on their database servers, they should be aware that the nature of AV software is that disk IO bandwidth is reduced slightly as most AV software checks disk writes/reads. Also, as the AV software runs, it will use CPU cycles that would normally be consumed by other server processes (e.g your database instance). As such, databases will have faster performance when not using AV software. As some AV software is known to lock the files whilst is scans then it is a good idea to exclude the Oracle Datafiles/controlfiles/logfiles from a regular AV scan Modified: 31-JAN-05 Ref #: ID-6595 -------------------------------------------------------------------------------- Can I use RAC in a distributed transaction processing environment? YES. Best practices is to have all tightly coupled branches of a distributed transaction running on a RAC database must run on the same instance. Between transactions and between services, transactions can be load balanced across all of the database instances. You can use services to manage DTP environments. By defining the DTP property of a service, the service is guaranteed to run on one instance at a time in a RAC database. All global distributed transactions performed through the DTP service are ensured to have their tightly-coupled branches running on a single RAC instance. Modified: 16-JUN-05 Ref #: ID-6864 -------------------------------------------------------------------------------- Why do we have a Virtual IP (VIP) in 10g? Why does it just return a dead connection when its primary node fails? Its all about availability of the application. When a node fails, the VIP associated with it is supposed to be automatically failed over to some other node. When this occurs, two things happen. (1) the new node re-arps the world indicating a new MAC address for the address. For directly connected clients, this usually causes them to see errors on their connections to the old address; (2) Subsequent packets sent to the VIP go to the new node, which will send error RST packets back to the clients. This results in the clients getting errors immediately. This means that when the client issues SQL to the node that is now down, or traverses the address list while connecting, rather than waiting on a very long TCP/IP time-out (~10 minutes), the client receives a TCP reset. In the case of SQL, this is ORA-3113. In the case of connect, the next address in tnsnames is used. Without using VIPs, clients connected to a node that died will often wait a 10 minute TCP timeout period before getting an error. As a result, you don't really have a good HA solution without using VIPs. Modified: 12-MAR-04 Ref #: ID-4609 -------------------------------------------------------------------------------- If I use Services with Oracle Database 10g, do I still need to set up Load Balancing ? Yes, Services allow you granular definition of workload and the DBA can dynamically define which instances provide the service. Connection Load Balancing still needs to be set up to allow the user connections to be balanced across all instances providing a service. Modified: 16-JUN-05 Ref #: ID-6731 -------------------------------------------------------------------------------- Can RMAN backup Real Application Cluster databases? Absolutely. RMAN can be configured to connect to all nodes within the cluster to parallelize the backup of the database files and archive logs. If files need to be restored, using set AUTOLOCATE ON alerts RMAN to search for backed up files and archive logs on all nodes. RAC with RMAN in the Oracle Documentation Modified: 26-NOV-02 Ref #: ID-4035 -------------------------------------------------------------------------------- I am receiving an ORA-29740 error. What should I do? This error can occur when problems are detected on the cluster: Error: ORA-29740 (ORA-29740) Text: evicted by member %s, group incarnation %s --------------------------------------------------------------------------- Cause: This member was evicted from the group by another member of the cluster database for one of several reasons, which may include a communications error in the cluster, failure to issue a heartbeat to the control file, etc. Action: Check the trace files of other active instances in the cluster group for indications of errors that caused a reconfiguration. For more information on troubleshooting this error, see the following Metalink note: Note 219361.1 Troubleshooting ORA-29740 in a RAC Environment Modified: 02-DEC-02 Ref #: ID-4093 -------------------------------------------------------------------------------- What does the Virtual IP service do? I understand it is for failover but do we need a separate network card? Can we use the existing private/public cards? What would happen if we used the public ip? The 10g Virtual IP Address (VIP) exists on every RAC node for public network communication. All client communication should use the VIPs in their TNS connection descriptions. The TNS ADDRESS_LIST entry should direct clienst to VIPs rather than using hostnames. During normal runtime, the behaviour is the same as hostnames, however when the node goes down or is shutdown the VIP is hosted elsewhere on the cluster, and does not accept connection requests. This results in a silent TCP/IP error and the client fails immediately to the next TNS address. If the network interface fails within the node, the VIP can be configured to use alternate interfaces in the same node. The VIP must use the public interface cards. There is no requirement to purchase additional public interface cards (unless you want to take advantage of within-node card failover.) Modified: 15-MAR-04 Ref #: ID-4636 -------------------------------------------------------------------------------- What do the VIP resources do once they detect a node has failed/gone down? Are the VIPs automatically acquired, and published, or is manual intervention required? Are VIPs mandatory? When a node fails, the VIP associated with the failed node is automatically failed over to one of the other nodes in the cluster. When this occurs, two things happen: The new node re-arps the world indicating a new MAC address for this IP address. For directly connected clients, this usually causes them to see errors on their connections to the old address; Subsequent packets sent to the VIP go to the new node, which will send error RST packets back to the clients. This results in the clients getting errors immediately. In the case of existing SQL conenctions, errors will typically be in the form of ORA-3113 errors, while a new connection using an address list will select the next entry in the list. Without using VIPs, clients connected to a node that died will often wait for a TCP/IP timeout period before getting an error. This can be as long as 10 minutes or more. As a result, you don't really have a good HA solution without using VIPs. Modified: 15-MAR-04 Ref #: ID-4638 -------------------------------------------------------------------------------- What are my options for load balancing with RAC? Why do I get an uneven number of connections on my instances? All the types of load balancing available currently (9i-10g) occur at connect time. This means that it is very important how one balances connections and what these connections do on a long term basis. Since establishing connections can be very expensive for your application, it is good programming practice to connect once and stay connected. This means one needs to be careful as to what option one uses. Oracle Net Services provides load balancing or you can use external methods such as hardware based or clusterware solutions. The following options exist: Random Either client side load balancing or hardware based methods will randomize the connections to the instances. On the negative side this method is unaware of load on the connections or even if they are up meaning they might cause waits on TCP/IP timeouts. Load Based Server side load balancing (by the listener) redirects connections by default depending on the RunQ length of each of the instances. This is great for short lived connections. Terrible for persistent connections or login storms. Do not use this method for connections from connection pools or applicaton servers Session Based Server side load balancing can also be used to balance the number of connections to each instance. Session count balancing is method used when you set a listener parameter, prefer_least_loaded_node_listener-name=off. Note listener name is the actual name of the listener which is different on each node in your cluster and by default is listener_nodename. Session based load balancing takes into account the number of sessions connected to each node and then distributes ne connections to balance the number of sessions across the different nodes. Modified: 15-MAR-05 Ref #: ID-4940 -------------------------------------------------------------------------------- Can I use ASM as mechanism to mirror the data in an Extended RAC cluster? Yes, but it cannot replicate everything that needs replication. ASM works well to replicate any object you can put in ASM. But you cannot put the OCR or Voting Disk in ASM. In 10gR1 they can either be mirrored using a different mechanism (which could then be used instead of ASM) or the OCR needs to be restored from backup and the Voting Disk can be recreated. In the future we are looking at providing Oracle redundancy for both. Modified: 18-OCT-04 Ref #: ID-5948 -------------------------------------------------------------------------------- Can our 10g VIP fail over from NIC to NIC as well as from node to node ? Yes the 10g VIP implementation is capable from failing over within a node from NIC to NIC and back if the failed NIC is back online again, and also we fail over between nodes. The NIC to NIC failover is fully redundant if redundant switches are installed. Modified: 10-DEC-04 Ref #: ID-6348 -------------------------------------------------------------------------------- What clients provide integration with FAN and FCF? With Oracle Database 10g Release 1, JDBC clients (both thick and thin driver) are integrated with FAN by providing FCF. With Oracle Database 10g Release 2, we have added ODP.NET and OCI. Other applications can integrate with FAN by using the API to subscribe to the FAN events. Modified: 28-APR-05 Ref #: ID-6735 -------------------------------------------------------------------------------- What is CLB_GOAL and how should I set it? CLB_GOAL is the connection load balancing goal for a service. There are 2 options, CLB_GOAL_SHORT and CLB_GOAL_LONG (default). Long is for applications that have long-lived connections. This is typical for connection pools and SQL*Forms sessions. Long is the default connection load balancing goal. Short is for applications that have short-lived connections. The GOAL for a service can be set with EM or DBMS_SERVICE. Note: You must still configure load balancing with Oracle Net Services Modified: 16-JUN-05 Ref #: ID-6854 -------------------------------------------------------------------------------- Can I use TAF and FAN/FCF? With Oracle Database 10g Release 1, NO. With Oracle Database 10g Release 2, the answer is YES for OCI and ODP.NET, it is recommended. For JDBC, you should not use TAF and FCF even with the Thick JDBC driver. Modified: 16-JUN-05 Ref #: ID-6866 -------------------------------------------------------------------------------- What is Server-side Transparent Application Failover (TAF) and how do I use it? Oracle Database 10g Release 2, introduces server-side TAF when using services. After you create a service, you can use the dbms_service.modify_service pl/sql procedure to define the TAF policy for the service. Only the basic method is supported. Note this is different than the TAF policy (traditional client TAF) that is supported by srvctl and EM Services page. If your service has a server side TAF policy defined, then you do not have to encode TAF on the client connection string. If the instance where a client is connected, fails, then the connection will be failed over to another instance in the cluster that is supporting the service. All restrictions of TAF still apply. NOTE: both the client and server must be 10.2 and aq_ha_notifications must be set to true for the service. Sample code to modify service: execute dbms_service.modify_service (service_name => 'gl.us.oracle.com' - , aq_ha_notifications => true - , failover_method => dbms_service.failover_method_basic - , failover_type => dbms_service.failover_type_select - , failover_retries => 180 - , failover_delay => 5 - , clb_goal => dbms_service.clb_goal_long); Modified: 07-JUL-05 Ref #: ID-6912 -------------------------------------------------------------------------------- I am seeing the wait events 'ges remote message', 'gcs remote message', and/or 'gcs for action'. What should I do about these? These are idle wait events and can be safetly ignored. The 'ges remote message' might show up in a 9.0.1 statspack report as one of the top wait events. To have this wait event not show up you can add this event to the PERFSTAT.STATS$IDLE_EVENT table so that it is not listed in Statspack reports. Modified: 02-APR-04 Ref #: ID-4092 -------------------------------------------------------------------------------- What are the changes in memory requirements from moving from single instance to RAC? If you are keeping the workload requirements per instance the same, then about 10% more buffer cache and 15% more shared pool is needed. The additional memory requirement is due to data structures for coherency management. The values are heuristic and are mostly upper bounds. Actual esource usage can be monitored by querying current and maximum columns for the gcs resource/locks and ges resource/locks entries in V$RESOURCE_LIMIT. But in general, please take into consideration that memory requirements per instance are reduced when the same user population is distributed over multiple nodes. In this case: Assuming the same user population N number of nodes M buffer cache for a single system then (M / N) + ((M / N )*0.10) [ + extra memory to compensate for failed-over users ] Thus for example with a M=2G & N=2 & no extra memory for failed-over users =( 2G / 2 ) + (( 2G / 2 )) *0.10 =1G + 100M Modified: 02-DEC-02 Ref #: ID-4030 -------------------------------------------------------------------------------- What is the Load Balancing Advisory? To assist in the balancing of application workload across designated resources, Oracle Database 10g Release 2 provides the Load Balancing Advisory. This Advisory monitors the current workload activity across the cluster and for each instance where a service is active; it provides a percentage value of how much of the total workload should be sent to this instance as well as service quality flag. The feedback is provided as an entry in the Automatic Workload Repository and a FAN event is published. Modified: 16-JUN-05 Ref #: ID-6858 -------------------------------------------------------------------------------- What is Runtime Connection Load Balancing? Runtime connection load balancing enables the connection pool to route incoming work requests to the available database connection that will provide it with the best service. This will provide the best service times globally, and routing responds fast to changing conditions in the system. Oracle has implemented runtime connection load balancing with ODP.NET and JDBC connection pools. Runtime Connection Load Balancing is tightly integrated with the automatic workload balancing features introduced with Oracle Database 10g I.E. Services, Automatic Workload Repository, and the new Load Balancing Advisory. Modified: 16-JUN-05 Ref #: ID-6860 -------------------------------------------------------------------------------- How do I enable the load balancing advisory? The load balancing advisory requires the use of services and Oracle Net connection load balancing. To enable it, on the server: set a goal (service_time or throughput, for ODP.NET enable AQ_HA_NOTIFICATIONS=>true, and set CLB_GOAL ) on your service. For client, you must be using the connection pool. For JDBC, enable the datasource parameter FastConnectionFailoverEnabled. For ODP.NET enable the datasource parameter Load Balancing=true. Modified: 16-JUN-05 Ref #: ID-6862 -------------------------------------------------------------------------------- How do I stop the GSD? If you are on 9.0 on Unix you would issue: $ ps -ef | grep jre $ kill -9 Stop the OracleGSDService on Windows. Note: Make sure that this is the process in use by GSD If you are on 9.2 you would issue: $ gsdctl stop Modified: 22-MAR-04 Ref #: ID-4091 -------------------------------------------------------------------------------- How should I deal with space management? Do I need to set free lists and free list groups? Manually setting free list groups is a complexity that is no longer required. We recommend using Automatic Segment Space Management rather than trying to manage space manually. Unless you are migrating from an earlier database version with OPS and have already built and tuned the necessary structures, Automatic Segment Space Management is the preferred approach. Automatic Segment Space Management is NOT the default, you need to set it. For more information see: Automatic Space Segment Management in RAC Environments Modified: 16-JUN-03 Ref #: ID-4074 -------------------------------------------------------------------------------- I was installing RAC and my Oracle files did not get copied to the remote node(s). What went wrong? First make sure the cluster is running and is available on all nodes. You should be able to see all nodes when running an 'lsnodes -v' command. If lsnodes shows that all members of the cluster are available, then you may have an rcp/rsh problem on Unix or shares have not been configured on Windows. You can test rcp/rsh on Unix by issuing the following from each node: [node1]/tmp> touch test.tst [node1]/tmp> rcp test.tst node2:/tmp [node2]/tmp> touch test.tst [node2]/tmp> rcp test.tst node1:/tmp On Windows, ensure that each node has administrative access to all these directories within the Windows environment by running the following at the command prompt: NET USE \\host_name\C$ Clustercheck.exe also checks for this. More information can be found in the Step-by-Step RAC notes available on Metalink. To find these search Metalink for 'Step-by-Step Installation of RAC'. Modified: 26-NOV-02 Ref #: ID-4094 -------------------------------------------------------------------------------- What are the implications of using srvctl disable for an instance in my RAC cluster? I want to have it available to start if I need it but at this time to not want to run this extra instance for this database. During node reboot, any disabled resources will not be started by the Clusterware, therefore this instance will not be restarted. It is recommended that you leave the vip, ons,gsd enabled in that node. For example, VIP address for this node is present in address list of database services, so a client connecting to these services will still reach some other database instance providing that service via listener redirection. J ust be aware that by disabling an Instance on a node, all that means is that the instance itself is not starting. However, if the database was originally created with 3 instances, that means there are 3 threads of redo. So, while the instance itself is disabled, the redo thread is still enabled, and will occasionally cause log switches. The archived logs for this 'disabled' instance would still be needed in any potential database recovery scenario. So, if you are going to disable the instance through srvctl, you may also want to consider disabling the redo thread for that instance. srvctl disable instance -d orcl -i orcl2 SQL> alter database disable public thread 2; Do the reverse to enable the instance. SQL> alter database enable public thread 2; srvctl enable instance -d orcl -i orcl2 Modified: 31-MAR-05 Ref #: ID-6672 -------------------------------------------------------------------------------- What is the Cluster Verification Utiltiy (cluvfy)? The Cluster Verification Utility (CVU) is a validation tool that you can use to check all the important components that need to be verified at different stages of deployment in a RAC environment. The wide domain of deployment of CVU ranges from initial hardware setup through fully operational cluster for RAC deployment and covers all the intermediate stages of installation and configuration of various components. Cluvfy does not take any corrective action following the failure of a verification task, does not enter into areas of performance tuning or monitoring, does not perform any cluster or RAC operation, and does not attempt to verify the internals of cluster database or cluster elements. Modified: 16-JUN-05 Ref #: ID-6850 -------------------------------------------------------------------------------- What versions of the database can I use the cluster verification utility (cluvfy) with? The cluster verification utility is release with Oracle Database 10g Release 2 but can also be used with Oracle Database 10g Release 1. Modified: 16-JUN-05 Ref #: ID-6852 -------------------------------------------------------------------------------- How many nodes can be had in an HP/Sun/IBM/Compaq/NT/Linux cluster? The number of nodes supported is not limited by Oracle, but more generally by the clustering software/hardware in question. When using Solely Oracle Clusterware: 63 nodes (9i or 10gR1) When using a third party clusterware: Sun: 8 HP UX: 16 HP Tru64: 8 IBM AIX: * 8 nodes for Physical Shared (CLVM) SSA disk * 16 nodes for Physical Shared (CLVM) non-SSA disk * 128 nodes for Virtual Shared Disk (VSD) * 128 nodes for GPFS * Subject to storage subsystem limitations Veritas: 8-16 nodes (check w/ Veritas) Modified: 21-OCT-04 Ref #: ID-4047 -------------------------------------------------------------------------------- Where I can find information about how to setup / install RAC on different platforms ? There is a roadmap for implementing Real Application Clusters' available at: A Smooth Transition to Real Application Clusters There are also Step-by-Step notes available for each platform available on the Metalink 'Top Tech Docs' for RAC: High Availability - Real Application Clusters Library Page Index Additional information can be found on OTN: http://technet.oracle.com/products/oracle9i/content.html --> 'Oracle Real Application Clusters' Modified: 08-AUG-02 Ref #: ID-4067 -------------------------------------------------------------------------------- Is it possible to run RAC on logical partitions (i.e. LPARs) or virtual separate servers. Yes, it is possible. The E10K and other high end servers can be partitioned into domains of smaller sizes, each domain with its own CPU(s) and operating system. Each domain is effectively a virtual server. RAC can be run on cluster comprises of domains. The benefits of using this is similar to a regular cluster, any domain failure will have little effect on other domains. Besides, the management of the cluster may be easier since there is only one physical server. Note however, since one E10K is still just one server. There are single points of failures. Any failures, such as back plane failure, that crumble the entire server will shutdown the virtual cluster. That is the tradeoff users have to make in how best to build a cluster database. Modified: 18-MAY-04 Ref #: ID-4075 -------------------------------------------------------------------------------- How do I check RAC certification? See the following Metalink note: Note 184875.1 How To Check The Certification Matrix for Real Application Clusters Please note that certifications for Real Application Clusters are performed against the Operating System and Clusterware versions. The corresponding system hardware is offered by System vendors and specialized Technology vendors. Some system vendors offer pre-installed, pre-configured RAC clusters. These are included below under the corresponding OS platform selection within the certification matrix. Modified: 26-NOV-02 Ref #: ID-4095 -------------------------------------------------------------------------------- Can the Oracle Database Configuration Assistant (DBCA) be used to create a database with Veritas DBE / AC 3.5? DBCA can be used to create databases on raw devices in 9i RAC Release 1 and 9i Release 2. Standard database creation scripts using SQL commands will work with file system and raw. DBCA cannot be used to create databases on file systems on Oracle 9i Release 1. The user can choose to set up a database on raw devices, and have DBCA output a script. The script can then be modified to use cluster file systems instead. With Oracle 9i RAC Release 2 (Oracle 9.2), DBCA can be used to create databases on a cluster filesystem. If the ORACLE_HOME is stored on the cluster filesystem, the tool will work directly. If ORACLE_HOME is on local drives on each system, and the customer wishes to place database files onto a cluster file system, they must invoke DBCA as follows: dbca -datafileDestination /oradata where /oradata is on the CFS filesystem. See 9iR2 README and bug 2300874 for more info. Modified: 10-JAN-03 Ref #: ID-4124 -------------------------------------------------------------------------------- Is crossover cable supported as an interconnect with 9iRAC/10gRAC on any platform ? NO. CROSS OVER CABLES ARE NOT SUPPORTED. The requirement is to use a switch: Detailed Reasons: 1) cross-cabling limits the expansion of RAC to two nodes 2) cross-cabling is unstable: a) Some NIC cards do not work properly with it. b) Instability. We have seen different problems e.g.. ORA-29740 at configurations using crossover cable, and other errors. Due to the benefits and stability provided by a switch, and their afforability, this is the only supported configuration. Please see certify.us.oracle.com as well. (content consolidated from that of Massimo Castelli, Roland Knapp and others) Modified: 21-FEB-05 Ref #: ID-4150 -------------------------------------------------------------------------------- Is Veritas Storage Foundation 4.0 supported with RAC? Veritas Storage Foundation 4.0 is certified on AIX, Solaris and HPUX for 9i RAC and Oracle Database 10g RAC. Veritas is production also on Linux, but it is not certified by Oracle. If customers choose Veritas on Linux, Oracle will support the Oracle products in the stack, but they do not qualify for Unbreakable Linux support. Modified: 05-OCT-04 Ref #: ID-5888 -------------------------------------------------------------------------------- Is 3rd Party Clusterware supported on Linux such as Veritas or Redhat? No, Oracle RAC 10g does not support 3rd Party clusterware on Linux. This means that if a cluster file system requires a 3rd party clusterware, the cluster file system is not supported. Modified: 11-MAY-05 Ref #: ID-6743 -------------------------------------------------------------------------------- Can you have multiple RAC $ORACLE_HOME's on Linux? No, there should be only one Oracle Cluster Manager (ORACM) running on each node. All RAC databases should run out of the $ORACLE_HOME that ORACM is installed in. Modified: 19-JUL-05 Ref #: ID-6931 -------------------------------------------------------------------------------- After installing patchset 9013 and patch_2313680 on Linux, the startup was very slow Please carefully read the following new information about configuring Oracle Cluster Management on Linux, provided as part of the patch README: Three parameters affect the startup time: soft_margin (defined at watchdog module load) -m (watchdogd startup option) WatchdogMarginWait (defined in nmcfg.ora). WatchdogMarginWait is calculated using the formula: WatchdogMarginWait = soft_margin(msec) + -m + 5000(msec). [5000(msec) is hardcoded] Note that the soft_margin is measured in seconds, -m and WatchMarginWait are measured in milliseconds. Based on benchmarking, it is recommended to set soft_margin between 10 and 20 seconds. Use the same value for -m (converted to milliseconds) as used for soft_margin. Here is an example: soft_margin=10 -m=10000 WatchdogMarginWait = 10000+10000+5000=25000 If CPU utilization in your system is high and you experience unexpected node reboots, check the wdd.log file. If there are any 'ping came too late' messages, increase the value of the above parameters. Modified: 20-DEC-04 Ref #: ID-4069 -------------------------------------------------------------------------------- Is CFS Available for Linux? Yes, OCFS (Oracle Cluster Filesystem) is now available for Linux. The following Metalink note has information for obtaining the latest version of OCFS: Note 238278.1 - How to find the current OCFS version for Linux Modified: 20-DEC-04 Ref #: ID-4089 -------------------------------------------------------------------------------- Where can I find more information about hangcheck-timer module on Linux ? And how do we configure hangcheck-timer module ? In releases 9.2.0.2.0 and later, Oracle recommends using a new I/O fencing model -- HangCheck-Timer module. Hangcheck-Timer module monitors the Linux kernel for long operating system hangs that could affect the reliability of a RAC node. You can configure hangcheck-timer module using 3 parameters -- hangcheck_tick, hangcheck_margin and MissCount. For more details, please review Note :: 259487.1 Modified: 20-DEC-04 Ref #: ID-4179 -------------------------------------------------------------------------------- Can RAC 10g and 9i RAC be installed and run on the same physical Linux cluster? Yes - CRS / CSS and oracm can coexist. Modified: 20-DEC-04 Ref #: ID-4408 -------------------------------------------------------------------------------- Is the hangcheck timer still needed with Oracle Database 10g RAC? YES! The hangcheck-timer module monitors the Linux kernel for extended operating system hangs that could affect the reliability of the RAC node ( I/O fencing) and cause database corruption. To verify the hangcheck-timer module is running on every node: as root user: /sbin/lsmod | grep hangcheck If the hangcheck-timer module is not listed enter the following command as the root user: /sbin/insmod hangcheck-timer hangcheck_tick=30 hangcheck_margin=180 To ensure the module is loaded every time the system reboots, verify that the local system startup file (/etc/rc.d/rc.local) contains the command above. For additional information please review the Oracle RAC Install and Configuration Guide (5-41). Modified: 20-DEC-04 Ref #: ID-6208 -------------------------------------------------------------------------------- How to configure bonding on Suse SLES8. Please see note:291958.1 Modified: 29-NOV-04 Ref #: ID-6288 -------------------------------------------------------------------------------- How to configure bonding on Suse SLES9. Please see note:291962.1 Modified: 29-NOV-04 Ref #: ID-6290 -------------------------------------------------------------------------------- Does RAC run faster with Sun-cluster or Veritas cluster-ware? (these being alternatives with Sun hardware) Is there some clusterware that would make RAC run faster? RAC scalability and performance are independent of the clusterware. However, we recommend that the customer uses a very fast memory based interconnect if one wants to optimize the performance. For Example, Sun can use FireLink, a very fast proprietary interconnect which is more optimal for RAC, while Veritas is limited to using Gigabit Ethernet. Starting with 10g there will be an alternative to SunCluster and Veritas Cluster than this is Oracle CRS/CSS. Modified: 20-DEC-04 Ref #: ID-4088 -------------------------------------------------------------------------------- Is HMP supported with 10g on all HP platforms ? - 10g RAC + HMP + PA-RISC = yes - 10g RAC + HMP + Itanium, "Oracle has no plans and will likely never support RAC over HMP on IPF." - 10g RAC + UDP + Itanium = yes (even over Hyperfabric) "Oracle recommends that HMP not be used. UDP is the recommended interconnect protocol across all platforms." Modified: 20-DEC-04 Ref #: ID-5488 -------------------------------------------------------------------------------- Does the Oracle Cluster File System (OCFS) support network access through NFS or Windows Network Shares? No, in the current release the Oracle Cluster File System (OCFS) is not supported for use by network access approaches like NFS or Windows Network Shares. Modified: 27-JAN-05 Ref #: ID-4122 -------------------------------------------------------------------------------- My customer wants to understand what type of disk caching they can use with their Windows RAC Cluster, the install guide tells them to disable disk caching? If the write cache identified is local to the node then that is bad for RAC. If the cache is visible to all nodes as a 'single cache', typically in the storage array, and is also 'battery backed' then that is OK. Modified: 31-MAR-05 Ref #: ID-6670 -------------------------------------------------------------------------------- Can I run my 9i RAC and RAC 10g on the same Windows cluster? Yes but the 9i RAC database must have the 9i Cluster Manager and you must run Oracle Clusterware for the Oracle Database 10g. 9i Cluster Manager can coexsist with Oracle Clusterware 10g. Modified: 01-JUL-05 Ref #: ID-6889 -------------------------------------------------------------------------------- Do I need HACMP/GPFS to store my OCR/Voting file on a shared device. The prerequisites doc for AIX clearly says: "If you are not using HACMP, you must use a GPFS file system to store the Oracle CRS files" ==> this is a documentation bug and this will be fixed with 10.1.0.3 ----- On AIX it is important to put the reserve_lock=no/reserve_policy =no_reserve in order to allow AIX to access the devices from more than one node simultaneously. Use the /dev/rhdisk devices (character special) for the crs and voting disk and change the attribute with the command chdev -l hdiskn -a reserve_lock=no (for ESS, EMC, HDS, CLARiiON, and MPIO-capable devices you have to do an chdev -l hdiskn -a reserve_policy=no_reserve) Modified: 20-DEC-04 Ref #: ID-5288 -------------------------------------------------------------------------------- Can I run Oracle RAC 10g on my IBM Mainframe Sysplex environment (z/OS)? YES! There is no separate documentation for RAC on z/OS. What you would call "clusterware" is built in to the OS and the native file systems are global. IBM z/OS documentation explains how to set up a Sysplex Cluster; once the customer has done that it is trivial to set up a RAC database. The few steps involved are covered in in Chapter 14 of the Oracle for z/OS System Admin Guide, which you can read here. There is also an Install Guide for Oracle on z/OS ( here) but I don't think there are any RAC-specific steps in the installation. By the way, RAC on z/OS does not use Oracle's clusterware (CSS/CRS/OCR). Modified: 07-JUL-05 Ref #: ID-6910 -------------------------------------------------------------------------------- What are the cdmp directories in the background_dump_dest used for? These directories are produced by the diagnosibility daemon process (DIAG). DIAG is a process related to RAC which as one of its tasks, performs cash dumping. The DIAG process dumps out tracing to file when it discovers the death of an essential process (foreground or background) in the local instance. A dump directory named something like cdmp_ is created in the bdump or background_dump_dest directory, and all the trace dump files DIAG creates are placed in this directory. Modified: 11-AUG-03 Ref #: ID-4152 -------------------------------------------------------------------------------- Is the Oracle E-Business Suite (Oracle Applications) certified against RAC? Yes. (There is no seperate certification required for RAC.) "" Modified: 04-JUN-03 Ref #: ID-4029 -------------------------------------------------------------------------------- What is the optimal migration path to be used while migrating the E-Business suite to RAC? Following is the recommended and most optimal path to migrate you E-Business suite to RAC environment: 1. Migrate the existing application to new hardware. (If applicable). 2. Use Clustered File System for all data base files or migrate all database files to raw devices. (Use dd for Unix or ocopy for NT) 3. Install/upgrade to the latest available e-Business suite. 4. Upgrade database to Oracle9i (Refer document 216550.1 on Metalink) 5. In step 4, install RAC option while installing Oracle9i and use Installer to perform install for all the nodes. 6. Clone Oracle Application code tree. Reference Documents: Oracle E-Business Suite Release 11i with 9i RAC: Installation and Configuration : Metalink Note# 279956.1 E-Business Suite 11i on RAC : Configuring Database Load balancing & Failover: Metalink Note# 294652.1 Oracle E-Business Suite 11i and Database - FAQ : Metalink# 285267.1 Modified: 08-JUL-05 Ref #: ID-4107 -------------------------------------------------------------------------------- How to configure concurrent manager in a RAC environment? Large clients commonly put the concurrent manager on a separate server now (in the middle tier) to reduce the load on the database server. The concurrent manager programs can be tied to a specific middle tier (e.g., you can have CMs running on more than one middle tier box). It is advisable to use specilize CM. CM middle tiers are set up to point to the appropriate database instance based on product module being used. Modified: 20-SEP-02 Ref #: ID-4108 -------------------------------------------------------------------------------- Should functional partitioning be used with Oracle Applications? We do not recommend functional partitioning unless throughput on your server architecture demands it. Cache fusion has been optimized to scale well with non-partitioned workload. If your processing requirements are extreme and your testing proves you must partition your workload in order to reduce internode communications, you can use Profile Options to designate that sessions for certain applications Responsibilities are created on a specific middle tier server. That middle tier server would then be configured to connect to a specific database instance. To determine the correct partitioning for your installation you would need to consider several factors like number of concurrent users, batch users, modules used, workload characteristics etc. Modified: 20-SEP-02 Ref #: ID-4109 -------------------------------------------------------------------------------- Which e-Business version is prefereable? Versions 11.5.5 onwards are certified with Oracle9i and hence with Oracle9i RAC. However we recommend the latest available version. Modified: 20-SEP-02 Ref #: ID-4110 -------------------------------------------------------------------------------- Can I use Automatic Undo Management with Oracle Applications? Yes. In a RAC environment we highly recommend it. Modified: 20-SEP-02 Ref #: ID-4111 -------------------------------------------------------------------------------- Can I use TAF with e-Business in a RAC environment? TAF itself does not work with e-Business suite due to Forms/TAF limitations, but you can configure the tns failover clause. On instance failure, when the user logs back into the system, their session will be directed to a surviving instance, and the user will be taken to the navigator tab. Their committed work will be available; any uncommitted work must be re-started. We also recommend you configure the forms error URL to identify a fallback middle tier server for Forms processes, if no router is available to accomplish switching across servers. Modified: 02-APR-03 Ref #: ID-4112 -------------------------------------------------------------------------------- Can I use OCFS with SE RAC? It is not supported to use OCFS with Standard Edition RAC. All database files must use ASM (redo logs, recovery area, datafiles, control files etc). We recommend that the binaries and trace files (non-ASM supported files) to be replicated on all nodes. This is done automatically by install. Modified: 01-SEP-04 Ref #: ID-5748 -------------------------------------------------------------------------------- What are the maximum number of nodes under OCFS on Linux ? Oracle 9iRAC on Linux, using OCFS for datafiles, can scale to a maximum of 32 nodes. Modified: 06-NOV-03 Ref #: ID-4118 -------------------------------------------------------------------------------- Where can I find documentation on OCFS ? For Main Page >>> http://oss.oracle.com/projects/ocfs/ For User Manual >>> http://oss.oracle.com/projects/ocfs/documentation/ For OCFS Files >>> http://oss.oracle.com/projects/ocfs/files/supported/ Modified: 06-NOV-03 Ref #: ID-4119 -------------------------------------------------------------------------------- What files can I put on Linux OCFS? For optimal performance, you should only put the following files on Linux OCFS: - Datafiles - Control Files - Redo Logs - Archive Logs - Shared Configuration File (OCR) - Quorum / Voting File - SPFILE Modified: 14-AUG-03 Ref #: ID-4156 -------------------------------------------------------------------------------- Is Sun QFS supported with RAC? What about Sun GFS? Sun QFS is supported with Oracle 9i RAC. Sun is planning to certify QFS with Oracle Database 10g and RAC but as of November 15,2004, this certification is "planned". For 9i, Software Stack details: For SVM you need Solaris 9 9/04 (Solaris 9 update 7),SVM Patch 116669-03(this is required SUN patch), Sun Cluster 3.1 Update 3, Oracle 9.2.0.5 + Oracle patch 3366258 For SharedQFS you need Solaris 9 04/03 and above or Solaris 8 02/02 and above, QFS 4.2, Sun Cluster 3.1 Update 2 or above, Oracle 9.2.0.5 + Oracle patch 3566420 Differently, Sun GFS (Global File System) is only supported for Oracle binary and archive logs only, but NOT for database files. Modified: 19-JAN-05 Ref #: ID-6128 -------------------------------------------------------------------------------- Is Red Hat GFS(Global File System) is certified by Oracle for use with Real Application Clusters? Sistina Cluster Filesystem is not part of the standard RedHat kernel and therefore is not certified under the unbreakable Linux but falls under a kernel extension. This however, does not mean that Oracle RAC is not certified with it. As a fact, Oracle RAC does not certify against a filesystem per se, but certifies against an operating system. If, as is the case with Sistina filesystem, the filesystem is certified with the operating system, this only means that the combination does not fall under the unbreakable Linux combination and Oracle does not provide direct support and fix the filesystem in case of an error. Customer will have to contact the filesystem provider for support. Modified: 22-NOV-04 Ref #: ID-6228 -------------------------------------------------------------------------------- How to move the OCR location ? - stop the CRS stack on all nodes using "init.crs stop" - Edit /var/opt/oracle/ocr.loc on all nodes and set up ocrconfig_loc=new OCR device - Restore from one of the automatic physical backups using ocrconfig -restore. - Run ocrcheck to verify. - reboot to restart the CRS stack. - additional information can be found at http://st-doc.us.oracle.com/10/101/rac.101/b10765/storage.htm#i1016535 Modified: 24-MAR-04 Ref #: ID-4728 -------------------------------------------------------------------------------- Is it supported to rerun root.sh from the Oracle Clusterware installation ? Rerunning root.sh after the initial install is expressly discouraged and unsupported. We strongly recommend not doing it. Modified: 05-MAY-05 Ref #: ID-4730 -------------------------------------------------------------------------------- Is it supported to allow 3rd Party Clusterware to manage Oracle resources (instances, listeners, etc) and turn off Oracle Clusterware management of these? In 10g we do not support using 3rd Party Clusterware for failover and restart of Oracle resources. Oracle Clusterware resources should not be disabled. Modified: 05-MAY-05 Ref #: ID-6528 -------------------------------------------------------------------------------- What is the High Availability API? An application-programming interface to allow processes to be put under the High Availability infrastructure that is part of the Oracle Clusterware distributed with Oracle Database 10g. A user written script defines how Oracle Clusterware should start, stop and relocate the process when the cluster node status changes. This extends the high availability services of the cluster to any application running in the cluster. Oracle Database 10g Real Application Clusters (RAC) databases and associated Oracle processes (E.G. listener) are automatically managed by the clusterware. Modified: 05-MAY-05 Ref #: ID-6741 -------------------------------------------------------------------------------- Is it possible to use ASM for the OCR and voting disk? No, the OCR and voting disk must be on raw or CFS (cluster filesystem). Modified: 19-JUL-05 Ref #: ID-6929 -------------------------------------------------------------------------------- During CRS installation, I am asked to define a private node name, and then on the next screen asked to define which interfaces should be used as private and public interfaces. What information is required to answer these questions? The private names on the first screen determine which private interconnect will be used by CSS. Provide exactly one name that maps to a private IP address, or just the IP address itself. If a logical name is used, then the IP address this maps to can be changed subsequently, but if you IP address is specified CSS will always use that IP address. CSS cannot use multiple private interconnects for its communication hence only one name or IP address can be specified. The private interconnect enforcement page determines which private interconnect will be used by the RAC instances. It's equivalent to setting the CLUSTER_INTERCONNECTS init.ora parameter, but is more convenient because it is a cluster-wide setting that does not have to be adjusted every time you add nodes or instances. RAC will use all of the interconnects listed as private in this screen, and they all have to be up, just as their IP addresses have to be when specified in the init.ora paramter. RAC does not fail over between cluster interconnects; if one is down then the instances using them won't start. Modified: 24-MAR-04 Ref #: ID-4724 -------------------------------------------------------------------------------- Can I change the name of my cluster after I have created it when I am using Oracle Database 10g Clusterware? No, you must properly deinstall CRS and then re-install. To properly de-install CRS, you MUST follow the directions in the Installation Guide Chapter 10. This will ensure the ocr gets cleaned out. Modified: 05-OCT-04 Ref #: ID-5890 -------------------------------------------------------------------------------- Can I change the public hostname in my Oracle Database 10g Cluster using Oracle Clusterware? Hostname changes are not supported in CRS, unless you want to perform a deletenode followed by a new addnode operation. Modified: 05-OCT-04 Ref #: ID-5892 -------------------------------------------------------------------------------- What should the permissions be set to for the voting disk and ocr when doing a RAC Install? The Oracle Real Application Clusters install guide is correct. It describes the PRE INSTALL ownership/permission requirements for ocr and voting disk. This step is needed to make sure that the CRS install succeeds. Please don't use those values to determine what the ownership/permmission should be POST INSTALL. The root script will change the ownership/permission of ocr and voting disk as part of install. The POST INSTALL permissions will end up being : OCR - root:oinstall - 640 Voting Disk - oracle:oinstall - 644 Modified: 22-OCT-04 Ref #: ID-5988 -------------------------------------------------------------------------------- Which processes access to OCR ? Oracle Cluster Registry (OCR) is used to store the cluster configuration information among other things. OCR needs to be accessible from all nodes in the cluster. If OCR became inaccessible the CSS daemon would soon fail, and take down the node. PMON never needs to write to OCR. To confirm if OCR is accessible, try ocrcheck from your ORACLE_HOME and ORA_CRS_HOME. Modified: 22-OCT-04 Ref #: ID-5990 -------------------------------------------------------------------------------- How do I restore OCR from a backup? On Windows, can I use ocopy? The only recommended way to restore an OCR from a backup is "ocrconfig -restore ". The ocopy command will not be able to perform the restore action for OCR. Modified: 27-OCT-04 Ref #: ID-6008 -------------------------------------------------------------------------------- Does the hostname have to match the public name or can it be anything else? When there is no vendor clusterware, only CRS, then the public node name must match the host name. When vendor clusterware is present, it determines the public node names, and the installer doesn't present an opportunity to change them. So, when you have a choice, always choose the hostname. Modified: 05-NOV-04 Ref #: ID-6050 -------------------------------------------------------------------------------- Is it a requirement to have the public interface linked to ETH0 or does it only need to be on a ETH lower than the private interface?: - public on ETH1 - private on ETH2 There is no requirement for interface name ordering. You could have - public on ETH2 - private on ETH0 Just make sure you choose the correct public interface in VIPCA, and in the installer's interconnect classification screen. Modified: 05-NOV-04 Ref #: ID-6052 -------------------------------------------------------------------------------- How to Restore a Lost Voting Disk used by Oracle Clusterware 10g Please read Note:279793.1 and for OCR Note:268937.1 Modified: 02-DEC-04 Ref #: ID-6308 -------------------------------------------------------------------------------- With Oracle Clusterware 10g, how do you backup the OCR? There is an automatic backup mechanism for OCR. The default location is : $ORA_CRS_HOME\cdata\"clustername"\ To display backups : ocrconfig -showbackup To restore a backup : ocrconfig -restore The automatic backup mechanism keeps upto about a week old copy. So, if you want to retain a backup copy more than that, then you should copy that "backup" file to some other name. Unfortunately there are a couple of bugs regarding backup file manipulation, and changing default backup dir on Windows. These will be fixed in 10.1.0.4. OCR backup on Windows are absent. Only file in the backup directory is temp.ocr which would be the last backup. You can restore this most recent backup by using the command ocr -restore temp.ocr If you want to take a logical copy of OCR at any time use : ocrconfig -export , and use -import option to restore the contents back. Modified: 02-DEC-04 Ref #: ID-6328 -------------------------------------------------------------------------------- How do I protect the OCR and Voting in case of media failure? In Oracle Database 10g Release 1 the OCR and Voting device are not mirrored within Oracle,hence both must be mirrored via a storage vendor method, like RAID 1. Starting with Oracle Database 10g Release 2 Oracle Clusterware will multiplex the OCR and Voting Disk (two for the OCR and three for the Voting). Please read Note:279793.1 and Note:268937.1 regarding backup and restore a lost Voting/OCR and FAQ 6238 regarding OCR backup. Modified: 05-MAY-05 Ref #: ID-6612 -------------------------------------------------------------------------------- How do I use multiple network interfaces to provide High Availability for my interconnect with Oracle Clusterware? This needs to be done externally to Oracle Clusterware usually by some OS provided nic bonding which gives Oracle Clusterware a single ip address for the interconnect but provide failover across multiple nic cards. There are several articles in Metalink on how to do this. For example for Sun Solaris search for IPMP. On Linux, read the doc on rac.us Configure Redundant Network Cards / Switches for Oracle Database 10g Release 1 Real Application Cluster on Linux Modified: 06-APR-05 Ref #: ID-6680 -------------------------------------------------------------------------------- How do I put my application under the control of Oracle Clusterware to achieve higher availability? First write a control agent. It must accept 3 different parameters: start-The control agent should start the application, check-The control agent should check the application, stop-The Control agent should start the application. Secondly you must create a profile for your application using crs_profile. Thirdly you must register your application as a resource with Oracle Clusterware (crs_register). See the RAC Admin and Deployment Guide for details. Modified: 16-JUN-05 Ref #: ID-6846 -------------------------------------------------------------------------------- Can I use Oracle Clusterware to provide cold failover of my 9i or 10g single instance Oracle Databases? Oracle does not provide the necessary wrappers to fail over single-instance databases using Oracle Clusterware 10g Release 2. But since it's possible for customers to use Oracle Clusterware to wrap arbitrary applications, it'd be possible for them to wrap single-instance databases this way. Modified: 01-JUL-05 Ref #: ID-6891 -------------------------------------------------------------------------------- Does Oracle Clusterware support application vips? Yes, with Oracle Database 10g Release 2, Oracle Clusterware now supports an "application" vip. This is to support putting applications under the control of Oracle Clusterware using the new high availability API and allow the user to use the same URL or connection string regardless of which node in the cluster the application is running on. The application vip is a new resource defined to Oracle Clusterware and is a functional vip. It is defined as a dependent resource to the application. There can be many vips defined, typically one per user application under the control of Oracle Clusterware. You must first create a profile (crs_profile), then register it with Oracle Clusterware (crs_register). The usrvip script must run as root. Modified: 11-JUL-05 Ref #: ID-6893 -------------------------------------------------------------------------------- Why is the home for Oracle Clusterware not recommended to be subdirectory of the Oracle base directory? If anyone other than root has write permissions to the parent directories of the CRS home, then they can give themselves root escalations. This is a security issue. The CRS home itself is a mix of root and non-root permissions, as appropriate to the security requirements. Please follow the install docs about who is your primary group and what other groups you need to create and be a member of. Modified: 11-JUL-05 Ref #: ID-6915 -------------------------------------------------------------------------------- . -------------------------------------------------------------------------------- Copyright © 9.6 JRE: ======== JRE: ---- Oracle 9.2 uses JRE 1.3.1 - Java Compiler (javac): Compiles programs written in the Java programming language into bytecodes. - Java Interpreter (java): Executes Java bytecodes. In other words, it runs programs written in the Java programming language. - Jave Runtime Interpreter (jre): Similar to the Java Interpreter (java), but intended for end users who do not require all the development-related options available with the java tool. The PATH statement enables Windows to find the executables (javac, java, javadoc, etc.) from any current directory. The CLASSPATH tells the Java virtual machine and other applications (which are located in the "jdk_\bin" directory) where to find the class libraries, such as classes.zip file (which is in the lib directory). Note 1: ------- Suppose on a Solaris 5.9 machine with Oracle 9.2, we search for jre: # find . -name "jre*" -print ./opt/app/oracle/product/9.2/inventory/filemap/jdk/jre ./opt/app/oracle/product/9.2/jdk/jre ./opt/app/oracle/jre ./opt/app/oracle/jre/1.1.8/bin/sparc/native_threads/jre ./opt/app/oracle/jre/1.1.8/bin/jre ./opt/app/oracle/jre/1.1.8/jre_config.txt ./usr/j2se/jre ./usr/iplanet/console5.1/bin/base/jre ./usr/java1.2/jre Suppose on a AIX 5.2 machine with Oracle 9.2, we search for jre: ./apps/oracle/product/9.2/inventory/filemap/jdk/jre ./apps/oracle/product/9.2/inventory/filemap/jre ./apps/oracle/product/9.2/jdk/jre ./apps/oracle/product/9.2/jre ./apps/oracle/oraInventory/filemap/apps/oracle/jre ./apps/oracle/oraInventory/filemap/apps/oracle/jre/1.3.1/jre ./apps/oracle/jre ./apps/oracle/jre/1.1.8/bin/jre ./apps/oracle/jre/1.1.8/bin/aix/native_threads/jre ./apps/oracle/jre/1.3.1/jre ./apps/ora10g/product/10.2/jdk/jre ./apps/ora10g/product/10.2/jre ./usr/java131/jre ./usr/idebug/jre Note 2: ------- jre - The Java Runtime Interpreter (Solaris) jre interprets (executes) Java bytecodes. SYNOPSIS jre [ options ] classname DESCRIPTION The jre command executes Java class files. The classname argument is the name of the class to be executed. Any arguments to be passed to the class must be placed after the classname on the command line. Class paths for the Solaris version of the jre tool can be specified using the CLASSPATH environment variable or by using the -classpath or -cp options. The Windows version of the jre tool ignores the CLASSPATH environment variable. For both Solaris and Windows, the -cp option is recommend for specifying class paths when using jre. OPTIONS -classpath path(s) Specifies the path or paths that jre uses to look up classes. Overrides the default or the CLASSPATH environment variable if it is set. If more than one path is specified, they must be separated by colons. Each path should end with the directory containing the class file(s) to be executed. However, if a file to be executed is a zip or jar file, the path to that file must end with the file's name. Here is an example of an argument for -classpath that specifies three paths consisting of the current directory and two additional paths: .:/home/xyz/classes:/usr/local/java/classes/MyClasses.jar -cp path(s) Prepends the specified path or paths to the base classpath or path given by the CLASSPATH environment variable. If more than one path is specified, they must be separated by colons. Each path should end with the directory containing the class file(s) to be executed. However, if a file to be executed is a zip or jar file, the path to that file must end with the file's name. Here is an example of an argument for -cp that specifies three paths consisting of the current directory and two additional paths: .:/home/xyz/classes:/usr/local/java/classes/MyClasses.jar -help Print a usage message. -mx x Sets the maximum size of the memory allocation pool (the garbage collected heap) to x. The default is 16 megabytes of memory. x must be greater than or equal to 1000 bytes. By default, x is measured in bytes. You can specify x in either kilobytes or megabytes by appending the letter "k" for kilobytes or the letter "m" for megabytes. -ms x Sets the startup size of the memory allocation pool (the garbage collected heap) to x. The default is 1 megabyte of memory. x must be > 1000 bytes. By default, x is measured in bytes. You can specify x in either kilobytes or megabytes by appending the letter "k" for kilobytes or the letter "m" for megabytes. -noasyncgc Turns off asynchronous garbage collection. When activated no garbage collection takes place unless it is explicitly called or the program runs out of memory. Normally garbage collection runs as an asynchronous thread in parallel with other threads. -noclassgc Turns off garbage collection of Java classes. By default, the Java interpreter reclaims space for unused Java classes during garbage collection. -nojit Specifies that any JIT compiler should be ignored and instead invokes the default Java interpreter. -ss x Each Java thread has two stacks: one for Java code and one for C code. The -ss option sets the maximum stack size that can be used by C code in a thread to x. Every thread that is spawned during the execution of the program passed to jre has x as its C stack size. The default units for x are bytes. The value of x must be greater than or equal to 1000 bytes. You can modify the meaning of x by appending either the letter "k" for kilobytes or the letter "m" for megabytes. The default stack size is 128 kilobytes ("-ss 128k"). -oss x Each Java thread has two stacks: one for Java code and one for C code. The -oss option sets the maximum stack size that can be used by Java code in a thread to x. Every thread that is spawned during the execution of the program passed to jre has x as its Java stack size. The default units for x are bytes. The value of x must be greater than or equal to 1000 bytes. You can modify the meaning of x by appending either the letter "k" for kilobytes or the letter "m" for megabytes. The default stack size is 400 kilobytes ("-oss 400k"). -v, -verbose Causes jre to print a message to stdout each time a class file is loaded. -verify Performs byte-code verification on the class file. Beware, however, that java -verify does not perform a full verification in all situations. Any code path that is not actually executed by the interpreter is not verified. Therefore, java -verify cannot be relied upon to certify class files unless all code paths in the class file are actually run. -verifyremote Runs the verifier on all code that is loaded into the system via a classloader. verifyremote is the default for the interpreter. -noverify Turns verification off. -verbosegc Causes the garbage collector to print out messages whenever it frees memory. -DpropertyName=newValue Defines a property value. propertyName is the name of the property whose value you want to change and newValue is the value to change it to. For example, this command line % jre -Dawt.button.color=green ... sets the value of the property awt.button.color to "green". jre accepts any number of -D options on the command line. ENVIRONMENT VARIABLES CLASSPATH You can use the CLASSPATH environment variable to specify the path to the class file or files that you want to execute. CLASSPATH consists of a colon-separated list of directories that contain the class files to be executed. For example: .:/home/xyz/classes If the file to be executed is a zip file or a jar file, the path should end with the file name. For example: .:/usr/local/java/classes/MyClasses.jar SEE ALSO CLASSPATH Note 3: ------- Solaris: Installing IBM JRE, Version 1.3.1 To install JRE 1.3.1 on Solaris, follow these steps: Log on as root. Insert the IBM Tivoli Access Manager for Solaris CD. Install the IBM JRE 1.3.1 package: pkgadd -d /cdrom/cdrom0/solaris -a /cdrom/cdrom0/solaris/pddefault SUNWj3rt where -d /cdrom/cdrom0/solaris specifies the location of the package and -a /cdrom/cdrom0/solaris/pddefault specifies the location of the installation administration script. Set the PATH environmental variable: PATH=/usr/j2se/jre/bin:$PATH export PATH After you install IBM JRE 1.3.1, no configuration is necessary. ################################################################################### ========= 30 LOBS: ========= 30.1 General LOB info: ---------------------- Note 1: ======= A LOB is a Large Object. LOBs are used to store large, unstructured data, such as video, audio, photo images etc. With a LOB you can store up to 4 Gigabytes of data. They are similar to a LONG or LONG RAW but differ from them in quite a few ways. LOBs offer more features to the developer than a LONG or LONG RAW. The main differences between the data types also indicate why you would use a LOB instead of a LONG or LONG RAW. These differences include the following: - · You can have more than one LOB column in a table, whereas you are restricted to just one LONG or LONG RAW column per table. · When you insert into a LOB, the actual value of the LOB is stored in a separate segment (except for in-line LOBs) and only the LOB locator is stored in the row, thus making it more efficient from a storage as well as query perspective. With LONG or LONG RAW, the entire data is stored in-line with the rest of the table row. · LOBs allow a random access to its data, whereas with a LONG you have to go in for a sequential read of the data from beginning to end. · The maximum length of a LOB is 4 Gig as compared to a 2 Gig limit on LONG · Querying a LOB column returns the LOB locator and not the entire value of the LOB. On the other hand, querying LONG returns the entire value contained within the LONG column You can have two categories of LOBs based on their location with respect to the database. The categories include internal LOBs and external LOBs. As the names suggest, internal LOBs are stored within the database, as table columns. External LOBs are stored outside the database as operating system files. Only a reference to the actual OS file is stored in the database. An internal LOB can also be persistent or temporary depending on the life of the internal LOB. An internal LOB can be one of three different data types as follows: - · CLOB – A Character LOB. Used to store character data. · BLOB – A Binary LOB. Used to store binary, raw data · NCLOB – A LOB that stores character data that corresponds to the national character set defined for the database. The only external LOB data type in Oracle 8i is called a BFILE. · BFILE - Short for Binary File. These hold references to large binary data stored as physical files in the OS outside the database. DBA_LOBS displays the BLOBs and CLOBs contained in all tables in the database. BFILEs are stored outside the database, so they are not described by this view. This view's columns are the same as those in "ALL_LOBS". NCLOB and CLOB, are both encoded a internal fixed-width Unicode character set. CLOB = Character Large Object 4Gigabytes NCLOB = National Character Large Object 4Gigabytes BLOB = Binary Large Object 4Gigabytes BFILE = pointer to binary file on disk 4Gigabytes - A limited number of BFILEs can be open simultaneously per session. The initialization parameter, SESSION_MAX_OPEN_FILES defines an upper limit on the number of simultaneously open files in a session. The default value for this parameter is 10. That is, you can open a maximum of 10 files at the same time per session if the default value is utilized. If you want to alter this limit, the database administrator can change the value of this parameter in the init.ora file. For example: SESSION_MAX_OPEN_FILES=20 If the number of unclosed files exceeds the SESSION_MAX_OPEN_FILES value then you will not be able to open any more files in the session. To close all open files, use the FILECLOSEALL call. - LOB locators Regardless of where the value of the internal LOB is stored, a locator is stored in the row. You can think of a LOB locator as a pointer to the actual location of the LOB value. A LOB locator is a locator to an internal LOB while a BFILE locator is a locator to an external LOB. When the term locator is used without an identifying prefix term, it refers to both LOB locators and BFILE locators. - Internal LOB Locators For internal LOBs, the LOB column stores a locator to the LOB's value which is stored in a database tablespace. Each LOB column/attribute for a given row has its own distinct LOB locator and copy of the LOB value stored in the database tablespace. - LOB Locator Operations Setting the LOB Column/Attribute to contain a locator Before you can start writing data to an internal LOB, the LOB column/attribute must be made non-null, that is, it must contain a locator. Similarly, before you can start accessing the BFILE value, the BFILE column/attribute must be made non-null. For internal LOBs, you can accomplish this by initializing the internal LOB to empty in an INSERT/UPDATE statement using the functions EMPTY_BLOB() for BLOBs or EMPTY_CLOB() for CLOBs and NCLOBs. For external LOBs, you can initialize the BFILE column to point to an external file by using the BFILENAME() function. Note 2: ======= From: Oracle, Kalpana Malligere 29-Aug-01 14:50 Subject: Re : What is my best LOB choice Hello, There are several articles/discussions available in the MetaLink Repository which discuss LOBs, including BFILEs. They are accessible via the Search option and the following articles should assist you to make you choice: 66431.1 LOBS - Storage, Redo and Performance Issues 66046.1 Oracle8i: LOBs 107441.1 Comparison between LOBs, and LONG & LONG Raw Datatypes To find any performance comparison between BFILEs and BLOBs, the best suggestion is to try a small scale test. One of the customer wrote that his rule of thumb is that a small number of large LOBs => bfile, and a large number of small LOBs => BLOB. The BLOB datatype can store up to 4Gb of data. BLOBs can participate fully in transactions. Changes made to a BLOB value by the DBMS_LOB package, PL/SQL, or the OCI can be committed or rolled back. The BFILE datatype stores unstructured binary data (such as image files) in operating-system files outside the database. A BFILE column or attribute stores a file locator that points to an external file containing the data. BFILEs can also store up to 4Gb of data. Howerver, BFILEs are read-only; you cannot modify them. They support only random (not sequential) reads, and they do not participate in transactions. The underlying operating system must maintain the file integrity and durability for BFILEs. The database administrator must ensure that the file exists and that Oracle processes have operating-system read permissions on the file. Your application will have an impact on which is preferable. BFILEs will really help if your application is WEB based because you can access them through an annonymous FTP connect into the browser by passing the URL to the HTML. You can also do this through a regular BLOB, but this would make you drag the entire image through the Oracle server buffer cache everytime it is requested. The separation of the backup can be beneficial especially if the the image files are mostly static. This reduces the backup volume of the database itself. You also don't need a special program for loading them into the database. You just copy the files to the OS and run a DML statement to add them. This way you also avoid the redo created by inserting them as an internal BLOB. On the other side of the coin, you will have to devise a file naming convention/directory structure to prevent overwriting the BFILE's. You may want to do only one backup instead of both. With BLOBs, if you backup the database, you have everything needed. You won't be able to update a BFILE through the database, you will always have to make modifcations through the OS. LOB types can be replicated, but not BFILE. The Oracle 8i Application Developer's Guide - Large Objects (LOBs), provides information on the various programmatic environments and how to operate on LOB and BFILE data. Questions on these capabilities should be posted to the appropriate forum (i.e. Oracle PL/SQL, Oracle Call Interface, Oracle Precompiler, etc.). To answer your question, it depends on how you want to use the data. A LOB is stored in line by default if it is less than 3,960 bytes, whereas an out-of-line LOB takes about 20 bytes per row. An inline LOB (i.e. one that is actually stored in the row) is always logged, but an out-of-line can be made non-logging. Preference is always to DISABLE STORAGE IN ROW, but if your LOBs are actually very small, and the way you use them is sufficiently special then you may want to store them in line. But if so, they could probably become simple varchar2(4000). Note - the minimum size an out-of-line LOB can use is one Oracle block (plus a bit of extra space in the LOBINDEX). Thanks! Kalpana Oracle Technical Support Note 3: ======= Doc ID : Note:66431.1 Content Type: TEXT/PLAIN Subject: LOBS - Storage, Redo and Performance Issues Creation Date: 05-NOV-1998 Type: BULLETIN Last Revision Date: 25-JUL-2002 Status: PUBLISHED Introduction ~~~~~~~~~~~~ This is a short note on the internal storage of LOBs. The information here is intended to supplement the documentation and other notes which describe how to use LOBS. The focus is on the storage characteristics and configuration issues which can affect performance. There are 4 types of LOB: CLOB, BLOB, NCLOB stored internally to Oracle BFILE stored externally The note mainly discusses the first 3 types of LOB which as stored INTERNALLY within the Oracle DBMS. BFILE's are pointers to external files and are only mentioned briefly. Examples of handling LOBs can be found in [NOTE:47740.1] Attributes ~~~~~~~~~~ There are many attributes associated with LOB columns. The aim here is to cover the fundamental points about each of the main attributes. The attributes for each LOB column are specified using the "LOB (lobcolname) STORE AS ..." syntax. A table containing LOBs (CLOB, NCLOB and BLOB) creates 2 additional disk segments per LOB column - a LOBINDEX and a LOBSEGMENT. These can be viewed, along with the LOB attributes, using the dictionary views: DBA_LOBS, ALL_LOBS or USER_LOBS which give the columns: OWNER Table Owner TABLE_NAME Table name COLUMN_NAME Column name in the table SEGMENT_NAME Segment name of the LOBSEGMENT INDEX_NAME Segment name of the LOBINDEX CHUNK Chunk size (bytes) PCTVERSION PctVersion CACHE Cache option of the LOB Segment (yes/no) LOGGING Logging mode of the LOB segment (yes/no) IN_ROW Whether storage in row is allowed (yes/no) SELECT l.table_name as "TABLE", l.column_name as "COLUMN", l.segment_name as "SEGMENT", l.index_name as "INDEX", l.chunk as "CHUNKSIZE", l.LOGGING, l.IN_ROW, t.tablespace_name FROM DBA_LOBS l, DBA_TABLES t WHERE l.table_name=t.table_name AND l.owner in ('VPOUSERDB','TRIDION_CM'); Storage Parameters ~~~~~~~~~~~~~~~~~~ By default LOB segments are created in the same tablespace as the base table using the tablespaces default storage details. You can specify the storage attributes of the LOB segments thus: Create table DemoLob ( A number, B clob ) LOB(b) STORE AS lobsegname ( TABLESPACE lobsegts STORAGE (lobsegment storage clause) INDEX lobindexname ( TABLESPACE lobidxts STORAGE ( lobindex storage clause ) ) ) TABLESPACE tables_ts STORAGE( tables storage clause ) ; CREATE TABLE t_lob (DOCUMENT_NR NUMBER(16,0) NOT NULL, DOCUMENT_BLOB BLOB NOT NULL ) STORAGE (INITIAL 100k NEXT 100K PCTINCREASE 0 MAXEXTENTS 100 ) TABLESPACE system lob (DOCUMENT_BLOB) store as DOCUMENT_LOB (tablespace ts storage (initial 30K next 30K pctincrease 30 maxextents 3) index (tablespace ts_index storage (initial 40K next 40K pctincrease 40 maxextents 4))); In 8.0 the LOB INDEX can be stored separately from the lob segment. If a tablespace is specified for the LOB SEGMENT then the LOB INDEX will be placed in the same tablespace UNLESS a different tablespace is explicitly specified. Unless you specify names for the LOB segments system generated names are used. In ROW Versus Out of ROW ~~~~~~~~~~~~~~~~~~~~~~~~ LOB columns can be allowed to store data within the row or not as detailed below. Whether in-line storage is allowed or not can ONLY be specified at creation time. "STORE AS ( enable storage in row )" Allows LOB data to be stored in the TABLE segment provided it is less than about 4000 bytes. The actual maximum in-line LOB is 3964 bytes. If the lob value is greater than 3964 bytes then the LOB data is stored in the LOB SEGMENT (ie: out of line). An out of line LOB behaves as described under 'disable storage in row' except that if its size shrinks to 3964 or less the LOB can again be stored inline. When a LOB is stored out-of-line in an 'enable storage in row' LOB column between 36 and 84 bytes of control data remain in-line in the row piece. In-line LOBS are subject to normal chaining and row migration rules within Oracle. Ie: If you store a 3900 byte LOB in a row with a 2K block size then the row piece will be chained across two or more blocks. Both REDO and UNDO are written for in-line LOBS as they are part of the normal row data. "STORE AS ( disable storage in row )" This option prevents any size of LOB from being stored in-line. Instead a 20 byte LOB locator is stored in the ROW which gives a unique identifier for a LOB in the LOB segment for this column. The Lob Locator actually gives a key into the LOB INDEX which contains a list of all blocks (or pages) that make up the LOB. The minimum storage allocation for an out of line LOB is 1 Database BLOCK per LOB ITEM and may be more if CHUNK is larger than a single block. UNDO is only written for the column locator and LOB INDEX changes. No UNDO is generated for pages in the LOB SEGMENT. Consistent Read is achieved by using page versions. Ie: When you update a page of a LOB the OLD page remains and a new page is created. This can appear to waste space but old pages can be reclaimed and reused. CHUNK size ~~~~~~~~~~ "STORE AS ( CHUNK bytes ) " Can ONLY be specified at creation time. In 8.0 values of CHUNK are in bytes and are rounded to the next highest multiple of DB_BLOCK_SIZE without erroring. Eg: If you specify a CHUNK of 3000 with a block size of 2K then CHUNK is set to 4096 bytes. "bytes" / DB_BLOCK_SIZE determines the unit of allocation of blocks to an 'out of line' LOB in the LOB segment. Eg: if CHUNK is 32K and the LOB is 'disable storage in row' then even if the LOB is only 10 bytes long 32K will be allocated in the LOB SEGMENT. CHUNK does NOT affect in-line LOBS. PCTVERSION ~~~~~~~~~~ "STORE AS ( PCTVERSION n )" PCTVERSION can be changed after creation using: ALTER TABLE tabname MODIFY LOB (lobname) ( PCTVERSION n ); PCTVERSION affects the reclamation of old copies of LOB data. This affects the ability to perform consistent read. If a session is attempting to use an OLD version of a LOB and that version gets overwritten (because PCTVERSION is too small) then the user will typically see the errors: ORA-01555: snapshot too old: rollback segment number with name "" too small ORA-22924: snapshot too old PCTVERSION can prevent OLD pages being used and force the segment to extend instead. Do not expect PCTVERSION to be an exact percentage of space as there is an internal fudge factor applied. CACHE ~~~~~ "STORE AS ( CACHE )" or "STORE AS ( NOCACHE )" This option can be changed after creation using: ALTER TABLE tabname MODIFY LOB (lobname) ( CACHE ); or ALTER TABLE tabname MODIFY LOB (lobname) ( NOCACHE ); With NOCACHE set (the default) reads from and writes to the LOB SEGMENT occur using direct reads and writes. This means that the blocks are never cached in the buffer cache and the the Oracle shadow process performs the reads/writes itself. The reads / writes show up under the wait events "direct path read" and "direct path write" and multiple blocks can be read/written at a time (provided the caller is using a large enough buffer size). When set the CACHE option causes the LOB SEGMENT blocks to be read / written via the buffer cache . Reads show up as "db file sequential read" but unlike a table scan the blocks are placed at the most-recently-used end of the LRU chain. The CACHE options for LOB columns is different to the CACHE option for tables as CACHE_SIZE_THRESHOLD does not limit the size of LOB read into the buffer cache. This means that extreme caution is required otherwise the read of a long LOB can effectively flush the cache. In-line LOBS are not affected by the CACHE option as they reside in the actual table block (which is typically accessed via the buffer cache any way). The cache option can affect the amount of REDO generated for out of line LOBS. With NOCACHE blocks are direct loaded and so entire block images are written to the REDO stream. If CHUNK is also set then enough blocks to cover CHUNK are written to REDO. If CACHE is set then the block changes are written to REDO. Eg: In the extreme case 'DISABLE STORAGE IN ROW NOCACHE CHUNK 32K' would write redo for the whole 32K even if the LOB was only 5 characters long. CACHE would write a redo record describing the 5 byte change (taking about 100-200 bytes). LOGGING ~~~~~~~ "STORE AS ( NOCACHE LOGGING )" or "STORE AS ( NOCACHE NOLOGGING )" This option can be changed after creation but the LOGGING / NOLOGGING attribute must be prefixed by the NOCACHE option. The CACHE option implicitly enables LOGGING. The default for this option is LOGGING. If a LOB is set to NOCACHE NOLOGGING then updates to the LOB SEGMENT are not logged to the redo logs. However, updates to in-line LOBS are still logged as normal. As NOCACHE operations use direct block updates then all LOB segment operations are affected. NOLOGGING of the LOB segment means that if you have to recover the database then sections of the LOB segment will be marked as corrupt during recovery. Space required for updates ~~~~~~~~~~~~~~~~~~~~~~~~~~ If a LOB is out-of-line then updates to pages if the LOB cause new versions of those pages to be created. Rollback is achieved by reverting back to the pre-updated page versions. This has implications on the amount of space required when a LOB is being updated as the LOB SEGMENT needs enough space to hold both the OLD and NEW pages concurrently in case your transaction rolls back. Eg: Consider the following: INSERT a large LOB LOB SEGMENT extends take the new pages COMMIT; DELETE the above LOB The LOB pages are not yet free as they will be needed in case of rollback. INSERT a new LOB Hence this insert may require more space in the LOB SEGMENT COMMIT; Only after this point could the deleted pages be used. Performance Issues ~~~~~~~~~~~~~~~~~~~ Working with LOBs generally requires more than one round trip to the database. The application first has to obtain the locator and only then can perform operations against that locator. This is true for inline or out of line LOBS. The buffer size used to read / write the LOB can have a significant impact on performance, as can the SQL*Net packet sizes. Eg: With OCILobRead() a buffer size is specified for handling the LOB. If this is small (say 2K) then there can be a round trip to the database for each 2K chunk of the LOB. To make the issue worse the server will only fetch the blocks needed to satisfy the current request so may perform single block reads against the LOB SEGMENT. If however a larger chunk size is used (say 32K) then the server can perform multiblock operations and pass the data back in larger chunks. There is a LOB buffering subsystem which can be used to help improve the transfer of LOBs between the client and server processes. See the documentation for details of this. BFILEs ~~~~~~ BFILEs are quite different to internal LOBS as the only real storage issue is the space required for the inline locator. This is about 20 bytes PLUS the length of the directory and filename elements of the BFILENAME. The performance implications of the buffer size are the same as for internal LOBS. References ~~~~~~~~~~ [NOTE:162345.1] LOBS - Storage, Read-consistency and Rollback Note 4: ======= Doc ID: Note:159995.1 Content Type: TEXT/X-HTML Subject: Different Behaviors of Lob and Lobindex Segments in 8.0, 8i and 9i Creation Date: 05-OCT-2001 Type: BULLETIN Last Revision Date: 27-MAR-2003 Status: PUBLISHED PURPOSE ------- This bulletin lists the different behaviors of a lob index segment regarding tablespace and storage values: -> When creating the table, the lob and lob index segments -> Altering the associated lob segment and/or lob index segment. SCOPE & APPLICATION ------------------- For all DBAs who manage different versions of Oracle with databases containing LOB segments, and who need to maintain the associated lob indexes. Under 8i and 9i In Oracle8i SQL Reference and Oracle9i SQL Reference, it is clearly stated that: lob_index_clause This clause is deprecated as of Oracle8i. Oracle generates an index for each LOB column. Oracle names and manages the LOB indexes internally. Although it is still possible for you to specify this clause, Oracle Corporation strongly recommends that you no longer do so. In any event, do not put the LOB index in a different tablespace from the LOB data. 1.Lob and lobindex specifications at table creation If you create a new table in release 8i and 9i and specify a tablespace and storage values for the LOB index for a non-partitioned table, the tablespace specification and storage values are ignored. The LOB index is located in the same tablespace as the LOB segment with the same storage values, except the NEXT and MAXEXTENTS values. the NEXT value of the lobindex = INITIAL default value of the tablespace (LOB segment) the MAXEXTENTS value of the lobindex = unlimited value (2Gb) SQL> CREATE TABLE t_lob 2 (DOCUMENT_NR NUMBER(16,0) NOT NULL, 3 DOCUMENT_BLOB BLOB NOT NULL 4 ) 5 STORAGE 6 (INITIAL 100k 7 NEXT 100K 8 PCTINCREASE 0 9 MAXEXTENTS 100 10 ) 11 TABLESPACE system 12 lob (DOCUMENT_BLOB) store as DOCUMENT_LOB 13 (tablespace ts storage 14 (initial 30K next 30K pctincrease 30 maxextents 3) 15 index (tablespace ts_index storage 16 (initial 40K next 40K pctincrease 40 maxextents 4))); Table created. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS 30720 10240 30 2147483645 DOCUMENT_LOB LOBSEGMENT TS 30720 30720 30 3 All storage modifications are based on this original table t_lob. 2.Lob and lobindex storage modifications When you modify the storage values for the lob and lob index segments, the values of the lob index are kept as initially set, except the PCT_INCREASE. The value of the lob segment PCTINCREASE spreads out on the lob index: SQL> alter table t_lob 2 modify lob (document_blob) 3 (storage (next 60K pctincrease 60 maxextents 6) 4 index (storage (next 70K pctincrease 70 maxextents 7))); Table altered. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS 30720 10240 60 2147483645 DOCUMENT_LOB LOBSEGMENT TS 30720 61440 60 6 3.Storage modifications of lob segment only If you modify the storage values for the lob segment only, you get the same behaviour: SQL> alter table t_lob 2 modify lob (document_blob) 3 (storage (next 60K pctincrease 60 maxextents 6)); Table altered. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS 30720 10240 60 2147483645 DOCUMENT_LOB LOBSEGMENT TS 30720 61440 60 3 4.Storage modifications of lobindex segment only If you modify the storage values for the lob index segment only, nothing is altered: SQL> alter table t_lob 2 modify lob (document_blob) 3 (index (storage (next 70K pctincrease 70 maxextents 7))) 4 ; Table altered. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS 30720 10240 30 2147483645 DOCUMENT_LOB LOBSEGMENT TS 30720 30720 30 3 If you attempt to modify the storage values of the lob index directly, you get an error message: SQL> alter index SYS_IL0000020297C00002$$ storage (pctincrease 80); alter index SYS_IL0000020297C00002$$ storage (pctincrease 80) * ERROR at line 1: ORA-22864: cannot ALTER or DROP LOB indexes SQL> alter index SYS_IL0000020297C00002$$ rebuild storage (pctincrease 60); alter index SYS_IL0000020297C00002$$ rebuild storage (pctincrease 60) * ERROR at line 1: ORA-02327: cannot create index on expression with datatype LOB Under 8.0 1.Lob and lobindex specifications at table creation If you create a new table in release 8.0 and specify a tablespace for the LOB index for a non-partitioned table, the tablespace specification and storage values are encountered. The LOB index is located in the defined tablespace with the user-defined storage values. SQL> CREATE TABLE t_lob 2 (DOCUMENT_NR NUMBER(16,0) NOT NULL, 3 DOCUMENT_BLOB BLOB NOT NULL 4 ) 5 STORAGE 6 (INITIAL 100k 7 NEXT 100K 8 PCTINCREASE 0 9 MAXEXTENTS 100 10 ) 11 TABLESPACE system 12 lob (DOCUMENT_BLOB) store as DOCUMENT_LOB 13 (tablespace ts storage 14 (initial 30K next 30K pctincrease 30 maxextents 3) 15 index (tablespace ts_index storage 16 (initial 40K next 40K pctincrease 40 maxextents 4))); Table created. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS_INDEX 40960 40960 40 4 DOCUMENT_LOB LOBSEGMENT TS 32768 30720 30 3 All storage modifications are based on this original table t_lob. 2.Lob and lobindex storage modifications When you modify the storage values for the lob and lob index segments, the values for the lobindex are kept as initially set: SQL> alter table t_lob 2 modify lob (document_blob) 3 (storage (next 60K pctincrease 60 maxextents 6) 4 index (storage (next 70K pctincrease 70 maxextents 7))); Table altered. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS_INDEX 40960 40960 40 4 DOCUMENT_LOB LOBSEGMENT TS 32768 61440 60 6 3.Storage modifications of lob segment only If you modify the storage values for the lob segment only, you get the same behavior: SQL> alter table t_lob 2 modify lob (document_blob) 3 (storage (next 60K pctincrease 60 maxextents 6)); Table altered. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS_INDEX 40960 40960 40 4 DOCUMENT_LOB LOBSEGMENT TS 32768 61440 60 6 Again, the lob segment storage values do not impact the lob index. 4.Storage modifications of lobindex segment only If you modify the storage values for the lob index segment only, nothing is altered: SQL> alter table t_lob 2 modify lob (document_blob) 3 (index (storage (next 70K pctincrease 70 maxextents 7))) 4 ; Table altered. SQL> select segment_name, segment_type, tablespace_name, 2 initial_extent, next_extent, pct_increase, max_extents 3 from user_segments; SEGMENT_NAME SEGMENT_TY TABLESPA INITIAL NEXT_EXT PCT_INC MAX_EXT ----------------------- ----------- --------- -------- -------- ------- --------- T_LOB TABLE SYSTEM 102400 102400 0 100 SYS_IL0000020297C00002$$ LOBINDEX TS_INDEX 40960 40960 40 4 DOCUMENT_LOB LOBSEGMENT TS 32768 30720 30 3 If you attempt to modify the storage values of the lob index directly, you get an error message: SQL> alter index SYS_IL0000020297C00002$$ storage (pctincrease 20); alter index SYS_IL0000020297C00002$$ storage (pctincrease 20) * ERROR at line 1: ORA-22864: cannot ALTER or DROP LOB indexes Migration from 7 to 9i The "Oracle9i Database Migration Release 1 (9.0.1)" documentation states: LOB Index Clause If you used the LOB index clause to store LOB index data in a tablespace separate from the tablespace used to store the LOB, the index data is relocated to reside in the same tablespace as the LOB. If you used Export/Import to migrate from Oracle7 to Oracle9i, the index data was relocated automatically during migration. However, the index data was not relocated if you used the Migration utility or the Oracle Data Migration Assistant. RELATED DOCUMENTS ----------------- LOBS - Storage, Redo and Performance Issues ALTER TABLE MODIFY DEFAULT ATTRIBUTES LOB DOES NOT UPDATE LOB INDEX DEFAULT TS LARGE LOB INDEX SEGMENT SIZE ALTER TABLE MODIFY LOB STORAGE PARAMETER DOES'T WORK UNABLE TO CHANGE STORAGE CHARACTERISTICS FOR LOB INDEXES Note 5: ======= Calculate sizes: Example ------- SQL> create table my_lob 2 (idx number null, a_lob clob null, b_lob blob null) 3 storage (initial 20k maxextents 121 pctincrease 0 ) 4 lob (a_lob, b_lob) store as 5 ( storage ( initial 100k next 100K maxextents 999 pctincrease 0)); Table created. SQL> select object_name,object_type,object_id from user_objects order by 2; OBJECT_NAME OBJECT_TYPE OBJECT_ID ---------------------------------------- ------------------ ---------- SYS_LOB0000004017C00002$$ LOB 4018 SYS_LOB0000004017C00003$$ LOB 4020 MY_LOB TABLE 4017 SQL> select bytes, s.segment_name,s.segment_type 2 from dba_segments s 3 where s.segment_name='MY_LOB'; BYTES SEGMENT_NAME SEGMENT_TYPE ---------- ------------------------------ ------------------ 65536 MY_LOB TABLE SQL> select sum(bytes), s.segment_name, s.segment_type 2 from dba_lobs l, dba_segments s 3 where s.segment_type = 'LOBSEGMENT' 4 and l.table_name = 'MY_LOB' 5 and s.segment_name = l.segment_name 6 group by s.segment_name,s.segment_type; SUM(BYTES) SEGMENT_NAME SEGMENT_TYPE ---------- ------------------------------ ------------------ 131072 SYS_LOB0000004017C00002$$ LOBSEGMENT 131072 SYS_LOB0000004017C00003$$ LOBSEGMENT Therefore the total size for the table MY_LOB is: 65536 (for the table) + 131072 (for CLOB segment) + 131072 (for BLOB segment) => 327680 bytes Note 6: ======= Doc ID: Note:268476.1 Subject: LOB Performance Guideline Type: WHITE PAPER Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 09-APR-2004 Last Revision Date: 22-JUN-2004 LOB Performance Guidelines An Oracle White Paper April 2004 LOB Performance Guidelines Executive Overview.............................................................................. 3 LOB Overview...................................................................................... 3 Important Storage Parameters................................................................ 4 CHUNK............................................................................................ 4 Definition...................................................................................... 4 Points to Note............................................................................... 4 Recommendation........................................................................... 4 In-line and Out-of-Line storage: ENABLE STORAGE IN ROW and DISABLE STORAGE IN ROW 4 Definition...................................................................................... 4 Points to Note............................................................................... 5 Recommendation........................................................................... 5 CACHE, NOCACHE....................................................................... 5 Definition...................................................................................... 5 Points to Note............................................................................... 6 Recommendation........................................................................... 6 Consistent Reads on LOBs: RETENTION and PCTVERSION...... 6 Definition...................................................................................... 6 Points to Note............................................................................... 6 Recommendation........................................................................... 7 LOGGING, NOLOGGING............................................................. 7 Definition...................................................................................... 7 Points to Note............................................................................... 7 Recommendation........................................................................... 7 Performance GUIDELINE ? LOB Loading.......................................... 8 Points to Note................................................................................... 8 Use array operations for LOB inserts............................................. 8 Scalability problem ? with LOB disable storage in row option...... 8 Row Chaining problem ? with the use of OCILobWrite API......... 8 High number of consistent read blocks created and examined...... 9 CPU time and Elapsed time - not reported accurately................... 9 Reads/Writes are done one chunk at a time in synchronous way 10 High CPU system time................................................................. 11 Buffer cache sizing problem......................................................... 11 Multi-byte character set conversion............................................. 11 HWM enqueue contention........................................................... 11 RAC environment issues.............................................................. 12 Other LOB performance related issues....................................... 12 APPENDIX A..................................................................................... 13 LONG API access to LOB datatype............................................... 13 APPENDIX B..................................................................................... 15 Migration from in-line to out-of-line (and out-of-line to in-line) storage 15 APPENDIX C..................................................................................... 16 How LOB data is stored.................................................................. 16 In-line LOB ? LOB size less than 3964 bytes............................. 16 In-line LOB ? LOB size = 3965 bytes (1 byte greater than 3964) 16 In-line LOB ? LOB size greater than 12 chunk addresses........... 17 Out-of-line LOBs ? All LOB sizes.............................................. 17 LOB Performance Guidelines Executive Overview This document gives a brief overview of Oracle?s LOB data structure, emphasizing various storage parameter options and describes scenarios where those storage parameters are best used. The purpose of the latter is to help describe the effects of readers select the appropriate LOB storage options. This paper assumes that most customers load LOB data once and retrieve many times (less than 10% of DML is update and delete), so performance guidelines provided here are for LOB loading. LOBs were designed to efficiently store and retrieve large amounts of data. Small LOBs (< 1MB) perform better than LONGs for inserts, and have comparable performance on selects. Large LOBs perform better than LONGs in general. Oracle recommends the use of LOBs to store unstructured or semi-structured data, and has provided a LONG API to allow ease of migration from LONGs to LOBs. Oracle plans to de-support LONGs in the future. LOB Overview Whenever a table containing a LOB column is created, two segments are created to hold the specified LOB column. These segments are of type LOBSEGMENT and LOBINDEX. The LOBINDEX segment is used to access LOB chunks/pages that are stored in the LOBSEGMENT segment. CREATE TABLE foo (pkey NUMBER, bar BLOB); SELECT segment_name, segment_type FROM user_extents; 9792 is the object_id of the parent table FOO (if a table has more than one LOB column, LOB segment names are generated differently, use dba|user_lobs view to get parent table association). SEGMENT_NAME SEGMENT_TYPE FOO TABLE SYS_IL0000009792C00002$$ LOBINDEX SYS_LOB0000009792C00002$$ LOBSEGMENT (also referred as LOB chunks/pages) The LOBSEGMENT and the LOBINDEX segments are stored in the same tablespace as the table containing the LOB, unless otherwise specified.[1] Important Storage Parameters This section defines the important storage parameters of a LOB column (or a LOB attribute) - . ?fFor each definition we describe the effects of the parameter, and give recommendations for on how to get better performance and to avoid errors. CHUNK Definition CHUNK is the smallest unit of LOBSEGMENT allocation. It is a multiple of DB_BLOCK_SIZE. Points to Note ? For example, if the value of CHUNK is 8K and an inserted LOB is only 1K in size, then 1 chunk is allocated and 7K are wasted in that chunk. The CHUNK option does NOT affect in-line LOBs (see the definition in the next section) ? Choose an appropriate chunk size for best performance also to avoid space wastage. The maximum chunk size is 32K. ? The CHUNK parameter cannot be altered. Recommendation Choose a chunk size for optimal performance and minimum space wastage. For LOBs that are less than 32K, a chunk size that is 60% (or more) of the LOB size is a good starting point. For LOBs larger than 32K, choose a chunk size equal to the frequent update size. In-line and Out-of-Line storage: ENABLE STORAGE IN ROW and DISABLE STORAGE IN ROW Definition LOB storage is said to be Inin-line when the LOB data is stored with the other column data in the row. A LOB can only be stored inline if its size is less than ~4000 bytes. For in-line LOB data, space is allocated in the table segment (the LOBINDEX and LOBSEGMENT segments are empty). LOB storage is said to be out-of-line when the LOB data is stored , in CHUNK sized blocks in the LOBSEGMENT segment, separate from the other columns? data. ENABLE STORAGE IN ROW allows LOB data to be stored in the table segment provided it is less than ~4000 bytes. DISABLE STORAGE IN ROW prevents LOB data from being stored in-line, regardless of the size of the LOB. Instead only a 20-byte LOB locator is stored with the other column data in the table segment. Points to Note ? In-line LOBs are subject to normal chaining and row migration rules within Oracle. If you store a 3900 byte LOB in a row with 2K block size then the row will be chained across two or more blocks. Both REDO and UNDO are written for in-line LOBs as they are part of the normal row data. The CHUNK option does not affect in-line LOBs. ? With out-of-line storage, UNDO is written only for the LOB locator and LOBINDEX changes. No UNDO is generated for chunks/pages in the LOBSEGMENT. Consistent Read is achieved by using page versions (see the RETENTION or PCTVERSION options). ? DML operations on out-of-line LOBs can generate high amounts of redo information, because redo is generated for the entire chunk. For example, in the extreme case, ?DISABLE STORAGE IN ROW CHUNK 32K? would write redo for the whole 32K even if the LOB changes were was only 5 bytes. ? When in-line LOB data is updated, and if the new LOB size is greater than 3964 bytes, then it is migrated and stored out-of-line. If this migrated LOB is updated again and its size becomes less than 3964 bytes, it is not moved back in-line (except when we use LONG API for update). ? ENABLE|DISABLE STORAGE IN ROW parameters cannot be altered. Recommendation Use ENABLE STORAGE IN ROW, except in cases where the LOB data is not retrieved as much as other columns? data. In this case, if the LOB data is stored out-of-line, the biggest gain is achieved while performing full table scans, as the operation does not retrieve the LOB?s data. CACHE, NOCACHE Definition The CACHE storage parameter causes LOB data blocks to be read/written via the buffer cache. With the NOCACHE storage parameter, LOB data is read/written using direct reads/writes. This means that the LOB data blocks are never in the buffer cache and the Oracle server process performs the reads/writes. Points to Note ? With the CACHE option, LOB data reads show up as wait event ?db file sequential read?, writes are performed by the DBWR process. With the NOCACHE option, LOB data reads/writes show up as wait events direct path read (lob)?/?direct path write (lob)?. Corresponding statistics are ?physical reads direct (lob)? and ?physical writes direct (lob)?. ? In-line LOBs are not affected by the CACHE option as they reside with the other column data, which is typically accessed via the buffer cache. ? The CACHE option gives better read/write performance than the NOCACHE option. ? The CACHE option for LOB columns is different from the CACHE option for tables. This means that caution is required otherwise the read of a large LOB can effectively flush the buffer cache. ? The CACHE|NOCACHE option can be altered. Recommendation Enable caching, except for cases where caching LOBs would severely impact performance for other online users, by forcing these users to perform disk reads rather than getting cache hits. Consistent Reads on LOBs: RETENTION and PCTVERSION Consistent Read (CR) on LOBs uses a different mechanism than that used for other data blocks in Oracle. Older versions of the LOB are retained in the LOB segment and CR is used on the LOB index to access these older versions (for in-line LOBs which are stored in the table segment, the regular UNDO mechanism is used). There are two ways to control how long older versions are maintained. Definition ? RETENTION ? time-based: this specifies how long older versions are to be retained. ? PCTVERSION ? space-based: this specifies what percentage of the LOB segment is to be used to hold older versions. Points to Note ? RETENTION is a keyword in the LOB column definition. No value can be specified for RETENTION. The RETENTION value is implicit,.. If a LOB is created with database compatibility set to 9.2.0.0 or higher, undo_management=TRUE and PCTVERSION is not explicitly specified, time-based retention is used. The LOB RETENTION value is always equal to the value of the UNDO_RETENTION database instance parameter. ? You cannot specify both PCTVERSION and RETENTION. ? PCTVERSION is applicable only to LOB chunks/pages allocated in LOBSEGMENTS. Other LOB related data in the table column and the LOBINDEX segment use regular undo mechanism. ? PCTVERSION=0: the space allocated for older versions of LOB data in LOBSEGMENTS can be reused by other transactions and can cause ?snapshot too old? errors. ? PCTVERSION=100: the space allocated by older versions of LOB data can never be reused by other transactions. LOB data storage space is never reclaimed and it always increases. ? RETENTION and PCTVERSION can be altered Recommendation Time-based retention using the RETENTION keyword is preferred. A high value for RETENTION or PCTVERSION may be needed to avoid ?snapshot too old? errors in environments with high concurrent read/write LOB access. LOGGING, NOLOGGING Definition LOGGING: enables logging of LOB data changes to the redo logs. NOLOGGING: changes to LOB data (stored in LOBSEGMENTs) are not logged into the redo logs, however in-line LOB changes are still logged as normal. Points to Note ? The CACHE option implicitly enables LOGGING. ? If NOLOGGING was set, and if you have to recover the database, then sections of the LOBSEGMENT will be marked as corrupt during recovery (LOBINDEX changes are logged to redo logs and are recovered, but the corresponding LOBSEGMENTs are not logged for recovery). ? LOGGING|NOLOGGING can be altered. The NOCACHE option is required to turn off LOGGING, e.g. (NOCACHE NOLOGGING). Recommendation Use NOLOGGING only when doing bulk loads or migrating from LONG to LOB. Backup is recommended after bulk operations. Performance GUIDELINE LOB Loading In the rest of the document, you will notice LOB API and LONG API methods being referenced many times. The difference between these APIs is as follows: LOB API: the LOB data is accessed by first selecting the LOB locator. LONG API: the LOB data is accessed without using the LOB locator. Points to Note Use array operations for LOB inserts Scalability problem with LOB disable storage in row option BUG 3180333 - LOB LOADING USING SQLLDR DOESN'T SCALE Problem scenario: 2 (or more) concurrent sqlldr processes trying to load LOB data (LOB column defined with DISABLE STORAGE IN ROW). Loading will run almost serially. Serialization point is getting a CR copy of the LOBINDEX block. Workaround: use ENABLE STORAGE IN ROW even for LOBs whose size is greater than 3964 bytes. With ENABLE STORAGE IN ROW, we store the first 12 chunk addresses in the table row and if the inserted LOB data size can be addressed within these first 12 chunk addresses, then LOBINDEX is empty. Generating a CR version of a table block is more efficient and,,, in some cases, not required. This code path provides much better scalability. Please note that if LOB data is larger than 12 chunk addresses, then we may see CR contention with the ENABLE STORAGE IN ROW option as well. Row Chaining problem with the use of OCILobWrite API TAR 2760194.995 (UK) - LOADING SMALL (AVG LEN 1120) CLOB DATA INTO TABLE PRODUCES MUCH CHAINING, WHY? Problem scenario: in 10gR1 (and older releases), SQL*Loader uses OCILobWrite API for LOB loading. This leads to a row chaining problem, as described below: CREATE TABLE foo (pkey NUMBER NOT NULL, bar BLOB); Load 3 rows with LOB data size as 3700, 3000 and 3400 respectively. SQL*Loader loads the LOB columns, first by inserting empty_blob, and second, by writing the LOB data using the LOB locator. In the first step, the average row length is pkey length + empty_blob length= 4 + 40 bytes = ~44 bytes. Assuming that DB_BLOCK_SIZE=8192, these 3 rows can be inserted into one data block. In the second step, loading LOB data, the 1st row, 3700 bytes of LOB, and the 2nd row, 3000 bytes of LOB, can be inserted into the same block. However, for the 3rd row of LOB data, there is no space left in that block, so the row must be chained. Workaround: the first workaround could be to increase the value of PCTFREE. It may help solve this problem, but it unnecessarily wastes space. The second workaround is to write a loader program using the LONG API method (please note that an enhancement request against sqlldr component is filed for this problem, and there is a plan to fix it in the future release). High number of consistent read blocks created and examined BUG 3297800 - SQLLDR MAY NEED TO USE LONG API INTERFACE FOR LOBS LESS THAN 2GB Problem scenario: 2 (or more) concurrent sqlldr processes loading LOB data in conventional mode. Using the LOB API method for loading the LOB data in a single user environment may also cause a high number of CR block creation to occur. As mentioned earlier, loading the LOB data is performed in 2 steps. . In the first step, sqlldr inserts empty_blob for LOB columns. Then, with this LOB locator, the LOB data is written using an OCILobWrite call. In a multi-user loading environment, before OCILobWrite is invoked, if other loading processes change the data block, it may be required to examine the block and, if required, a CR version of the block is created. Workaround: None, other than writing a loader program using he LONG API method CPU time and Elapsed time - not reported accurately BUG 3504487 - DBMS_LOB/OCILob* CALL RESOURCE USAGE IS NOT REPORTED, AS THEY ARE NOT PART OF A CURSOR Problem scenario: the work done using LOB API calls is not part of the cursor, so reporting resource usage while collecting statistics for the LOB workload, such as the CPU time or the elapsed time, may not be accurate. Example to illustrate this situation: (We have already a table created as: CREATE TABLE foo (pkey NUMBER, bar BLOB);) Declare lob_loc blob; buffer raw(32767); lob_amt binary_integer := 16384; begin buffer := utl_raw.cast_to_raw(rpad('FF', 32767, 'FF')); for j in 1..10000 loop select bar into lob_loc from foo where pkey = j for update; dbms_lob.write(lob_loc, lob_amt, 1, buffer ); commit; end loop; dbms_output.put_line ('Write test finished '); end; / After executing the above PL/SQL, query V$SQL to measure cpu_time and elapsed time resource usage. select sql_text, cpu_time/100000, elapsed_time/100000 from v$sql where sql_text like '%foo%' or sql_text like ?%dbms_lob%?; SQL_TEXT ------------------------------------------------------------------------------------------------ CPU_TIME/1000000 ELAPSED_TIME/1000000 -------------------------- ----------------------------------- declare lob_loc blob; buffer raw(32767); lob_amt binary_integer := 16384 ; begin buffer := utl_raw.cast_to_raw(rpad('FF', 32767, 'FF')); for j in 1..10000 loop select bar into lob_loc from foo where pkey = j for update; dbms_lob.write(lob_loc, lob_amt, 1, buffer ); commit; end loop; dbms_output.put_line ('Write test finished '); end; 19.54 19.28 SELECT bar from foo where pkey = :b1 for update 5.00 4.81 As you can see, the PL/SQL block took about 19.54 seconds in CPU time and 19.28 seconds in elapsed time respectively. Out of 19.54 secondss , the SELECT statement contributed to 5.00 seconds, so the remaining 14 seconds (approximately) were spent in dbms_lob.write. This is not reported, because the work done by dbms_lob.write is not part of a cursor. Similarly OCILOB API calls were not part of a cursor as well. Workaround: None Reads/Writes are done one chunk at a time in synchronous way BUG 3437770 - LOB DIRECT PATH READ/WRITES ARE LIMITED BY CHUNK SIZE Problem scenario: The Oracle server process does NOCACHE LOB reads/writes using a direct path mechanism. The limitation here is that reads/writes are done one chunk at a time in a synchronous way. Consider the example below: Assuming CHUNK size=8K, DB_BLOCK_SIZE=2k, LOB data = 64K, 8 writes are done (each doing 4 blocks of write at a time) to load the entire LOB data, waiting for each write to complete before issuing another write. Workaround: use as many loader processes as possible to maximize disk throughput. High CPU system time BUG 3437770 - LOB DIRECT PATH READ/WRITES ARE LIMITED BY CHUNK SIZE This is probably due to the above limitation (reads/writes are done one chunk at time in synchronous way) Buffer cache sizing problem Problem scenario: loading LOB data with the CACHE option will most likely fill up even a large buffer cache. Under this condition, a degradation in the load rate can be seen if the database writer doesn?t keep up with the foreground free buffer requests. Workaround: follow the general instance tuning guidelines - use asynchronous I/O (if not possible, use multiple db writer processes) - stripe datafiles across many spindles - use the NOCACHE option The CACHE option will also force other online users to perform physical disk reads. This can be avoided by using multiple block sizes. For example, keep online user objects in 4k (or 8k) block size tablespace and and cached LOB data in 8kK (or 16k) block size tablespace. Allocate the required amount of buffer cache for each block sizes (e.g. db_4k_block_buffer=500M, db_8k_block_buffer=2000M) Multi-byte character set conversion BUG 3324897 - LOBS LESS THAN 3964 BYTES ARE STORED OUT-OF-LINE WHILE LOADING USING SQLLDR Problem scenario: wWhen dealing with multi-byte character set, additional bytes are required for CLOB data. This may cause client side CLOB data of ~ 4000 bytes, being stored out-of-line in the database. Workaround: None Use array operations for LOB inserts HWM enqueue contention BUG 3537749 - HW ENQUEUE CONTENTION WHEN LOADING LOB DATA Problem scenario: given the large size of LOB data (compare to relational table row size), blocks under HWM are filled rapidly (under high concurrent load condition) and can cause HW enqueue contention. Workaround: ASSM with larger extent size may help. RAC environment issues BUG 3429986 - CONVENTIONAL LOAD OF LOB FROM 2 RAC NODE DO NOT SCALE DUE TO LOG FLUSH LATENCIES Problem scenario: In a RAC environment, when loading LOB data into one partition, you may notice contention on 1st level bitmap and LOB header segment with ASSM. You may notice the same contention on a single instance (with a large number of CPUs) with a high number of concurrent loaders. Workaround: loading into separate partitions will avoid this situation. If this is not possible, use range-hash partition instead of just range partitions. FREEPOOLS should help in this situation, but we need to do more testing to see the effect of this parameter.but didn?t provide any improvement in our testing. Other LOB performance related issues BUG 3234751 - EXCESSIVE USAGE OF TEMP TS WHILE LOADING LOB USING SQLLDR IN CONVENTIONAL MODE BUG 3230541 - LOB LOADING USING SQLLDR DIRECT PATH SLOWER THAN CONVENTIONAL BUG 3189083 - OPEN/CLOSE OF DATAFILE FOR EVERY LOB CHUNK WRITEWRITES APPENDIX A APPENDIX A LONG API access to LOB datatype Oracle provides transparent access to LOBs from applications that use LONG and LONG RAW datatypes. If your application uses DML (INSERT, UPDATE, DELETE) statements from OCI or PL/SQL (PRO*C etc) for LONG or LONG RAW data, no application changes are required after the column is converted to a LOB. For example, you can SELECT a CLOB into a character variable, or a BLOB into a RAW variable. You can define a CLOB column as SQLT_CHR or a BLOB column as SQLT_BIN and select the LOB data directly into a CHARACTER or RAW buffer without selecting out the locator first. The following example demonstrates this concept: create table foo ( pkey number(10) not null, bar long raw ); set serveroutput on declare in_buf raw(32767); out_buf raw(32767); out_pkey number; begin in_buf := utl_raw.cast_to_raw (rpad('FF', 32767, 'FF')); for j in 1..10 loop insert into foo values (j, in_buf) ; commit; end loop; dbms_output.put_line ('Write test finished '); for j in 1..10 loop select pkey, bar into out_pkey, out_buf from foo where pkey=j ; end loop; dbms_output.put_line ('Read test finished '); end; / Now migrate LONG RAW column to BLOB column alter table foo modify (bar blob); That works. alter table foo modify (bar long raw); ERROR at line 1: ORA-22859: invalid modification of columns So that does not work. There are few things customer should note when doing the LONG to LOB migration. This alter table migration statement runs serially in 9i. i (what about 8i,10g). Indexes need to be rebuilt and statistics recollected. After the LONG to LOB migration, the above PL/SQL block will work without any modifications. Advanced LOB features may require the use of the LOB API, described in the Oracle Documentation[2] APPENDIX B Migration FROM from in-line to out-of-line (and out-of-line to in-line) STORAGE This section explains one major difference between the LOB API and LONG API methods. If a change to the in-line LOB data makes it larger than 3964 bytes, then it is automatically moved out of table segment and stored out-of-line. If during future operations, the LOB data shrinks to under 3964 bytes, it will remain out-of-line. In other words, once a LOB is migrated out, it is always stored out-of-line irrespective of its size, with the following exception scenario. Consider a scenario where you used the LONG API to update the LOB datatype [..] begin in_buf := utl_raw.cast_to_raw (rpad('FF', 3964, 'FF')); insert into foo values (1, in_buf) ; commit; [..] Above LOB is stored in-line, update the LOB to a size more than 3964 bytes [..] in_buf := utl_raw.cast_to_raw (rpad('FF', 4500, 'FF')); update foo set bar=buffer where pkey=1; commit; [..] After the update LOB is stored out-of-line, now update the LOB to a size smaller than 3964 bytes [..] in_buf := utl_raw.cast_to_raw (rpad('FF', 3000, 'FF')); update foo set bar=buffer where pkey=1; commit; [..] LOB is stored in-line again. When using the LONG API for update, the older LOB is deleted (or space is reclaimed as per RETENTION or PCTVERSION setting) and a new LOB is created, with a new LOB locator. This is different from using LOB API, where DML on LOB is possible only using the LOB locator (the LOB locator doesn?t change) APPENDIX C How LOB data is stored The purpose of this section is to differentiate how the ENABLE STORAGE IN ROW option is different from the DISABLE STORAGE IN ROW option for LOB data size greater than 3964 bytes. It also highlights customers when LOBINDEX is really used (following example scenarios assume Solaris OS and Oracle 9204 32 bit version).. In-line LOB LOB size less than 3964 bytes LOB can be NULL, EMPTY_BLOB, and actual LOB data create table foo ( pkey number(10) not null, bar BLOB ) lob (bar) store as (enable storage in row chunk 2k); declare inbuf raw(3964); begin inbuf := utl_raw.cast_to_raw(rpad('FF', 3964, 'FF')); insert into foo values (1, NULL); insert into foo values (2, EMPTY_BLOB() ); insert into foo values (3, inbuf ); commit; end; / note: RPAD ('-', 60, '-')==>'------------------------------------------------------------' Now Foo table rows are: Pkey=1 Bar=0 byte (nothing is stored) Pkey=2 Bar=36 byte (10 byte metadata + 10 byte LobId + 16 byte Inode) Pkey=3 Bar=4000 byte (36 byte + 3964 byte of LOB data, nothing stored in LOBINDEX and LOBSEGMENT LobId - LOB Locator In-line LOB ? LOB size = 3965 bytes (1 byte greater than 3964) LOB is defined as in-line, but actual data is greater than 3964 bytes, so moved out ? please note this is different from LOB being defined as out-of-line. [..] inbuf := utl_raw.cast_to_raw(rpad('FF', 3965, 'FF')); insert into foo values (4, inbuf ); [..] Foo table row Pkey=4 Bar=40 bytes (36 byte + 4 byte for one chunk RDBA). Using this RDBA, we directly access LOB data in LOBSEGMENT. Nothing stored in LOBINDEX RDBA ? Relative Database Block Address In-line LOB ? LOB size greater than 12 chunk addresses With in-line LOB option, we store the first 12 chunk addresses in the table row. This takes 84 bytes (36+4*12) of size in table row. LOBs that are less than 12 chunks in size will not have entries in the LOBINDEX if ENABLE STORAGE IN ROW is used [..] inbuf := utl_raw.cast_to_raw(rpad('FF', 32767, 'FF')); insert into foo values (5, inbuf ); [..] Here, we are inserting 32767 bytes of LOB data, given our chunk size of 2k, we need approximately 16 blocks (32767/2048). So we store first 12 chunk RDBAs in table row and the rest in LOBINDEX Foo table row Pkey=5 Bar=84 bytes (36 byte + 4*12 byte for first 12 chunk RDBA). Using this RDBA, we directly access 12 LOB chunks in LOBSEGMENT. Then using the LobId, we lookup LOBINDEX to get rest of the LOB chunk RDBAs. Out-of-line LOBs ? All LOB sizes With out-of-line LOB option, only LOB locator is stored in table row. Using LOB locator, we lookup LOBINDEX and find the range of chunk RDBAs, using this RDBAs we read LOB data from LOBSEGMENT create table foo (pkey number(10) not null, bar BLOB) lob (bar) store as (disable storage in row chunk 2k); [..] inbuf := utl_raw.cast_to_raw(rpad('FF', 20, 'FF')); insert into foo values (6, inbuf ); [..] Foo table rows Pkey=6 Bar=20 bytes (10 byte metadata + 10 byte LobId). Please note Inode and chunk RDBAs are stored in LOBINDEX. LOB Performance Guidelines April 2004 Author: V. Jegraj (Vinayagam.Djegaradjane) Acknowledgements: Vishy Karra, Krishna Kunchithapadam, Cecilia Gervasio Oracle Corporation World Headquarters 500 Oracle Parkway Redwood Shores, CA 94065 U.S.A. Worldwide Inquiries: Phone: +1.650.506.7000 Fax: +1.650.506.7200 www.oracle.com Copyright ? 2004 Oracle Corporation All rights reserved. -------------------------------------------------------------------------------- [1] In Oracle8i, users can specify storage parameters for LOB index, but from Oracle9i Database onwards, specifying storage parameters for a LOB index is ignored without any error and the index is stored in the same tablespace as the LOB segment, with an Oracle generated index name. [2] Large Objects (LOBs) in Oracle9i Application Developer's Guide, DBMS_LOB package in Oracle9i Supplied PL/SQL Packages and Types Reference, LOB and FILE Operations in Oracle Call Interface Programmer?s guide . -------------------------------------------------------------------------------- Copyright © 2005, Oracle. All rights reserved. Legal Notices and Terms of Use. Note 7: ======= Doc ID: Note:1071540.6 Content Type: TEXT/PLAIN Subject: Converting a Long datatype to Clob in Oracle8i? Creation Date: 27-MAY-1999 Type: BULLETIN Last Revision Date: 24-JUN-2004 Status: PUBLISHED PURPOSE This note describes the Oracle 8.1.x function that converts data stored in LONG and LONG RAW datatypes to CLOB and BLOB datatypes respectively. This is done using the TO_LOB function. Converting a long datatype to a Clob: ========================================= The TO_LOB function is provided in Oracle 8.1.x to convert LONG and LONG RAW datatypes to CLOB and BLOB datatypes respectively. Note: The TO_LOB function is not provided in Oracle 8.0.x. Oracle recommends that long datatypes be converted to CLOBs, NCLOB or BLOBs. Note: When a LOB is stored in a table, the data (LOB VALUE) and a pointer to the data called a LOB LOCATOR, are stored separately. The data may be stored along with the locator in the table itself or in a separate table. The LOB clause in the create table command can be used to specify whether an attempt should be made to store data in the main table or a separate one. The LOB clause may also be used to specify a separate tablespace and storage clause for both the LOB table and its associated index. Example: SQL> create table long_data (c1 number, c2 long); Table created. SQL> desc long_data Name Null? Type ------------------------------- -------- ---- C1 NUMBER C2 LONG SQL> insert into long_data values 2 (1, 'This is some long data to be migrated to a CLOB'); 1 row created. Note: The TO_LOB function may be used in CREATE TABLE AS SELECT or INSERT...SELECT statements: Example: SQL> create table test_lobs 2 (c1 number, c2 clob); Table created. SQL> desc test_lobs Name Null? Type ------------------------------- -------- ---- C1 NUMBER C2 CLOB SQL> insert into test_lobs 2 select c1, to_lob(c2) from long_data; 1 row created. SQL> select c2 from test_lobs; C2 ----------------------------------------------- This is some long data to be migrated to a CLOB References: =========== Oracle8i SQL Reference Volume 1 [NOTE:66046.1] Oracle8i: LOBs 30.2 How to access LOB data: ============================ 30.2.1 SQL DML: --------------- Using SQL DML for Basic Operations on LOBs SQL DML provides basic operations -- INSERT, UPDATE, SELECT, DELETE -- that let you make changes to the entire values of internal LOBs within the Oracle ORDBMS. To work with parts of internal LOBs, you will need to use one of the interfaces that have been developed to handle more complex requirements. Oracle8 supports read-only operations on external LOBs. So if you need to update/write to external LOBs, you will have to develop client side applications suited to your needs Suppose you have the following table: create table multimedia_tab ( clip_id number, story clob, flsub nclob, photo bfile, frame blob, sound blob, voiced_ref voiced_type, inseg_ntab inseg_type, music bfile, map_obj map_typ ); create table multimedia_tab ( clip_id number, story clob, flsub nclob, photo bfile, frame blob, sound blob, music bfile ); The following INSERT statement populates story with the character string 'JFK interview', sets flsub, frame and sound to an empty value, sets photo to NULL, and initializes music to point to the file 'JFK_interview' located under the logical directory 'AUDIO_DIR' (see the CREATE DIRECTORY command in the Oracle8i Reference. Character strings are inserted using the default character set for the instance. INSERT INTO Multimedia_tab VALUES (101, 'JFK interview', EMPTY_CLOB(), NULL, EMPTY_BLOB(), EMPTY_BLOB(), NULL, NULL, BFILENAME('AUDIO_DIR', 'JFK_interview'), NULL); Similarly, the LOB attributes for the Map_typ column in Multimedia_tab can be initialized to NULL or set to empty as shown below. Note that you cannot initialize a LOB object attribute with a literal. INSERT INTO Multimedia_tab VALUES (1, EMPTY_CLOB(), EMPTY_CLOB(), NULL, EMPTY_BLOB(), EMPTY_BLOB(), NULL, NULL, NULL, Map_typ('Moon Mountain', 23, 34, 45, 56, EMPTY_BLOB(), NULL); SELECTing a LOB Performing a SELECT on a LOB returns the locator instead of the LOB value. In the following PL/SQL fragment you select the LOB locator for story and place it in the PL/SQL locator variable Image1 defined in the program block. When you use PL/SQL DBMS_LOB functions to manipulate the LOB value, you refer to the LOB using the locator. DECLARE Image1 BLOB; ImageNum INTEGER := 101; BEGIN SELECT story INTO Image1 FROM Multimedia_tab WHERE clip_id = ImageNum; DBMS_OUTPUT.PUT_LINE('Size of the Image is: ' || DBMS_LOB.GETLENGTH(Image1)); /* more LOB routines */ END; DECLARE Image1 BLOB; ImageNum INTEGER := 101; BEGIN SELECT content INTO Image1 FROM binaries2 WHERE id = 1211; DBMS_OUTPUT.PUT_LINE('Size of the Image is: ' || DBMS_LOB.GETLENGTH(Image1)); /* more LOB routines */ END; / XXX So you can retrieve all kinds of info with DBMS_LOB 30.2.2 The EMPTY_BLOB and EMPTY_CLOB functions: ----------------------------------------------- The EMPTY_BLOB function returns an empty locator of type BLOB (binary large object). The specification for the EMPTY_BLOB function is: FUNCTION EMPTY_BLOB RETURN BLOB; You can call this function without any parentheses or with an empty pair. Here are some examples: INSERT INTO family_member (name, photo) VALUES ('Steven Feuerstein', EMPTY_BLOB()); DECLARE my_photo BLOB := EMPTY_BLOB; BEGIN Use EMPTY_BLOB to initialize a BLOB to "empty." Before you can work with a BLOB, either to reference it in SQL DML statements such as INSERTs or to assign it a value in PL/SQL, it must contain a locator. It cannot be NULL. The locator might point to an empty BLOB value, but it will be a valid BLOB locator. The EMPTY_CLOB function returns an empty locator of type CLOB. The specification for the EMPTY_CLOB function is: FUNCTION EMPTY_CLOB RETURN CLOB; You can call this function without any parentheses or with an empty pair. Here are some examples: INSERT INTO diary (entry, text) VALUES (SYSDATE, EMPTY_CLOB()); DECLARE the_big_novel CLOB := EMPTY_CLOB; BEGIN Use EMPTY_CLOB to initialize a CLOB to "empty". Before you can work with a CLOB, either to reference it in SQL DML statements such as INSERTs or to assign it a value in PL/SQL, it must contain a locator. It cannot be NULL. The locator might point to an empty CLOB value, but it will be a valid CLOB locator. 30.2.3 DBMS_LOB --------------- Simple example to get the length of a lob: DECLARE Image1 BLOB; ImageNum INTEGER := 101; BEGIN SELECT content INTO Image1 FROM binaries2 WHERE id = 1211; DBMS_OUTPUT.PUT_LINE('Size of the Image is: ' || DBMS_LOB.GETLENGTH(Image1)); /* more LOB routines */ END; / DBMS_LOB The DBMS_LOB package provides subprograms to operate on BLOBs, CLOBs, NCLOBs, BFILEs, and temporary LOBs. You can use DBMS_LOB to access and manipulation specific parts of a LOB or complete LOBs. DBMS_LOB can read and modify BLOBs, CLOBs, and NCLOBs; it provides read-only operations for BFILEs. The bulk of the LOB operations are provided by this package. Example: Load Text Files to CLOB then Write Back Out to Disk - (PL/SQL) Overview The following example is part of the Oracle LOB Examples Collection. This example provides two PL/SQL procedures that demonstrate how to populate a CLOB column with a text file (an XML file) then write it back out to the file system as a different file name. - Load_CLOB_From_XML_File: This PL/SQL procedure loads an XML file on disk to a CLOB column using a BFILE reference variable. Notice that I use the new PL/SQL procedure DBMS_LOB.LoadCLOBFromFile(), introduced in Oracle 9.2, that handles uploading to a multi-byte UNICODE database. - Write_CLOB_To_XML_File: This PL/SQL procedure writes the contents of the CLOB column in the database piecewise back to the file system. Let's first take a look at an example XML file: DatabaseInventoryBig.xml: - - production.iDevelopment.info production iDevelopment.info Jeffrey Hunter The following database should be considered the most stable for up-to-date data. The backup strategy includes running the database in Archive Log Mode and performing nightly backups. All new accounts need to be approved by the DBA Group before being created. - development.iDevelopment.info development iDevelopment.info Jeffrey Hunter Melody Hunter The following database should contain all hosted applications. Production data will be exported on a weekly basis to ensure all development environments have stable and current data. - testing1.iDevelopment.info testing1 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host more than half of the testing for our hosting environment. - testing2.iDevelopment.info testing2 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the HR department only. - testing3.iDevelopment.info testing3 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the Finance department only. - testing4.iDevelopment.info testing4 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the HQ department only. - testing5.iDevelopment.info testing5 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the Engineering department only. - testing6.iDevelopment.info testing6 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the IT department only. - testing7.iDevelopment.info testing7 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the Marketing department only. - testing8.iDevelopment.info testing8 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the Purchasing department only. - testing9.iDevelopment.info testing9 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the Accounts Payable department only. - testing10.iDevelopment.info testing10 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing OEM. - testing11.iDevelopment.info testing11 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing XMLDB. - testing12.iDevelopment.info testing12 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for tuning. - testing13.iDevelopment.info testing13 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for UAT. - testing14.iDevelopment.info testing14 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for additional monitoring. - testing15.iDevelopment.info testing15 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing upgrades. - testing16.iDevelopment.info testing16 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for certification tesing. - testing17.iDevelopment.info testing17 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing18.iDevelopment.info testing18 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing19.iDevelopment.info testing19 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing20.iDevelopment.info testing20 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing21.iDevelopment.info testing21 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing22.iDevelopment.info testing22 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing23.iDevelopment.info testing23 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. + testing24.iDevelopment.info testing24 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. + testing25.iDevelopment.info testing25 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. + testing26.iDevelopment.info testing26 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. + testing27.iDevelopment.info testing27 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. + testing28.iDevelopment.info testing28 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing29.iDevelopment.info testing29 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing30.iDevelopment.info testing30 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing31.iDevelopment.info testing31 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing32.iDevelopment.info testing32 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing33.iDevelopment.info testing33 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing34.iDevelopment.info testing34 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing35.iDevelopment.info testing35 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing36.iDevelopment.info testing36 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing37.iDevelopment.info testing37 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing38.iDevelopment.info testing38 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing39.iDevelopment.info testing39 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing40.iDevelopment.info testing40 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing41.iDevelopment.info testing41 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing42.iDevelopment.info testing42 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing43.iDevelopment.info testing43 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing44.iDevelopment.info testing44 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing45.iDevelopment.info testing45 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing46.iDevelopment.info testing46 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing47.iDevelopment.info testing47 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing48.iDevelopment.info testing48 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing49.iDevelopment.info testing49 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing50.iDevelopment.info testing50 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing51.iDevelopment.info testing51 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing52.iDevelopment.info testing52 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing53.iDevelopment.info testing53 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing54.iDevelopment.info testing54 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing55.iDevelopment.info testing55 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing56.iDevelopment.info testing56 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing57.iDevelopment.info testing57 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing58.iDevelopment.info testing58 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing59.iDevelopment.info testing59 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the DBA department for testing of all ERP application modules. - testing60.iDevelopment.info testing60 iDevelopment.info Jeffrey Hunter Melody Hunter Alex Hunter The following database will host a testing database to be used by the Sales Force Automation department. dest_clob -- , SRC_LOB => src_clob -- , AMOUNT => DBMS_LOB.GETLENGTH(src_clob) -- ); -- -- ----------------------------------------------------------------------- INSERT INTO test_clob(id, file_name, xml_file, timestamp) VALUES(1001, 'DatabaseInventoryBig.xml', empty_clob(), sysdate) RETURNING xml_file INTO dest_clob; -- ------------------------------------- -- OPENING THE SOURCE BFILE IS MANDATORY -- ------------------------------------- DBMS_LOB.OPEN(src_clob, DBMS_LOB.LOB_READONLY); DBMS_LOB.LoadCLOBFromFile( DEST_LOB => dest_clob , SRC_BFILE => src_clob , AMOUNT => DBMS_LOB.GETLENGTH(src_clob) , DEST_OFFSET => dst_offset , SRC_OFFSET => src_offset , BFILE_CSID => DBMS_LOB.DEFAULT_CSID , LANG_CONTEXT => lang_ctx , WARNING => warning ); DBMS_LOB.CLOSE(src_clob); COMMIT; DBMS_OUTPUT.PUT_LINE('Loaded XML File using DBMS_LOB.LoadCLOBFromFile: (ID=1001).'); END; / SQL> @load_clob_from_xml_file.sql Procedure created. CREATE OR REPLACE PROCEDURE Write_CLOB_To_XML_File IS clob_loc CLOB; buffer VARCHAR2(32767); buffer_size CONSTANT BINARY_INTEGER := 32767; amount BINARY_INTEGER; offset NUMBER(38); file_handle UTL_FILE.FILE_TYPE; directory_name CONSTANT VARCHAR2(80) := 'EXAMPLE_LOB_DIR'; new_xml_filename CONSTANT VARCHAR2(80) := 'DatabaseInventoryBig_2.xml'; BEGIN DBMS_OUTPUT.ENABLE(100000); -- ---------------- -- GET CLOB LOCATOR -- ---------------- SELECT xml_file INTO clob_loc FROM test_clob WHERE id = 1001; -- -------------------------------- -- OPEN NEW XML FILE IN WRITE MODE -- -------------------------------- file_handle := UTL_FILE.FOPEN( location => directory_name, filename => new_xml_filename, open_mode => 'w', max_linesize => buffer_size); amount := buffer_size; offset := 1; -- ---------------------------------------------- -- READ FROM CLOB XML / WRITE OUT NEW XML TO DISK -- ---------------------------------------------- WHILE amount >= buffer_size LOOP DBMS_LOB.READ( lob_loc => clob_loc, amount => amount, offset => offset, buffer => buffer); offset := offset + amount; UTL_FILE.PUT( file => file_handle, buffer => buffer); UTL_FILE.FFLUSH(file => file_handle); END LOOP; UTL_FILE.FCLOSE(file => file_handle); END; / SQL> @write_clob_to_xml_file.sql Procedure created. Now lets test it: SQL> set serveroutput on SQL> exec Load_CLOB_From_XML_File Loaded XML File using DBMS_LOB.LoadCLOBFromFile: (ID=1001). PL/SQL procedure successfully completed. SQL> exec Write_CLOB_To_XML_File PL/SQL procedure successfully completed. SQL> SELECT id, DBMS_LOB.GETLENGTH(xml_file) Length FROM test_clob; ID LENGTH ---------- ---------- 1001 41113 SQL> host ls -l DatabaseInventory* -rw-r--r-- 1 oracle dba 41113 Sep 20 15:02 DatabaseInventoryBig.xml -rw-r--r-- 1 oracle dba 41113 Sep 20 15:48 DatabaseInventoryBig_2.xml 30.2.4 REMOTE SELECTS, INSERTS, UPDATES: ---------------------------------------- Valid operations on LOB columns in remote tables include: CREATE TABLE as select * from table1@remote_site; INSERT INTO t select * from table1@remote_site; UPDATE t set lobcol = (select lobcol from table1@remote_site); INSERT INTO table1@remote... UPDATE table1@remote... DELETE table1@remote... 30.2.5: Export a BLOB to a file with Java: ------------------------------------------ First we create a Java stored procedure that accepts a file name and a BLOB as parameters: CREATE OR REPLACE JAVA SOURCE NAMED "BlobHandler" AS import java.lang.*; import java.sql.*; import oracle.sql.*; import java.io.*; public class BlobHandler { public static void ExportBlob(String myFile, BLOB myBlob) throws Exception { // Bind the image object to the database object // Open streams for the output file and the blob File binaryFile = new File(myFile); FileOutputStream outStream = new FileOutputStream(binaryFile); InputStream inStream = myBlob.getBinaryStream(); // Get the optimum buffer size and use this to create the read/write buffer int size = myBlob.getBufferSize(); byte[] buffer = new byte[size]; int length = -1; // Transfer the data while ((length = inStream.read(buffer)) != -1) { outStream.write(buffer, 0, length); outStream.flush(); } // Close everything down inStream.close(); outStream.close(); } }; / ALTER java source "BlobHandler" compile; show errors java source "BlobHandler" Next we publish the Java call specification so we can access it via PL/SQL: CREATE OR REPLACE PROCEDURE ExportBlob (p_file IN VARCHAR2, p_blob IN BLOB) AS LANGUAGE JAVA NAME 'BlobHandler.ExportBlob(java.lang.String, oracle.sql.BLOB)'; / Next we grant the Oracle JVM the relevant filesystem permissions: EXEC Dbms_Java.Grant_Permission( - 'SCHEMA-NAME', - 'java.io.FilePermission', - '<>', - 'read ,write, execute, delete'); Finally we can test it: CREATE TABLE tab1 (col1 BLOB); INSERT INTO tab1 VALUES(empty_blob()); COMMIT; DECLARE v_blob BLOB; BEGIN SELECT col1 INTO v_blob FROM tab1; ExportBlob('c:\MyBlob',v_blob); END; / 30.2.6 Import into a BLOB from a file: -------------------------------------- Import BLOB Contents The following article presents a simple methods for importing a file into a BLOB datatype. First a directory object is created to point to the relevant filesystem directory: CREATE OR REPLACE DIRECTORY images AS 'C:\'; Next we create a table to hold the BLOB: CREATE TABLE tab1 (col1 BLOB); Finally we import the file into a BLOB datatype and insert it into the table: DECLARE v_bfile BFILE; v_blob BLOB; BEGIN INSERT INTO tab1 (col1) VALUES (empty_blob()) RETURN col1 INTO v_blob; v_bfile := BFILENAME('IMAGES', 'MyImage.gif'); Dbms_Lob.Fileopen(v_bfile, Dbms_Lob.File_Readonly); Dbms_Lob.Loadfromfile(v_blob, v_bfile, Dbms_Lob.Getlength(v_bfile)); Dbms_Lob.Fileclose(v_bfile); COMMIT; END; / Hope this helps. Regards Tim... 30.2.7 Import into a CLOB from a file: -------------------------------------- Import CLOB Contents The following article presents a simple methods for importing a file into a CLOB datatype. First a directory object is created to point to the relevant filesystem directory: CREATE OR REPLACE DIRECTORY documents AS 'C:\'; Next we create a table to hold the CLOB: CREATE TABLE tab1 (col1 CLOB); Finally we import the file into a CLOB datatype and insert it into the table: DECLARE v_bfile BFILE; v_clob CLOB; BEGIN INSERT INTO tab1 (col1) VALUES (empty_clob()) RETURN col1 INTO v_clob; v_bfile := BFILENAME('DOCUMENTS', 'Sample.txt'); Dbms_Lob.Fileopen(v_bfile, Dbms_Lob.File_Readonly); Dbms_Lob.Loadfromfile(v_clob, v_bfile, Dbms_Lob.Getlength(v_bfile)); Dbms_Lob.Fileclose(v_bfile); COMMIT; END; / Hope this helps. Regards Tim... Note 5: ------- You Asked (Jump to Tom's latest followup) I have a table with a blob column. It's possible to specify an extra storage clause for this column ? and we said... Yes, the following example is cut and pasted from the SQL Reference Manual, the CREATE TABLE command: CREATE TABLE lob_tab (col1 BLOB, col2 CLOB) STORAGE (INITIAL 512 NEXT 256) LOB (col1, col2) STORE AS (TABLESPACE lob_seg_ts STORAGE (INITIAL 6144 NEXT 6144) CHUNK 4 NOCACHE LOGGING INDEX (TABLESPACE lob_index_ts STORAGE (INITIAL 256 NEXT 256) ) ); The table will be stored in the users default tablespace with (INITIAL 512 NEXT 256). The actual lob data will be in LOB_SEG_TS with (INITIAL 6144 NEXT 6144). The lob index built on the pages constituting the lob will be stored in yet a 3'rd tablespace -- lob_index_ts with (INITIAL 256 NEXT 256). Reviews GOTO a page to Bookmark Review | Bottom | Top lob storage recovery May 07, 2004 Reviewer: bob from PA Tom, If the LOB tablespace is not backed up, can the table data (a different tablespace) be recovered in a failure scenario? I know with TSPITR the process validates that no objects cross tablespaces that are not included in the set being recovered with the TSPITR check/validate funcion. This doesn't mean the tablespace won't be recovered in the auxiliary db, it just means the automated process won't continue through to export the objects, and re-import unless you pass the check. (or at least that was what happened in the test I ran). I am just curious about what would happen to this table if its lob tablespace was lost and non-recoverable. can just the regular data be recovered Followup: well, it's going to be problematic as the lob locators will point to "garbage". You cannot really TSPITR a table with lobs without doing the *same* to the lob segments. You'd have to sort of update the lobs to NULL and pull it manually -- but then I would ask "why have the lobs in the first place, must not be very important"? so yes, we'd be able to get the scalar data back (complete recovery would be best here), update the lob to null and go forward with that. 30.3 Errors in LOB: =================== 30.3.1: ------- Doc ID: Note:293515.1 Subject: ORA-1578 ORA-26040 in a LOB segment - Script to solve the errors Type: PROBLEM Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 09-DEC-2004 Last Revision Date: 25-FEB-2005 Purpose ============ - The purpose of this article is to provide a script to solve errors ORA-1578 / ORA-26040 when a lob block is accessed by a sql statement. - Note that the data inside the corrupted lob blocks is not salvageable. This procedure will update the lob column with an empty lob to avoid errors ORA-1578 / ORA-26040. - After applying this solution dbverify would still produce error DBV-200 until block marked as corrupted is reused and reformatted. Symptoms =========== - ORA-1578 and ORA-26040 are produced when accesing a lob column in a table: ORA-1578 : ORACLE data block corrupted (file # %s, block # %s) ORA-26040: Data block was loaded using the NOLOGGING option - dbverify for the datafile that produces the errors fails with: DBV-00200: Block, dba , already marked corrupted Example: dbv file=/oracle/oradata/data.dbf blocksize=8192 DBV-00200: Block, dba 54528484, already marked corrupted ..... The dba can be used to get the relative file number and block number: Relative File number: SQL> select dbms_utility.data_block_address_file(54528484) from dual; DBMS_UTILITY.DATA_BLOCK_ADDRESS_FILE(54528484) ---------------------------------------------- 13 Block Number: SQL> select dbms_utility.data_block_address_block(54528484) from dual; DBMS_UTILITY.DATA_BLOCK_ADDRESS_BLOCK(54528484) ----------------------------------------------- 2532 Cause ========== - LOB segment has been defined as NOLOGGING - LOB Blocks were marked as corrupted by Oracle after a datafile restore / recovery. Identify the table referencing the lob segment - Example ========================================================= Error example when accessing the lob column by a sql statement: ORA-01578 : ORACLE data block corrupted (file #13 block # 2532) ORA-01110 : datafile 13: '/oracle/oradata/data.dbf' ORA-26040 : Data block was loaded using the NOLOGGING option. 1. Query dba_extents to find out the lob segment name select owner, segment_name, segment_type from dba_extents where file_id = 13 and 2532 between block_id and block_id + blocks - 1; In our example it returned: owner=SCOTT segment_name=SYS_LOB0000029815C00006$$ segment_type=LOBSEGMENT 2. Query dba_lobs to identify the table_name and lob column name: select table_name, column_name from dba_lobs where segment_name = 'SYS_LOB0000029815C00006$$' and owner = 'SCOTT'; In our example it returned: table_name = EMP column_name = EMPLOYEE_ID_LOB Fix ====== 1. Identify the table rowid's referencing the corrupted lob segment blocks by running the following plsq script: rem ********************* Script begins here ******************** create table corrupted_data (corrupted_rowid rowid); set concat # declare error_1578 exception; pragma exception_init(error_1578,-1578); n number; begin for cursor_lob in (select rowid r, &&lob_column from &table_owner.&table_with_lob) loop begin n:=dbms_lob.instr(cursor_lob.&&lob_column,hextoraw('8899')) ; exception when error_1578 then insert into corrupted_data values (cursor_lob.r); commit; end; end loop; end; / undefine lob_column rem ********************* Script ends here ******************** When prompted by variable values and following our example: Enter value for lob_column: EMPLOYEE_ID_LOB Enter value for table_owner: SCOTT Enter value for table_with_lob: EMP 2. Update the lob column with empty lob to avoid ORA-1578 and ORA-26040: SQL> set concat # SQL> update &table_owner.&table_with_lob set &lob_column = empty_blob() where rowid in (select corrupted_rowid from corrupted_data); if &lob_column is a CLOB datatype, replace empty_blob by empty_clob. Reference ============== Note 290161.1 - The Gains and Pains of Nologging Operations 30.3.2: ------- Displayed below are the messages of the selected thread. Thread Status: Closed From: Neil Bullen 26-Mar-02 08:26 Subject: How do you alter NOLOGGING in lob index partition RDBMS Version: 8.1.7.2.1 Operating System and Version: Compaq Tru64 Unix 5.2 Error Number (if applicable): Product (i.e. SQL*Loader, Import, etc.): Product Version: How do you alter NOLOGGING in lob index partition I have discovered that a lob index partition is set to NOLOGGING, how can I alter this to LOGGING. The lob is set to CACHE and LOGGING, the index def_logging is set to NONE and the tablespace is set to LOGGING. -------------------------------------------------------------------------------- From: Oracle, Rowena Serna 02-Apr-02 03:26 Subject: Re : How do you alter NOLOGGING in lob index partition You could find the system generated lobindex name and use the "alter index" command. Regards, Rowena Serna Oracle Corporation ------------------------------------------------------------------------------- From: Neil Bullen 03-Apr-02 23:42 Subject: Re : How do you alter NOLOGGING in lob index partition Using alter index on a lob segment index results in error ORA-22864 cannot ALTER or DROP LOB indexes, the solution I found was to alter the lob caching setting, even though dba_lobs showed the CACHE and LOGGING settings to be 'YES' by issuing the ALTER TABLE MODIFY LOB() (CACHE); command all partitions of the associated index were changed to LOGGING. What threw me was the CACHE and LOGGING settings in dba_lobs already being set correctly, however resetting these again was the key. -------------------------------------------------------------------------------- From: Oracle, Rowena Serna 09-Apr-02 02:46 Subject: Re : How do you alter NOLOGGING in lob index partition Thanks for updating. Regards, Rowena Serna Oracle Corporation 30.3.4 exp/imp errors and LOBS: ------------------------------- Note 1: ------- Doc ID: Note:48023.1 Subject: OERR: IMP 64 Definition of LOB was truncated by export Type: REFERENCE Status: PUBLISHED Content Type: TEXT/PLAIN Creation Date: 07-NOV-1997 Last Revision Date: 26-MAR-2001 Error: IMP 64 Text: Definition of LOB was truncated by export --------------------------------------------------------------------------- Cause: While producing the dump file, Export was unable to write the * entire contents of a LOB. Import is therefore unable to * reconstruct the contents of the LOB. The remainder of the * import of the current table will be skipped. Action: Delete the offending row in the exported database and retry the * export. . Note 2: ------- An export or import of a table with a Large Object (LOB) column, has a slower performance than an export or import of a table without LOB columns: -- create two tables: TESTTAB1 with a VARCHAR2 column, and TESTTAB2 with a -- CLOB column: connect / as sysdba create table scott.testtab1 (nr number, txt varchar2(2000)); create table scott.testtab2 (nr number, txt clob); -- populate both tables with the same 500,000 rows: declare x varchar2(50); begin for i in 1..500000 loop x := 'This is a line with the number: ' || i; insert into scott.testtab1 values(i,x); insert into scott.testtab2 values(i,x); commit; end loop; end; / -- export both tables: % exp system/manager file=exp_testtab1.dmp tables=scott.testtab1 direct=y % exp system/manager file=exp_testtab1a.dmp tables=scott.testtab1 % exp system/manager file=exp_testtab2.dmp tables=scott.testtab2 No CLOB No CLOB With CLOB DIRECT CONVENTIONAL column ------------ ------------ ------------ 8.1.7.4.0 0:13 0:20 7:49 9.2.0.4.0 0:14 0:18 7:37 9.2.0.5.0 0:12 0:15 7:03 10.1.0.2.0 0:16 0:31 7:15 Note 3: ------- Doc ID: Note:157024.1 Content Type: TEXT/X-HTML Subject: Insert/Import of Table with Lob Fails IMP-00003 ORA-3237 Creation Date: 24-MAY-2001 Type: PROBLEM Last Revision Date: 21-OCT-2003 Status: PUBLISHED fact: Oracle Server - Enterprise Edition fact: Import Utility (IMP) symptom: Import fails with error symptom: Insert fails symptom: Table with LOB column symptom: Locally managed tablespace symptom: IMP-00003: ORACLE error %lu encountered symptom: IMP-00017: following statement failed with ORACLE error %lu: symptom: ORA-03237: Initial Extent of specified size cannot be allocated cause: Extent size specified for the tablespace is not large enough. fix: For LOBS, ensure that the extent size specification in the tablespace is least three times the db_block_size. For example: If the db_block_size is 8192, then the extent size for the tablespace should be at least 24576. Explaination: Certain objects may require larger extents by virtue of how they are built internally (Example: an RBS requires at least four blocks and a LOB at least three). References: SQL Reference Guide, Create Tablespace Note 4: ------- Doc ID: Note:211721.1 Content Type: TEXT/X-HTML Subject: Unable to Import Tables with LOB Columns Creation Date: 13-SEP-2002 Type: PROBLEM Last Revision Date: 03-OCT-2003 Status: PUBLISHED fact: Oracle Server - Enterprise Edition 9+- fact: Oracle Server - Enterprise Edition 8.1 fact: Oracle Server - Enterprise Edition 8 fact: Import Utility (IMP) symptom: Import fails symptom: ORA-01658: unable to create INITIAL extent for segment in tablespace '%s' symptom: ORA-01652: unable to extend temp segment by %s in tablespace %s symptom: Table contains LOB column symptom: Problem does not occur for tables without LOB columns cause: No LOB storage specifications were specified on the table creation for those tables with LOB columns. LOB data is stored both within and outwith the table depending on how much data the column contains. A new database was created and the data reimported into a tablespace with 1.7GB default initial extent size. The LOB storage outwith the table defaults to the initial extent of the tablespace and this storage requirement could not be fulfilled. fix: As a user with dba privileges issue alter tablespace default storage (initial M)& #059; where and are replaced with appropriate values. See also : Note:1074731.6 ORA-01658 During 'Create Table' Statement Note 5: ------- Doc ID: Note:197699.1 Subject: "IMP-00003 ORA-00959 ON IMPORT OF TABLE WITH CLOB DATATYPES" Type: PROBLEM Status: PUBLISHED Content Type: TEXT/PLAIN Creation Date: 31-MAY-2002 Last Revision Date: 29-AUG-2002 Problem Description ------------------- You are attempting to import a table that has CLOB datatype and you receive the following errors: IMP-00003: ORACLE error 959 encountered ORA-00959: tablespace does not exist Solution Description -------------------- Create the table that has CLOB datatypes before the import, specifying tablespaces that exist on the target system, and import using IGNORE=Y. Here is a simple example where you can get this problem and how to resolve it: I have a user "TEST" with default tablespace has "USERS" Step-1: Create tst Tablespace ================================= SQL> create tablespace tst datafile 'c:\temp\tst1.dbf' size 5m; Tablespace created. Step-2: Create table with CLOB datatype by login to "TEST" user ================================================================= SQL> CREATE TABLE "TEST"."PX2000" ("ID" NUMBER(*,0), "SUBMITDATE" DATE, "COMMENTS" VARCHAR2(4000),"RECOMMENDEDTIMELONG" CLOB) PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 STORAGE(INITIAL 65536 FREELISTS 1 FREELIST GROUPS 1) TABLESPACE "TST" LOGGING LOB ("RECOMMENDEDTIMELONG") STORE AS (TABLESPACE "TST" ENABLE STORAGE IN ROW CHUNK 8192 PCTVERSION 10 NOCACHE STORAGE(INITIAL 65536 FREELISTS 1 FREELIST GROUPS 1)) ; SQL> select table_name,tablespace_name from user_tables 2 where table_name='PX2000'; TABLE_NAME TABLESPACE_NAME ------------------------------ ------------------------------ PX2000 TST SQL> select username,default_tablespace from user_users; USERNAME DEFAULT_TABLESPACE ------------------------------ ------------------------------ TEST USERS Step-3: Export the Table ========================= exp test/test file=px2000.dmp tables=px2000 . . exporting table PX2000 0 rows exported Step-4: Drop the "TST" tablespace including contents: Please note that 'AND datafiles' is a new option in version 9i. Omit this clause if running version prior to 9i. ============================================================ SQL> drop tablespace tst including contents and datafiles; Tablespace dropped. Step-5: Import the table back to test schema ============================================== imp test/test file=px2000.dmp tables=px2000 IMP-00017: following statement failed with ORACLE error 959: "CREATE TABLE "PX2000" ("ID" NUMBER(*,0), "SUBMITDATE" DATE, "COMMENTS" VARC" "HAR2(4000), "RECOMMENDEDTIMELONG" CLOB) PCTFREE 10 PCTUSED 40 INITRANS 1 M" "AXTRANS 255 STORAGE(INITIAL 65536 FREELISTS 1 FREELIST GROUPS 1) TABLESPACE" " "TST" LOGGING LOB ("RECOMMENDEDTIMELONG") STORE AS (TABLESPACE "TST" ENAB" "LE STORAGE IN ROW CHUNK 8192 PCTVERSION 10 NOCACHE STORAGE(INITIAL 65536 F" "REELISTS 1 FREELIST GROUPS 1))" IMP-00003: ORACLE error 959 encountered ORA-00959: tablespace 'TST' does not exist Import terminated successfully with warnings. Step-6: Workaround is to extract the DDL from the dumpfile,change the tablespace to target database. Create the table manually and import with ignore=y option ================================================================================== % imp test/test file=px2000.dmp full=y show=y log= Step-7: Use the logFile to pre-create the table, then ignore object creation errors. ==================================================================================== % imp test/test file=px2000.dmp full=y ignore=y Explanation ----------- For most of the DDL's (except for Partitioned tables,tables without CLOB datatypes), i mport will automatically create the objects to the users default tablespace if the specified tablespace does not exist. DDL's with tables with CLOB datatypes and partitioned tables an IMP-00003 and ORA-00959 will result if the tablespace does not exists in target database. References ---------- [NOTE:1058330.6] "IMP-00003 ORA-00959 ON IMPORT OF PARTITIONED TABLE" [BUG:1982168] "IMP-3 / ORA-959 importing table with CLOB using IGNORE=Y into variable width charset DB" [BUG:2398272] "IMPORT TABLE WITH CLOB DATATYPE FAILS WITH IMP-3 AND ORA-959" Oracle Utilites Manual . Note 6: ------- Displayed below are the messages of the selected thread. Thread Status: Closed From: Helmut Daiminger 12-Dec-00 21:50 Subject: MOVE table with LOB column to another tablespace RDBMS Version: 8.1.6.1.2 Operating System and Version: Win2k, SP1 Error Number (if applicable): Product (i.e. SQL*Loader, Import, etc.): Product Version: MOVE table with LOB column to another tablespace Hi! I'm having a problem here: I want to move a table with a LOB column (i.e. LOB index segment) to a different tablespace. In the beginning the table and the LOB segment were in the USERS tablespace. I then exported the table using the EXP tool. Then I revoked the user's quota to the USERS tablespace and only gave him quota on the default tablespace. Then I run IMP and import that LOB-table. The table gets recreated in the new tablespace, but the creation of the LOB index fails with an error message that I don't have privileges to write to the USERS tablespace. How do I completey move the table and the LOB index segment to a new tablespace? This is 8.1.6 on Windows 2000 Server. Thanks, Helmut From: Oracle, Ken Robinson 14-Dec-00 21:05 Subject: Re : MOVE table with LOB column to another tablespace I believe you can do the following: ALTER TABLE foo MOVE TABLESPACE new_tbsp STORAGE(new_storage) LOB (lobcol) STORE AS lobsegment (TABLESPACE new_tbsp STORAGE (new_storage)); Regards, Ken Robinson Oracle Server EE Analyst Note 7. ------- Doc ID: Note:176898.1 Content Type: TEXT/X-HTML Subject: Import Fails with IMP-00032 and IMP-00008 Creation Date: 15-FEB-2002 Type: PROBLEM Last Revision Date: 24-JUN-2003 Status: PUBLISHED fact: Oracle Server - Enterprise Edition fact: Import Utility (IMP) symptom: IMP-00032: SQL statement exceeded buffer length symptom: IMP-00008: unrecognized statement in the export file cause: The insert statement run when importing exceeds the default or specified buffer size. For import of tables containing LONG, LOB, BFILE, REF, ROWID, LOGICALROWID or type columns, rows are inserted individually. The size of the buffer must be large enough to contain the entire row inserted. fix: Increase the buffer size, and make sure that it is big enough to contain the biggest row in the table(s) imported. For example: imp system/manager file=test.dmp full=y log=test.log buffer= 10000000 Note 8: ------- For tables with LOB columns, make sure the tablespace already exists in the target database before the import is done. Also, make sure the extent size is large enough. Note 9: ------- With imp/exp I hit a problem that on remote database users tablespace is called 'users', while on local it's 'users_data'. Now I have to go to documentation to figure out if those stupid switches would save the day... Also with schlobs the elegant insert into t2 select * from t1@remote_db_link; doesn't work. I wonder why export/import is not plain sqlplus statements where I can just specify the right 'where' clause... Followup: Yes, when you deal with multi segment objects (tables with LOBS, partitioned table, IOTs with overflows for example), using EXP/IMP is complicated if the target database doesn't have the same tablespace structure. That is because the CREATE statement contains many tablespaces and IMP will only "rewrite" the first TABLESPACE in it (it will not put multi-tablespace objects into a single tablespace, the object creation will fail of the tablespaces needed by that create do not exist). I dealt with this issue in my book, in there, I recommend you do an: imp .... full=y indexfile=temp.sql In temp.sql, you will have all of the DDL for indexes and tables. Simply delete all index creates and uncomment any table creates you want. Then, you can specify the tablespaces for the various components -- precreate the objects and run imp with ignore=y. The objects will now be populated. You are incorrect with the "schlobs" comment (both in spelling and in conclusion). scott@ORA815.US.ORACLE.COM> create table t ( a int, b blob ); Table created. scott@ORA815.US.ORACLE.COM> desc t Name Null? Type ----------------------------------- -------- ------------------------ A NUMBER(38) B BLOB scott@ORA815.US.ORACLE.COM> select a, dbms_lob.getlength(b) from t; no rows selected scott@ORA815.US.ORACLE.COM> insert into t select x, y from t@ora8i.world; 1 row created. scott@ORA815.US.ORACLE.COM> select a, dbms_lob.getlength(b) from t; A DBMS_LOB.GETLENGTH(B) ---------- --------------------- 1 1000011 So, the "elegant insert into select * from" does work. imp/exp can be plain sqlplus statements -- use indexfile=y (if you get my book, I use this over and over in various places to get the DDL). In 9i, there is a simple stored procedure interface as well. Note 10: -------- Tom Without using the export import( show=y) Is there any query to find out in which Tablespace the LOB column is stored Thanks in advance Followup: select * from user_segments you can join user_segments to user_lobs if you like as well. user_segments will give you tablespace info. user_lobs will give you the lob segment name. Note 11: -------- IMP-00003 ORACLE error number encountered Cause: Import encountered the referenced Oracle error. Action: Look up the Oracle message in the ORA message chapters of this manual, and take appropriate action. IMP-00020 long column too large for column buffer size (number) Cause: The column buffer is too small. This usually occurs when importing LONG data. Action: Increase the insert buffer size 10,000 bytes at a time (for example). Use this step-by-step approach because a buffer size that is too large may cause a similar problem. IMP-00064 Definition of LOB was truncated by export Cause: While producing the dump file, Export was unable to write the entire contents of a LOB. Import is therefore unable to reconstruct the contents of the LOB. The remainder of the import of the current table will be skipped. Action: Delete the offending row in the exported database and retry the export. IMP-00070 Lob definitions in dump file are inconsistent with database. Cause: The number of LOBS per row in the dump file is different than the number of LOBS per row in the table being populated. Action: Modify the table being imported so that it matches the column attribute layout of the table that was exported. Note 12: -------- we have a 10 Mill rows table with a BLOB column in it the size of the lob varies: from 1K up ward to a few megabytes, but most are in the 2K-3K range. So currently, we have ENABLE STORAGE IN ROW. and want to do DISABLE STORAGE IN ROW b/c we are starting to do a lot of range scan on the table. When we export/import the table and during import have moved all the lobs out of line.. the total space used during the import bloated 5 times from a 2GIG tablespace into a 10GIG tablespace??? Why? The database block size is 8K, running 9.2.0.6 with auto sgement management in the tablespace CREATE TABLESPACE "BLOB_DATA" LOGGING DATAFILE 'D:ORACLEORADATATESTDBBLOB_DATA01.ora' SIZE 2048M REUSE AUTOEXTEND OFF EXTENT MANAGEMENT LOCAL UNIFORM SIZE 8M SEGMENT SPACE MANAGEMENT AUTO Note 13: -------- To relocate tables using lobs: Method 1: ========= 1. export data using exp cmd 2. drop all tables 3. create a new LOB tablespace 4. re-create all the tables with the LOB Storage clause, for example create table FOO ( col1 NUMBER ,col2 BLOB ) tablespace DATA_TBLSPCE LOB ( col2 ) STORE AS col2_blob ( tablespace BLOB_TBLSPCE disable storage in row chunk 8192 pctversion 10 cache storage (initial 64K next 64K minextents 1 maxextents unlimited pctincrease 0 ) 5. import data with ignore=y Method 2: ======= Doc ID: Note:130814.1 Subject: How to move LOB Data to Another Tablespace Type: HOWTO Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 19-DEC-2000 Last Revision Date: 05-AUG-2003 Purpose ------- The purpose of this article is to provide the syntax for altering the storage parameters of a table that contains one or more LOB columns. Scope & Application ------------------- This article will be useful for Oracle DBSs, Developers, and Support Analysts. How to move LOB Data to Another Tablespace ------------------------------------------ If you want to make no other changes to the table containing a lob other than to rebuild it, use: ALTER TABLE foo MOVE; This will rebuild the table segment. It does not affect any of the lob segments associated with the lob columns which is the desired optimization. If you want to change one or more of the physical attibutes of the table containing the lob, however no attributes of the lob columns are to be changed, use the following syntax: ALTER TABLE foo MOVE TABLESPACE new_tbsp STORAGE(new_storage); This will rebuild the table segment. It does not rebuild any of the lob segments associated with the lob columns which is the desired optimization. If a table containing a lob needs no changes to the physical attributes of the table segment, but you want to change one or more lob segments; for example, you want to move the lob column to a new tablespace as well as the lob's storage attributes, use the following syntax: ALTER TABLE foo MOVE LOB(lobcol) STORE AS lobsegment (TABLESPACE new_tbsp STORAGE (new_storage)); Note that this will also rebuild the table segment (although, in this case, in the same tablespace and without changing the table segment physical attributes). If a table containing a lob needs changes to both the table attributes as well as the lob attributes then use the following syntax: ALTER TABLE foo MOVE TABLESPACE new_tbsp STORAGE(new_storage) LOB (lobcol) STORE AS lobsegment (TABLESPACE new_tbsp STORAGE (new_storage)); Explanation ----------- The 'ALTER TABLE foo MODIFY LOB (lobcol) ...' syntax does not allow for a change of tablespace ALTER TABLE my_lob MODIFY LOB (a_lob) (TABLESPACE new_tbsp); (TABLESPACE new_tbsp) * ORA-22853: invalid LOB storage option specification You have to use the MOVE keyword instead as shown in the examples. References ---------- Note 66431.1 LOBS - Storage, Redo and Performance Issues Bug 747326 ALTER TABLE MODIFY LOB STORAGE PARAMETER DOES'T WORK Additional Search Words ----------------------- ora-1735 ora-906 ora-2143 ora-22853 clob nclob blob Method 3: ========= Move doesnt support Long datatypes. You can either convert them to LOBs and then move or do exp/imp of the table with the LONG column or create the table with LONG in the locally managed tablespace and copy the data from the old table using PL/SQL loop or CTAS with to_lob in the locally managed tablespace.. SQL> desc t Name Null? Type ----------------------------------------- -------- ---------------------------- X NUMBER(38) Y LONG SQL> alter table t move; alter table t move * ERROR at line 1: ORA-00997: illegal use of LONG datatype -- You can create the new table in the Locally Managed tablespace SQL> create table t_lob tablespace users as select x,to_lob(y) y from t; Table created. SQL> desc t_lob Name Null? Type ----------------------------------------- -------- ---------------------------- X NUMBER(38) Y CLOB -- Now you can drop the old table and rename the new table -- Or you can move the LOB table to the locally managed tablespace SQL> alter table t_lob move; Table altered. -- Or you can precreate the new table with LONG in the locally managed tablespace and do exp/imp -- export the Long table SQL> !exp / file=t.dmp tables=t compress=n Export: Release 9.2.0.3.0 - Production on Tue Mar 2 09:32:30 2004 Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved. Connected to: Oracle9i Enterprise Edition Release 9.2.0.3.0 - Production With the Partitioning, OLAP and Oracle Data Mining options JServer Release 9.2.0.3.0 - Production Export done in WE8ISO8859P1 character set and AL16UTF16 NCHAR character set About to export specified tables via Conventional Path ... . . exporting table T 2 rows exported Export terminated successfully without warnings. -- just rename the old table for reference purposes SQL> rename t to tbak; Table renamed. -- Create the LONG table in the locally managed tablespace SQL> create table t(x int,y long) tablespace users; Table created. -- now import the data SQL> !imp / file=t.dmp tables=t ignore=y Import: Release 9.2.0.3.0 - Production on Tue Mar 2 09:33:43 2004 Copyright (c) 1982, 2002, Oracle Corporation. All rights reserved. Connected to: Oracle9i Enterprise Edition Release 9.2.0.3.0 - Production With the Partitioning, OLAP and Oracle Data Mining options JServer Release 9.2.0.3.0 - Production Export file created by EXPORT:V09.02.00 via conventional path import done in WE8ISO8859P1 character set and AL16UTF16 NCHAR character set . importing OPS$ORACLE's objects into OPS$ORACLE . . importing table "T" 2 rows imported Import terminated successfully without warnings. SQL> desc t Name Null? Type ----------------------------------------- -------- ---------------------------- X NUMBER(38) Y LONG Note 14: -------- Doc ID : Note:225337.1 Content Type: TEXT/PLAIN Subject: ORA-22285 ON ACCESSING THE BFILE COLUMN OF A TABLE Creation Date: 08-JAN-2003 Type: PROBLEM Last Revision Date: 17-DEC-2004 Status: PUBLISHED Fact(s) ~~~~~~~ *The directory alias for the relevant directory exists. *This condition might be encountered in general or particularly after successful export/import of 'table with bfile column' from one schema to another. *Non-bfile columns of the table could be accessed but not the bfile column. Symptom(s) ~~~~~~~~~~ Accessing the bfile column of table gives the following errors: ORA-22285: non-existent directory or file for ..... ORA-06512: at "SYS.DBMS_LOB", line ... Diagnosis: ~~~~~~~~~~ -- create the exporting user schema and the table with bfile data-- SQL>connect system/manager SQL>create user test2 identified by test2 default tablespace users quota 50 m on users / SQL>grant connect, create table, create any directory to test2 / SQL>conn test2/test2 SQL>create table test_lobs ( c1 number, c2 clob, c3 bfile, c4 blob ) LOB (c2) STORE AS (ENABLE STORAGE IN ROW) LOB (c4) STORE AS (DISABLE STORAGE IN ROW) / create two files (rec2.txt , rec3.txt) using OS utilities in some directory say ( /tmp ) --create the directory alias -- SQL>create directory tmp_dir as '/tmp' / -- Populate the table-- SQL>insert into test_lobs values (1,null,null,null) / SQL>insert into test_lobs values (2,EMPTY_CLOB(),BFILENAME('TMP_DIR','rec2.txt'),EMPTY_BLOB()) / SQL>insert into test_lobs values (3,'Some data for record3.', BFILENAME('TMP_DIR','rec2.txt'), '48656C6C6F'||UTL_RAW.CAST_TO_RAW('there!')) / -- access the table-- SQL>column len_c2 format 9999 SQL>column len_c3 format 9999 SQL>column len_c4 format 9999 SQL>select c1, DBMS_LOB.GETLENGTH(c2) len_c2, DBMS_LOB.GETLENGTH(c3) len_c3, DBMS_LOB.GETLENGTH(c4) len_c4 from test_lobs / C1 LEN_C2 LEN_C3 LEN_C4 -------------------------- ------ ------ ------ 1 2 0 124 0 3 22 124 11 -- carry out the schema level export-- $ exp system/manager file=exp44.dmp log=logr44.log owner=test2 IMPORTING DATABASE: create same two files (rec2.txt , rec3.txt) using OS utilities in some directory say ( /tmp ) --create the directory alias -- SQL>conn system/manager SQL>create directory tmp_dir as '/tmp' / -- create the importing user schema-- SQL>create user test3 identified by test3 default tablespace users quota 50 m on users / SQL>grant connect, create table, create any directory to test3 / --carry out the successful schema level import-- $ imp system/manager fromuser=test2 touser=test3 file=exp44.dmp log=log44.log --try to access the imported table as below (same statement as by the exporting user-- SQL>select c1, DBMS_LOB.GETLENGTH(c2) len_c2, DBMS_LOB.GETLENGTH(c3) len_c3, DBMS_LOB.GETLENGTH(c4) len_c4 from test_lobs / ERROR: ORA-22285: non-existent directory or file for GETLENGTH operation ORA-06512: at "SYS.DBMS_LOB", line 547 -- However non bfile columns could be accessed-- Cause ~~~~~ The importing user lacks the read access on the corresponding directory/ directory alias. Solution(s) ~~~~~~~~~~~ grant read access on the corressponding directory to the user who tries to access the bfile table as below: SQL> conn system/manager Connected. SQL> grant read on directory tmp_dir to test3; ( please see the example above ) Once the read permission is granted ,the bfile column of the said table is accessible since the corresponding directory (/alias) is accessible. Refrences: ~~~~~~~~~~ [NOTE:66046.1] : LOBs, Longs, and other Datatypes Note 15: -------- Doc ID: Note:279722.1 Subject: IMPORT OF TABLE WITH LOB GENERATES CORE DUMP Type: PROBLEM Status: MODERATED Content Type: TEXT/X-HTML Creation Date: 31-JUL-2004 Last Revision Date: 02-AUG-2004 The information in this article applies to: Oracle Server - Enterprise Edition - Version: 9.2.0.3 This problem can occur on any platform. Symptoms IMPORT OF TABLE WITH LOB GENERATES CORE DUMP Cause Importing a table having a clob created with chunksize = 32k Error Details: ------------- . importing DBAPIDB1's objects into DBAPIDB1 . . importing table "TE2006"Segmentation fault Trace from the Core Dump: ------------------------ lmmstrmlr 44 lmmstmrg D4 lmmstmrg D4 lmmstfree 104 lmmfree C0 impmfr 24 impplb 5BC impins 22B8 do_insert 48C imptabwrk F4 impdta 41C impdrv 2D68 main 14 __start 94 Fix FIX: --- Apply the patch for Bug:3091499 WORKAROUND: ---------- Before import, create the table with chunksize <= 16K and run import setting ignore=y References - Import Of Table With Lob Generates Core Dump Note 16: keep LOBS at manageble size. ------------------------------------- (1) Look at PCTVERSION: Since the LOB segments are usually very large, they are treated differently from other columns. While other columns can be guaranteed to give consistent reads, these columns are not. This is because, it is difficult to manage with LOB data rollback segments due to their size unlike other columns. So they do not use rollback segments. Usually only one copy exists, so the queries reading that column may not get consistent reads while other queries modify them. In these cases, the other queries will get "ORA-22924 snapshot too old" errors. To maintain read consistency Oracle creates new LOB page versions every time a lob changes. PCTVERSION is the percentage of all used LOB data space that can be occupied by old versions of LOB data pages. As soon as old versions of LOB data pages start to occupy more than the PCTVERSION amount of used LOB space, Oracle tries to reclaim the old versions and reuse them. In other words, PCTVERSION is the percent of used LOB data blocks that is available for versioning old LOB data. The PCTVERSION can be set to the percentage of LOB's that are occasionally updated. Often a table's a LOB column usually gets the data uploaded only once, but is read multiple times. Hence it is not necessary to keep older versions of LOB data. It is recommended that this value be changed to "0". By default PCTVERSION is set to 10%. So, most of the instances usually have it set to 10%, it must be set to 0% explicitly. The value can be changed any time in a running system. Use the following query to find out currently set value for PCTVERSION: SQL> select PCTVERSION from dba_lobs where TABLE_NAME = 'table_name' and COLUMN_NAME='column_name'; PCTVERSION ---------- 10 PCTVERSION can be changed using the following SQL (it can be run anytime in a running system): ALTER TABLE FND_LOBS MODIFY LOB (FILE_DATA) ( PCTVERSION 0 ); Note 17: difference 9iR1 9iR2 with respect to Locally managed tablespace ------------------------------------------------------------------------ Doc ID: Note:159078.1 Subject: Cannot Create Table with LOB Column in Locally Managed Tablespace Type: PROBLEM Status: PUBLISHED Content Type: TEXT/X-HTML Creation Date: 26-SEP-2001 Last Revision Date: 04-AUG-2004 fact: Oracle Server - Enterprise Edition 9.0.1 symptom: Creating new OEM repository fails symptom: Create table SMP_LMV_SEARCH_OBJECT fails symptom: ORA-03001: unimplemented feature symptom: Table with LOB column cause: You try to create a LOB segment in a bitmapped (locally managed) tablespace. This is a limitation for bitmapped segments in 9i. This is being documented in the SQL Reference- the restriction will be lifted in 9i Release 2. fix: Create the table in a tablespace that was created with clause SEGMENT SPACE MANAGEMENT MANUAL Note 18: -------- In a trace file you either get ORA-00600: internal error code, arguments: [kkdoilsn1], [], [], [], [], [], [], [] or ORA-00600: internal error code, arguments: [15265], [], [], [], [], [], [], [] description: in a 9.2 database, a table with lob and indexsegments was moved to another tablespace. Explanation: 9202 2405258 Dictionary corruption / OERI:15265 from MOVE LOB to existing segment name 2405258 Dictionary corruption / OERI:15265 from MOVE LOB to existing segment name This is Bug 2405258 Fixed: 9202 Corruption LOB Related (CLOB/BLOB/BFILE) Dictionary corruption / ORA-600 [15265] from MOVE LOB toexisting segment name. Eg: ALTER TABLE mytab MOVE LOB (artist_bio) STORE AS lobsegment (STORAGE(INITIAL 1M NEXT 1M)); corrupts the dictionary if "logsegment" already exists. Bug 2405258 Dictionary corruption / OERI:15265 from MOVE LOB to existing segment name This note gives a brief overview of bug 2405258. Affects: Product (Component) Oracle Server (RDBMS) Range of versions believed to be affected Versions >= 8 but < 10G Versions confirmed as being affected 9.2.0.1 Platforms affected Generic (all / most platforms affected) Fixed: This issue is fixed in 9.2.0.2 (Server Patch Set) 10G Production Base Release Symptoms: Corruption (Dictionary) Internal Error may occur (ORA-600) ORA-600 [15265] Related To: Datatypes - LOBs (CLOB/BLOB/BFILE) Description Dictionary corruption / ORA-600 [15265] from MOVE LOB to existing segment name. Eg: ALTER TABLE mytab MOVE LOB (artist_bio) STORE AS lobsegment (STORAGE(INITIAL 1M NEXT 1M)); corrupts the dictionary if "logsegment" already exists. ===================== 31. BLOCK CORRUPTION: ===================== Note 1: ======= Doc ID : Note:47955.1 Content Type: TEXT/PLAIN Subject: Block Corruption FAQ Creation Date: 14-NOV-1997 Type: FAQ Last Revision Date: 17-AUG-2004 Status: PUBLISHED ORACLE SERVER ------------- BLOCK CORRUPTION ---------------- FREQUENTLY ASKED QUESTIONS -------------------------- 25-JAN-2000 CONTENTS -------- 1. What does the error ORA-01578 mean? 2. How to determine what object is corrupted? 3. What are the recovery options if the object is a table? 4. What are the recovery options if the object is an index? 5. What are the recovery options if the object is a rollback segment? 6. What are the recovery options if the object is a data dictionary object? 7. What methods are available to assist in pro-actively identifying corruption? 8. How can corruption be prevented? 9. What are the common causes of corruption? QUESTIONS & ANSWERS 1. What does the error ORA-01578 mean? An Oracle data block is written in an internal binary format which conforms to a defined structure. The size of the physical data block is determined by the "init.ora" parameter DB_BLOCK_SIZE set at the time of database creation. The format of the block is similar regardless of the type of data contained in the block. Each formatted block on disk has a wrapper which consists of a block header and footer. Unformatted blocks should be zero throughout. Whenever a block is read into the buffer cache, the block wrapper information is checked for validity. The checks include verifying that the block passed to Oracle by the operating system is the block requested (data block address) and also that certain information stored in the block header matches information stored in the block footer in case of a split (fractured) block. On a read from disk, if an inconsistency in this information is found, the block is considered to be corrupt and ORA-01578: ORACLE data block corrupted (file # %s, block # %s) is signaled where file# is the file ID of the Oracle datafile and block# is the block number, in Oracle blocks, within that file. However, this does not always mean that the block on disk is truely physically corrupt. That fact needs to be confirmed. 2. How to determine what object is corrupted? The following query will display the segment name, type, and owner: SELECT SEGMENT_NAME, SEGMENT_TYPE, OWNER FROM SYS.DBA_EXTENTS WHERE FILE_ID = AND BETWEEN BLOCK_ID AND BLOCK_ID + BLOCKS - 1; Where is the file number and is the block number reported in the ORA-01578 error message. Suppose block 82817 from table 'USERS' is corrupt: SQL> select extent_id, block_id, blocks from dba_extents where segment_name='USERS'; EXTENT_ID BLOCK_ID BLOCKS ---------- ---------- ---------- 0 82817 8 1 82825 8 2 82833 8 3 82841 8 4 82849 8 SQL> SELECT SEGMENT_NAME, SEGMENT_TYPE, OWNER 2 FROM SYS.DBA_EXTENTS 3 WHERE FILE_ID = 9 4 AND 82817 BETWEEN BLOCK_ID AND BLOCK_ID + BLOCKS - 1; SEGMENT_NAME SEGMENT_TYPE OWNER --------------------------------------------------------------------------------- ------------------ USERS TABLE VPOUSERDB 3. What are the recovery options if the object is a table? The following options exist for resolving non-index block corruption in a table which is not part of the data dictionary: o Restore and recover the database from backup (recommended). o Recover the object from an export. o Select the data out of the table bypassing the corrupted block(s). If the table is a Data Dictionary table, you should contact Oracle Support Services. The recommended recovery option is to restore the database from backup. [NOTE:28814.1] contains information on how to handle ORA-1578 errors in Oracle7. References: ----------- [NOTE:28814.1] TECH ORA-1578 and Data Block Corruption in Oracle7 4. What are the recovery options if the object is an index? If the object is an index which is not part of the data dictionary and the base table does not contain any corrupt blocks, you can simply drop and recreate the index. If the index is a Data Dictionary index, you should contact Oracle Support Services. The recommended recovery option is to restore the database from backup. There is a possibility you might be able to drop the index and then recreate it based on the original create SQL found in the administrative scripts. Oracle Support Services will be able to make the determination as to whether this is a viable option for you. 5. What are the recovery options if the object is a rollback segment? If the object is a rollback segment, you should contact Oracle Support Services. The recommended recovery option is to restore the database from backup. 6. What are the recovery options if the object is a data dictionary object? If the object is a Data Dictionary object, you should contact Oracle Support Services. The recommended recovery option is to restore the database from backup. If the object is an index on a Data Dictionary table, you might be able to drop the index and then recreate it based on the original create SQL found in the administrative scripts. Oracle Support Services will be able to make the determination as to whether this is a viable option. 7. What methods are available to assist in pro-actively identifying corruption? ANALYZE TABLE/INDEX/CLUSTER ... VALIDATE STRUCTURE is a SQL command which can be executed against a table, index, or cluster which scans every block and reports a failure upon encountering any potentially corrupt blocks. The CASCADE option checks all associated indices and verifies the 1 to 1 correspondence between data and index rows. This is the most detailed block check available, but requires the database to be open. DB Verify is a utility which can be run against a datafile of a database that will scan every block in the datafile and generate a report identifying any potentially corrupt blocks. DB Verify performs basic block checking steps, however it does not provide the capability to verify the 1 to 1 correspondence between data and index rows. It can be run when the database is closed. Export will read the blocks allocated to each table being exported and report any potential block corruptions encountered. References: ----------- [NOTE:35512.1] DBVERIFY - Database file Verification Utility (7.3.2 onwards) 8. How can corruption be prevented? Unfortunately, there is no way to totally eliminate the risk of corruption. You can only minimize the risk and plan accordingly. 9. What are the common causes of corruption? o Bad I/O, H/W, Firmware. o Operating System I/O or caching problems. o Memory or paging problems. o Disk repair utilities. o Part of a datafile being overwritten. o Oracle incorrectly attempting to access an unformatted block. o Oracle or operating system bug. Note 77587.1 discusses block corruptions in Oracle and how they are related to the underlying operating system and hardware. References: ----------- [NOTE:77587.1] BLOCK CORRUPTIONS ON ORACLE AND UNIX Note 2: ======= ORA-00600: Internal message code, arguments: [01578] [...] [...] [] [] []. ORA-01578: Oracle data block corrupted (file ..., block ...). Having encountered the Oracle data block corruption, we must firstly investigate which database segment (name and type) the corrupted block is allocated to. Chances are that the block belongs either to an index or to a table segment, since these two type of segments fill the major part of our databases. The following query will reveil the segment that holds the corrupted block identified by and (which were given to you in the error message): SELECT ds.* FROM dba_segments ds, sys.uet$ e WHERE ds.header_file=e.segfile# and ds.header_block=e.segblock# and e.file#= and between e.block# and e.block#+e.length-1; If the segment turns out to be an index segment, then the problem can be very quickly solved. Since all the table data required for recreating the index is still accessable, we can drop and recreate the index (since the block will reformatted, when taken FROM the free-space list and reused for the index). If the segment turns out to be a table segment a number of options for solving the problem are available: - restore and recovery of datafile the block is in - imp table - sql The last option involves using SQL to SELECT as much data as possible FROM the current corrupted table segment and save the SELECTed rows into a new table. SELECTing data that is stored in segment blocks that preceede the corrupted block can be easily done using a full table scan (via a cursor). Rows stored in blocks after the corrupted block cause a problem. A full table scan will never reach these. However these rows can still be fetched using rowids (single row lookups). 2.1 Table was indexed Using an optimizer hint we can write a query that SELECTs the rows FROM the table via an index scan (using rowid's), instead of via a full table scan. Let's assume our table is named X with columns a, b and c. And table X is indexed uniquely on columns a and b by index X_I, the query would look like: SELECT /*+index(X X_I) */ a, b, c FROM X; We must now exclude the corrupt block FROM being accessed to avoid the internal exception ORA-00600[01578]. Since the blocknumber is a substring of the rowid ( ) this can very easily be achieved: SELECT /*+index(X X_I) */ a, b, c FROM X WHERE rowid not like ||'.%.'||; But it is important to realize that the WHERE-clause gets evaluated right after the index is accessed and before the table is accessed. Otherwise we would still get the ORA-00600[01578] exception. Using the above query as a subquery in an insert statement we can restore all rows of still valid blocks to a new table. Since the index holds the actual column values of the indexed columns we could also use the index to restore all indexed columns of rows that reside in the corrupt block. The following query, SELECT /*+index(X X_I) */ a, b FROM X WHERE rowid like ||'.%.'||; retreives only indexed columns a and b FROM rows inside the corrupt block. The optimizer will not access the table for this query. It can retreive the column values using the index segment only. Using this technique we are able to restore all indexed column values of the rows inside the corrupt block, without accessing the corrupt block at all. Suppose in our example that column c of table X was also indexed by index X_I2. This enables us to completely restore rows inside the corrupt block. First restore columns a and b using index X_I: create table X_a_b(rowkey,a,b) as SELECT /*+index(X X_I) */ rowid, a, b FROM X WHERE rowid like ||'.%.'||; Then restore column c using index X_I2: create table X_c(rowkey,c) as SELECT /*+index(X X_I2) */ rowid, c FROM X WHERE rowid like ||'.%.'||; And finally join the columns together using the restored rowid: SELECT x1.a, x1.b, x2.c FROM X_a_b x1, X_c x2 WHERE x1.rowkey=x2.rowkey; In summary: Indexes on the corrupted table segment can be used to restore all columns of all rows that are stored outside the corrupted data blocks. Of rows inside the corrupted data blocks, only the columns that were indexed can be restored. We might even be able to use an old version of the table (via Import) to further restore non-indexed columns of these records. 2.2 Table has no indexes This situation should rarely occur since every table should have a primary key and therefore a unique index. However when no index is present, all rows of corrupted blocks should be considered lost. All other rows can be retrieved using rowid's. Since there is no index we must build a rowid generator ourselves. The SYS.UET$ table shows us exactly which extents (file#, startblock, endblock) we need to inspect for possible rows of our table X. If we make an estimate of the maximum number of rows per block for table X, we can build a PL/SQL-loop that generates possible rowid's of records inside table X. By handling the 'invalid rowid' exception, and skipping the corrupted data block, we can restore all rows except those inside the corrupted block. declare v_rowid varchar2(18); v_xrec X%rowtype; e_invalid_rowid exception; pragma exception_init(e_invalid_rowid,-1410); begin for v_uetrec in (SELECT file# file, block# start_block, block#+length#-1 end_block FROM uet$ WHERE segfile#=6 and segblock#=64) -- Identifies our segment X. loop for v_blk in v_uetrec.start_block..v_uetrec.end_block loop if v_uetrec.file<>6 and v_blk<>886 -- 886 in file 6 is our corrupted block. then for v_row in 0..200 -- 200 is maximum number of rows per block for segment X. loop begin SELECT a,b,c into v_rec FROM x WHERE rowid=chartorowid('0000'||hex(v_blk)||'.'|| hex(v_row)||'.'||hex(v_uetrec.file); insert into x_saved(a,b,c) values(v_rec.a,v_rec.b,v_rec.c); commit; exception when e_invalid_rowid then null; end; end loop; /*row-loop*/ end if; end loop; /*blk-loop*/ end loop; /*uet-loop*/ end; / The above code assumes that block id's never exceed 4 hexadecimal places. A definition of the hex-function which is used in the above code can be found in the appendix. Note 3: ======= Doc ID : Note:33405.1 Content Type: TEXT/PLAIN Subject: Extracting Data from a Corrupt Table using SKIP_CORRUPT_BLOCKS or Event 10231 Creation Date: 24-JAN-1996 Type: BULLETIN Last Revision Date: 13-SEP-2000 Status: PUBLISHED ***************** *** *** ***************** This note is an extension to article [NOTE:28814.1] about handling block corruption errors where the block wrapper of a datablock indicates that the block is bad. (Typically for ORA-1578 errors). The details here will not work if only the block internals are corrupt (eg: for ORA-600 or other errors). Please read [NOTE:28814.1] before reading this note. Introduction ~~~~~~~~~~~~ This short article explains how to skip corrupt blocks on an object either using the Oracle8i SKIP_CORRUPT table flag or the special Oracle event number 10231 which is available in Oracle releases 7 through 8.1 inclusive. The information here explains how to use these options. Before proceeding you should: a) Be certain that the corrupt block is on a USER table. (i.e.: not a data dictionary table) b) Have contacted Oracle Support Services and been advised to use event 10231 or the SKIP_CORRUPT flag. c) Have decided how you are to recreate the table. Eg: Export , and disk space is available etc.. d) You have scheduled down-time to attempt the salvage operation. e) Have a backup of the database. f) Have the SQL to rebuild the problem table, its indexes constraints, triggers, grants etc... This SQL should include relevant storage clauses. What is event 10231 ? ~~~~~~~~~~~~~~~~~~~~~ This event allows Oracle to skip certain types of corrupted blocks on full table scans ONLY hence allowing export or "create table as select" type operations to retrieve rows from the table which are not in the corrupt block. Data in the corrupt block is lost. The scope of this event is limited for Oracle versions prior to Oracle 7.2 as it only allows you to skip 'soft corrupt' blocks. Most ORA 1578 errors are a result of media corruptions and in such cases event 10231 is useless. From Oracle 7.2 onwards the event allows you to skip many forms of media corrupt blocks in addition to soft corrupt blocks and so is far more useful. It is still *NOT* guaranteed to work. [NOTE:28814.1] describes alternatives which can be used if this event fails. What is the SKIP_CORRUPT flag ? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In Oracle8i the functionality of the 10231 event has been externalised on a PER-SEGMENT basis such that it is possible to mark a TABLE or PARTITION to skip over corrupt blocks when possible. The flag is set or cleared using the DBMS_REPAIR package. DBA_TABLES has a SKIP_CORRUPT column which indicates if this flag is set for an object or not. Setting the event or flag ~~~~~~~~~~~~~~~~~~~~~~~~~ The event can either be set within the session or at database instance level. If you intend to use a CREATE TABLE AS SELECT then setting the event in the session may suffice. If you want to EXPORT the table data then it is best to set the event at instance level, or set the SKIP_CORRUPT table attribute if on Oracle8i. Oracle8i ~~~~~~~~ Connect as a DBA user and mark the table as needing to skip corrupt blocks thus: execute DBMS_REPAIR.SKIP_CORRUPT_BLOCKS('',''); or for a table partition: execute DBMS_REPAIR.SKIP_CORRUPT_BLOCKS('',''.''); Now you should be able to issue a CREATE TABLE AS SELECT operation against the corrupt table to extract data from all non-corrupt blocks, or EXPORT the table. Eg: CREATE TABLE salvage_emp AS SELECT * FROM corrupt_emp; To clear the attribute for a table use: execute DBMS_REPAIR.SKIP_CORRUPT_BLOCKS('','', flags=>dbms_repair.noskip_flag); execute DBMS_REPAIR.SKIP_CORRUPT_BLOCKS('VPOUSERDB','USERS', flags=>dbms_repair.noskip_flag); Setting the event in a Session ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Connect to Oracle as a user with access to the corrupt table and issue the command: ALTER SESSION SET EVENTS '10231 TRACE NAME CONTEXT FOREVER, LEVEL 10'; Now you should be able to issue a CREATE TABLE AS SELECT operation against the corrupt table to extract data from all non-corrupt blocks, but an export would still fail as the event is only set within your current session. Eg: CREATE TABLE salvage_emp AS SELECT * FROM corrupt_emp; Setting the event at Instance level ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ This requires that the event be added to the init$ORACLE_SID.ora file used to start the instance: shutdown the database Edit your init.ora startup configuration file and ADD a line that reads: event="10231 trace name context forever, level 10" Make sure this appears next to any other EVENT= lines in the init.ora file. STARTUP If the instance fails to start check the syntax of the event parameter matches the above exactly. Note the comma as it is important. SHOW PARAMETER EVENT To check the event has been set in the correct place. You should see the initial portion of text for the line in your init.ora file. If not check which parameter file is being used to start the database. Select out the data from the table using a full table scan operation. Eg: Use a table level export or create table as select. Export Warning: If the table is very large then some versions of export may not be able to write more than 2Gb of data to the export file. See [NOTE:62427.1] for general information on 2Gb limits in various Oracle releases. Salvaging data from the corrupt block itself ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ SKIP_CORRUPT and event 10231 extract data from good blocks but skip over corrupt blocks. To extract information from the corrupt block there are three main options: - Select column data from any good indexes This is discussed towards the end of the following 2 articles: Oracle7 - using ROWID range scans [NOTE:34371.1] Oracle8/8i - using ROWID range scans [NOTE:61685.1] - See if Oracle Support can extract any data from HEX dumps of the corrupt block. - It may be possible to salvage some data using Log Miner Once you have the data extracted ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Once you have the required data extracted either into an export file or into another table make sure you have a valid database backup before proceeding. The importance of this cannot be over-emphasised. Double check you have the SQL to rebuild the object and its indexes etc.. Double check that you have any diagnostic information if requested by Oracle support. Once you proceed with dropping the object certain information is destroyed so it is important to capture it now. Now you can: If 10231 was set at instance level: Remove the 'event' line from the init.ora file SHUTDOWN and RESTART the database. SHOW PARAMETER EVENT Make sure the 10231 event is no longer shown RENAME or DROP the problem table If you have space it is advisable to RENAME the problem table rather than DROP it at this stage. Recreate the table. Eg: By importing. Take special care to get the storage clauses correct when recreating the table. Create any indexes, triggers etc.. required Again take care with storage clauses. Re-grant any access to the table. If you RENAMEd the original table you can drop it once the new table has been tested. . Note 4: Analyze table validate structure: ========================================= validate structure table: ANALYZE TABLE CHARLIE.CUSTOMERS VALIDATE STRUCTURE; validate structure index: ANALYZE INDEX CHARLIE.PK_CUST VALIDATE STRUCTURE; Als er geen corrupte blocks worden gevonden, is de output slechts "table analyzed". Als er wel corrupte blocks worden gevonden, moet een aangemaakte trace file worden bekeken. Note 5: DBVERIFY Utility: ========================= Vanaf de OS prompt kan het dbv utility gedraaid worden om een datafile te onderzoeken. $ dbv FILE=/u02/oracle/cc1/data01.dbf BLOCKSIZE=8192 Note 6: DBMS_REPAIR package: ============================ Het DBMS_REPAIR package wordt aangemaakt door bmprpr.sql script. Stap 1. via ANALYZE TABLE ben je er achter gekomen dat van een table een of meer blocks corrupt zijn. Stap 2. Gebruik eerst DBMS_REPAIR.ADMIN_TABLES om de REPAIR_TABLE aan te maken. Deze table zal dan gegevens gaan bevatten over de blocks, en of die gemarkeerd zijn als zijnde corrupt e.d. declare begin dbms_repair.admin_tables('REPAIR_TABLE, dbms_repair.repair_table, dbms_repair.create_action, 'users'); end; / Stap 3. Gebruik nu de DBMS_REPAIR.CHECK_OBJECT procedure op het object om de repair_table uit stap 2 te vullen met corruptie gegevens. set serveroutput on declare rpr_count int; begin rpr_count:=0; dbms_repair.check_object('CHARLIE', 'CUSTOMERS', 'REPAIR_TABLE', rpr_count); dbms_output.put_line('repair_block_count :'||to_char(rpr_count)); end; / Note 7: ======= Tom, If I have this information: select * from V$DATABASE_BLOCK_CORRUPTION; FILE# BLOCK# BLOCKS CORRUPTION_CHANGE# CORRUPTIO ---------- ---------- ---------- ------------------ --------- 11 12357 12 197184960 LOGICAL and select * from v$backup_corruption; RECID STAMP SET_STAMP SET_COUNT PIECE# FILE# BLOCK# BLOCKS CORRUPTION_CHANGE# MAR CO ---------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ------------ 1 533835361 533835140 3089 1 11 12357 12 197184960 NO LOGICAL How can I get more details of what data resides on this blocks? and being 'Logical' can they be recoverd without loosing that data at all? Any extra details would be appreciated. Thanks, Orlando Followup: select * from dba_extents where file_id = 11 and 12357 between block_id an block_id+blocks-1; if it is something "rebuildable" -- like an index, drop and recreate might be the path of least resistance, else you would go back to your backups -- to before this was detected and restore that file/range of blocks (rman can do block level recovery) Tom trace file generated by analyze contained table scan: segment: file# 55 block# 229385 skipping corrupt block file# 55 block# 251372 This is repeated every day (analyzed each morning) but daily direct export / import succeeds. SQL> select segment_type from dba_extents where file_id=55 and 229385 between block_id and (block_id +( blocks -1)); SEGMENT_TYPE ---------------------------------------- TABLE $ dbv file=/u03/oradata/emu/emu_data_large02.dbf \ blocksize=8192 logfile=/dbv.log DBVERIFY: Release 8.1.7.2.0 - Production on Mon Aug 10 10:10:13 2004 (c) Copyright 2000 Oracle Corporation. All rights reserved. DBVERIFY - Verification starting : FILE = /u03/oradata/emu/emu_data_large02.dbf Block Checking: DBA = 230938092, Block Type = KTB-managed data block Found block already marked corrupted DBVERIFY - Verification complete Total Pages Examined : 256000 Total Pages Processed (Data) : 253949 Total Pages Failing (Data) : 0 Total Pages Processed (Index): 0 Total Pages Failing (Index): 0 Total Pages Processed (Other): 11 Total Pages Empty : 2040 Total Pages Marked Corrupt : 0 Total Pages Influx : 0 Any thoughts ? Thanks Note 6: ------- Detect And Correct Corruption Oracle provides a number of methods to detect and repair corruption within datafiles: DBVerify ANALYZE .. VALIDATE STRUCTURE DB_BLOCK_CHECKING. DBMS_REPAIR. Other Repair Methods. DBVerify DBVerify is an external utility that allows validation of offline datafiles. In addition to offline datafiles it can be used to check the validity of backup datafiles: C:>dbv file=C:\Oracle\oradata\TSH1\system01.dbf feedback=100 blocksize=4096 ANALYZE .. VALIDATE STRUCTURE The ANALYZE command can be used to verify each data block in the analyzed object. If any corruption is detected rows are added to the INVALID_ROWS table: -- Create the INVALID_ROWS table. SQL> @C:\Oracle\901\rdbms\admin\UTLVALID.SQL -- Validate the table structure. SQL> ANALYZE TABLE scott.emp VALIDATE STRUCTURE; -- Validate the table structure along with all it's indexes. SQL> ANALYZE TABLE scott.emp VALIDATE STRUCTURE CASCADE; -- Validate the index structure. SQL> ANALYZE INDEX scott.pk_emp VALIDATE STRUCTURE; DB_BLOCK_CHECKING When the DB_BLOCK_CHECKING parameter is set to TRUE Oracle performs a walk through of the data in the block to check it is self-consistent. Unfortunately block checking can add between 1 and 10% overhead to the server. Oracle recommend setting this parameter to TRUE if the overhead is acceptable. DBMS_REPAIR Unlike the previous methods dicussed, the DBMS_REPAIR package allows you to detect and repair corruption. The process requires two administration tables to hold a list of corrupt blocks and index keys pointing to those blocks. These are created as follows: BEGIN Dbms_Repair.Admin_Tables ( table_name => 'REPAIR_TABLE', table_type => Dbms_Repair.Repair_Table, action => Dbms_Repair.Create_Action, tablespace => 'USERS'); Dbms_Repair.Admin_Tables ( table_name => 'ORPHAN_KEY_TABLE', table_type => Dbms_Repair.Orphan_Table, action => Dbms_Repair.Create_Action, tablespace => 'USERS'); END; / With the administration tables built we are able to check the table of interest using the CHECK_OBJECT procedure: SET SERVEROUTPUT ON DECLARE v_num_corrupt INT; BEGIN v_num_corrupt := 0; Dbms_Repair.Check_Object ( schema_name => 'SCOTT', object_name => 'DEPT', repair_table_name => 'REPAIR_TABLE', corrupt_count => v_num_corrupt); Dbms_Output.Put_Line('number corrupt: ' || TO_CHAR (v_num_corrupt)); END; / Assuming the number of corrupt blocks is greater than 0 the CORRUPTION_DESCRIPTION and the REPAIR_DESCRIPTION columns of the REPAIR_TABLE can be used to get more information about the corruption. At this point the currupt blocks have been detected, but are not marked as corrupt. The FIX_CORRUPT_BLOCKS procedure can be used to mark the blocks as corrupt, allowing them to be skipped by DML once the table is in the correct mode: SET SERVEROUTPUT ON DECLARE v_num_fix INT; BEGIN v_num_fix := 0; Dbms_Repair.Fix_Corrupt_Blocks ( schema_name => 'SCOTT', object_name=> 'DEPT', object_type => Dbms_Repair.Table_Object, repair_table_name => 'REPAIR_TABLE', fix_count=> v_num_fix); Dbms_Output.Put_Line('num fix: ' || to_char(v_num_fix)); END; / Once the corrupt table blocks have been located and marked all indexes must be checked to see if any of their key entries point to a corrupt block. This is done using the DUMP_ORPHAN_KEYS procedure: SET SERVEROUTPUT ON DECLARE v_num_orphans INT; BEGIN v_num_orphans := 0; Dbms_Repair.Dump_Orphan_Keys ( schema_name => 'SCOTT', object_name => 'PK_DEPT', object_type => Dbms_Repair.Index_Object, repair_table_name => 'REPAIR_TABLE', orphan_table_name=> 'ORPHAN_KEY_TABLE', key_count => v_num_orphans); Dbms_Output.Put_Line('orphan key count: ' || to_char(v_num_orphans)); END; / If the orphan key count is greater than 0 the index should be rebuilt. The process of marking the table block as corrupt automatically removes it from the freelists. This can prevent freelist access to all blocks following the corrupt block. To correct this the freelists must be rebuilt using the REBUILD_FREELISTS procedure: BEGIN Dbms_Repair.Rebuild_Freelists ( schema_name => 'SCOTT', object_name => 'DEPT', object_type => Dbms_Repair.Table_Object); END; / The final step in the process is to make sure all DML statements ignore the data blocks marked as corrupt. This is done using the SKIP_CORRUPT_BLOCKS procedure: BEGIN Dbms_Repair.Skip_Corrupt_Blocks ( schema_name => 'SCOTT', object_name => 'DEPT', object_type => Dbms_Repair.Table_Object, flags => Dbms_Repair.Skip_Flag); END; / The SKIP_CORRUPT column in the DBA_TABLES view indicates if this action has been successful. At this point the table can be used again but you will have to take steps to correct any data loss associated with the missing blocks. Other Repair Methods Other methods to repair corruption include: Full database recovery. Individual datafile recovery. Block media recovery (BMR), available in Oracle9i when using RMAN. Recreate the table using the CREATE TABLE .. AS SELECT command, taking care to avoid the corrupt blocks by retricting the where clause of the query. Drop the table and restore it from a previous export. This may require some manual effort to replace missing data. Hope this helps. Regards Tim... Note 7: ------- If you know the file number and the block number indicating the corruption, you can salvage the data in the corrupt table by selecting around the bad blocks. Set event 10231 in the init.ora file to cause Oracle to skip software- and media- corrupted blocks when performing full table scans: Event="10231 trace name context forever, level 10" Set event 10233 in the init.ora file to cause Oracle to skip software- and media- corrupted blocks when performing index range scans: Event="10233 trace name context forever, level 10" Note 8: ------- Detecting and reporting data block corruption using the DBMS_REPAIR package: Note: Note that this event can only be used if the block "wrapper" is marked corrupt. Eg: If the block reports ORA-1578. 1. Create DBMS_REPAIR administration tables: To Create Repair tables, run the below package. SQL> EXEC DBMS_REPAIR.ADMIN_TABLES(‘REPAIR_ADMIN’, 1,1, ‘REPAIR_TS’); Note that table names prefix with ‘REPAIR_’ or ‘ORPAN_’. If the second variable is 1, it will create ‘REAIR_key tables, if it is 2, then it will create ‘ORPAN_key tables. If the thread variable is 1 then package performs ‘create’ operations. 2 then package performs ‘delete’ operations. 3 then package performs ‘drop’ operations. 2. Scanning a specific table or Index using the DBMS_REPAIR.CHECK_OBJECT procedure: In the following example we check the table employee for possible corruption’s that belongs to the schema TEST. Let’s assume that we have created our administration tables called REPAIR_ADMIN in schema SYS. To check the table block corruption use the following procedure: SQL> VARIABLE A NUMBER; SQL> EXEC DBMS_REPAIR.CHECK_OBJECT (‘TEST’,’EMP’, NULL, 1,’REPAIR_ADMIN’, NULL, NULL, NULL, NULL,:A); SQL> PRINT A; To check which block is corrupted, check in the REPAIR_ADMIN table. SQL> SELECT * FROM REPAIR_ADMIN; 3. Fixing corrupt block using the DBMS_REPAIR.FIX_CORRUPT_BLOCK procedure: SQL> VARIABLE A NUMBER; SQL> EXEC DBMS_REPAIR.FIX.CORRUPT_BLOCKS (‘TEST’,’EMP’, NULL, 1,’REPARI_ADMIN’, NULL,:A); SQL> SELECT MARKED FROM REPAIR_ADMIN; If u select the EMP table now you still get the error ORA-1578. 4. Skipping corrupt blocks using the DBMS_REPAIR. SKIP_CORRUPT_BLOCK procedure: SQL> EXEC DBMS_REPAIR. SKIP_CORRUPT.BLOCKS (‘TEST’, ‘EMP’, 1,1); Notice the verification of running the DBMS_REPAIR tool. You have lost some of data. One main advantage of this tool is that you can retrieve the data past the corrupted block. However we have lost some data in the table. 5. This procedure is useful in identifying orphan keys in indexes that are pointing to corrupt rows of the table: SQL> EXEC DBMS_REPAIR. DUMP ORPHAN_KEYS (‘TEST’,’IDX_EMP’, NULL, 2, ‘REPAIR_ADMIN’, ‘ORPHAN_ADMIN’, NULL,:A); If u see any records in ORPHAN_ADMIN table you have to drop and re-create the index to avoid any inconsistencies in your queries. 6. The last thing you need to do while using the DBMS_REPAIR package is to run the DBMS_REPAIR.REBUILD_FREELISTS procedure to reinitialize the free list details in the data dictionary views. SQL> EXEC DBMS_REPAIR.REBUILD_FREELISTS (‘TEST’,’EMP’, NULL, 1); NOTE Setting events 10210, 10211, 10212, and 10225 can be done by adding the following line for each event in the init.ora file: Event = "event_number trace name errorstack forever, level 10" - When event 10210 is set, the data blocks are checked for corruption by checking their integrity. Data blocks that don't match the format are marked as soft corrupt. - When event 10211 is set, the index blocks are checked for corruption by checking their integrity. Index blocks that don't match the format are marked as soft corrupt. - When event 10212 is set, the cluster blocks are checked for corruption by checking their integrity. Cluster blocks that don't match the format are marked as soft corrupt. - When event 10225 is set, the fet$ and uset$ dictionary tables are checked for corruption by checking their integrity. Blocks that don't match the format are marked as soft corrupt. - Set event 10231 in the init.ora file to cause Oracle to skip software- and media-corrupted blocks when performing full table scans: Event="10231 trace name context forever, level 10" - Set event 10233 in the init.ora file to cause Oracle to skip software- and media-corrupted blocks when performing index range scans: Event="10233 trace name context forever, level 10" To dump the Oracle block you can use below command from 8.x on words: SQL> ALTER SYSTEM DUMP DATAFILE 11 block 9; This command dumps datablock 9 in datafile11, into USER_DUMP_DEST directory. Dumping Redo Logs file blocks: SQL> ALTER SYSTEM DUMP LOGFILE ‘/usr/oracle8/product/admin/udump/rl. log’; Rollback segments block corruption, it will cause problems (ORA-1578) while starting up the database. With support of oracle, can use below under source parameter to startup the database. _CORRUPTED_ROLLBACK_SEGMENTS=(RBS_1, RBS_2) DB_BLOCK_COMPUTE_CHECKSUM This parameter is normally used to debug corruption’s that happen on disk. The following V$ views contain information about blocks marked logically corrupt: V$ BACKUP_CORRUPTION, V$COPY_CORRUPTION When this parameter is set, while reading a block from disk to catch, oracle will compute the checksum again and compares it with the value that is in the block. If they differ, it indicates that the block is corrupted on disk. Oracle makes the block as corrupt and signals an error. There is an overhead involved in setting this parameter. DB_BLOCK_CACHE_PROTECT=‘TRUE’ Oracle will catch stray writes made by processes in the buffer catch. Oracle 9i new RMAN futures: Obtain the datafile numbers and block numbers for the corrupted blocks. Typically, you obtain this output from the standard output, the alert.log, trace files, or a media management interface. For example, you may see the following in a trace file: ORA-01578: ORACLE data block corrupted (file # 9, block # 13) ORA-01110: data file 9: '/oracle/dbs/tbs_91.f' ORA-01578: ORACLE data block corrupted (file # 2, block # 19) ORA-01110: data file 2: '/oracle/dbs/tbs_21.f' $rman target =rman/rman@rmanprod RMAN> run { 2> allocate channel ch1 type disk; 3> blockrecover datafile 9 block 13 datafile 2 block 19; 4> } Recovering Data blocks Using Selected Backups: # restore from backupset BLOCKRECOVER DATAFILE 9 BLOCK 13 DATAFILE 2 BLOCK 19 FROM BACKUPSET; # restore from datafile image copy BLOCKRECOVER DATAFILE 9 BLOCK 13 DATAFILE 2 BLOCK 19 FROM DATAFILECOPY; # restore from backupset with tag "mondayAM" BLOCKRECOVER DATAFILE 9 BLOCK 13 DATAFILE 2 BLOCK 199 FROM TAG = mondayAM; # restore using backups made before one week ago BLOCKRECOVER DATAFILE 9 BLOCK 13 DATAFILE 2 BLOCK 19 RESTORE UNTIL 'SYSDATE-7'; # restore using backups made before SCN 100 BLOCKRECOVER DATAFILE 9 BLOCK 13 DATAFILE 2 BLOCK 19 RESTORE UNTIL SCN 100; # restore using backups made before log sequence 7024 BLOCKRECOVER DATAFILE 9 BLOCK 13 DATAFILE 2 BLOCK 19 RESTORE UNTIL SEQUENCE 7024; Note 9: ======= Displayed below are the messages of the selected thread. Thread Status: Closed From: nitinpawar@birlasunlife.com 23-Feb-05 11:51 Subject: ORA-01578 on system datafile RDBMS Version: Oracle9i Enterprise Edition Release 9.2.0.1.0 Operating System and Version: Windows 2000 Error Number (if applicable): ORA-01578 Product (i.e. SQL*Loader, Import, etc.): Product Version: ORA-01578 on system datafile A data block in SYSTEM tablespace datafile is corrupted. The error has been occuring since past 7 months. I noticed it recently when I took over the support. The database is in archivelog mode. We don't have any old hot backups of the database files. Both export and alert log indicate corrupt block to be # 7873, but dbverify declares block #7875 to be corrupt. It seems there is no object using the block. Following is the extract from the alert log. *** Corrupt block relative dba: 0x00401ec1 (file 1, block 7873) Fractured block found during buffer read Data in bad block - type: 16 format: 2 rdba: 0x00401ec1 last change scn: 0x0000.00007389 seq: 0x1 flg: 0x04 consistency value in tail: 0x23430601 check value in block header: 0x5684, computed block checksum: 0x396b spare1: 0x0, spare2: 0x0, spare3: 0x0 *** Reread of rdba: 0x00401ec1 (file 1, block 7873) found same corrupted data From: Oracle, Fahad Abdul Rahman 25-Feb-05 08:18 Subject: Re : ORA-01578 on system datafile Nitin, I would suggest you to relocate the system datafiles to a new location on disk and see if the corruption is removed. If the issue still persist ,then I would suggest you to log a TAR with Oracle Support for further research. ======================== 32. iSQL*Plus and EM 10: ======================== 32.1 iSQL*Plus: =============== Note 1: ------- How to start iSql*Plus: ----------------------- lsnrctl start emctl start dbconsole isqlplusctl start http://localhost:5561/isqlplus/ Note 2: ------- Doc ID: Note:281946.1 Content Type: TEXT/X-HTML Subject: How to Verify that iSQL*Plus 10i is Running and How to Restart the Processes? Creation Date: 31-AUG-2004 Type: HOWTO Last Revision Date: 06-APR-2005 Status: PUBLISHED The information in this document applies to: SQL*Plus - Version: 10.1.0 Information in this document applies to any platform. Goal How to verify that iSQL*Plus 10i is running, and how to restart the processes? Fix How to Verify that iSQL*Plus is running? ======================================= UNIX Platform ------------------- Check whether the iSQL*Plus process is running by entering the following command: ps -eaf |grep java The iSQL*Plus process looks something like the following: oracle 18488 1 0 16:01:30 pts/8 0:36 $ORACLE_HOME/jdk/bin/java -Djava.awt.headless=true -Doracle.oc4j.localhome=/ora Windows Platform -------------------------- Check whether the iSQL*Plus process is running by opening the Windows services dialog from the Control Panel and checking the status of the iSQL*Plus service. The iSQL*Plus service will be called "OracleOracle_Home_NameiSQL*Plus". How to Start and Stop iSQL*Plus? =============================== UNIX Platform -------------------- To start iSQL*Plus, enter the command: $ORACLE_HOME/bin/isqlplusctl start To stop iSQL*Plus, enter the command: $ORACLE_HOME/bin/isqlplusctl stop Windows Platform -------------------------- Use the Windows service to start and stop iSQL*Plus. The service is set to start automatically on installation and when the operating system is started. Note 3: ------- Doc ID: Note:281847.1 Content Type: TEXT/X-HTML Subject: How do I configure or test iSQL*Plus 10i? Creation Date: 30-AUG-2004 Type: HOWTO Last Revision Date: 25-MAR-2005 Status: PUBLISHED The information in this document applies to: SQL*Plus - Version: 10.1.0.0 to 10.1.0 Information in this document applies to any platform. Goal How do I configure or test?iSQL*Plus after the install or Oracle Enterprise Edition 10i? Fix iSQL*Plus 10.x is automatically installed and configured with Enterprise Edition 10i. At the end of the installation process a file called $ORACLE_HOME/install/readme.txt has the information needed to configure or test iSQL*Plus: readme.txt example: ---------------- The following J2EE Applications have been deployed and are accessible at the URLs listed below. Your database configuration files have been installed in?$ORACLE_HOME while other components selected for installation have been installed in $ORACLE_HOME\Db_1.? Be cautious not to accidentally delete these configuration files. Ultra Search URL: :5620/ultrasearch"http://:5620/ultrasearch Ultra Search Administration Tool URL: :5620/ultrasearch/admin"http://:5620/ultrasearch/admin iSQL*Plus URL: :5560/isqlplus"http://:5560/isqlplus Enteprise Manager 10g Database Control URL: :5500/em"http://:5500/em ---------------- The URL for your iSQL*Plus server is: :port/isqlplus" target=_blankhttp://:port /isqlplus :port/isqlplus/dba" target=_blankhttp://:port /isqlplus/dba The port number is likely to be 5560. If this URL does not display the iSQL*Plus log in page, check that iSQL*Plus has been started For more additional information about iSQL*Plus please check the following Metalink notes: Note 281947.1 How to Troubleshoot iSQLPlus 10i when it is not Starting on Unix? Note 281946.1?How to Verify that iSQLPlus 10i is Running and How to Restart the Processes? Note 283114.1?How to connect as sysdba/sysoper through iSQL*Plus in Oracle 10g Note 4: ------- Doc ID: Note:283114.1 Content Type: TEXT/X-HTML Subject: How to connect as sysdba/sysoper through iSQL*Plus in Oracle 10g Creation Date: 16-SEP-2004 Type: HOWTO Last Revision Date: 12-JAN-2005 Status: MODERATED This document is being delivered to you via Oracle Support's Rapid Visibility (RaV) process, and therefore has not been subject to an independent technical review. The information in this document applies to: SQL*Plus - Version: 10.0.1 Information in this document applies to any platform. Goal Enabling iSQL*Plus DBA Access. Fix Inorder to connect as SYSDBA through iSQL*Plus you will have to use iSQL*Plus DBA URL. Given below is a sample DBS URL in iSQL*Plus. " target=_blankhttp://Hostname:Port/isqlplus/dba Enabling iSQL*Plus DBA Access ============================= To access the iSQL*Plus DBA URL, you must set up the OC4J user manager. You can set up OC4J to use: The XML-based provider type, jazn-data.xml The LDAP-based provider type, Oracle Internet Directory This document discusses how to set up the iSQL*Plus DBA URL to use the XML-based provider. For information on how to set up the LDAP-based provider, see the Oracle9iAS Containers for J2EE documentation. To set up the iSQL*Plus DBA URL ================================= 1. Create users for the iSQL*Plus DBA URL. 2. Grant the webDba role to users. 3. Test iSQL*Plus DBA Access The Oracle JAAS Provider, otherwise known as JAZN (Java AuthoriZatioN), is Oracle's implementation of the Java Authentication and Authorization Service (JAAS). Oracle's JAAS Provider is referred to as JAZN in the remainder of this document. See the Oracle9iAS Containers for J2EE documentation for more information about JAZN, the Oracle JAAS Provider. Create and Manage Users for the iSQL*Plus DBA URL ================================================= The actions available to manage users for the iSQL*Plus DBA URL are: 1. Create users 2. List users 3. Grant the webDba role 4. Remove users 5. Revoke the webDba role 6. Change user passwords You perform these actions from the $ORACLE_HOME/oc4j/j2ee/isqlplus/application-deployments/isqlplus directory. $JAVA_HOME is the location of your JDK (1.4 or above). It should be set to $ORACLE_HOME/jdk, but you may use another JDK. admin_password is the password for the iSQL*Plus DBA realm administrator user, admin. The password for the admin user is set to 'welcome' by default. You should change this password as soon as possible. A JAZN shell option, and a command line option are given for all steps. To start the JAZN shell, enter: $JAVA_HOME/bin/java -Djava.security.properties=$ORACLE_HOME/sqlplus/admin/iplus/provider -jar $ORACLE_HOME/oc4j/j2ee/home/jazn.jar -user "iSQL*Plus DBA/admin" -password admin_password -shell To exit the JAZN shell, enter: EXIT Create Users You can create multiple users who have access to the iSQL*Plus DBA URL. To create a user from the JAZN shell, enter: JAZN> adduser "iSQL*Plus DBA" username password To create a user from the command-line, enter: $JAVA_HOME/bin/java -Djava.security.properties=$ORACLE_HOME/sqlplus/admin/iplus/provider -jar $ORACLE_HOME/oc4j/j2ee/home/jazn.jar -user "iSQL*Plus DBA/admin" -password admin_password -adduser "iSQL*Plus DBA" username password username and password are the username and password used to log into the iSQL*Plus DBA URL. To create multiple users, repeat the above command for each user. List Users You can confirm that users have been created and added to the iSQL*Plus DBA realm. To confirm the creation of a user using the JAZN shell, enter: JAZN> listusers "iSQL*Plus DBA" To confirm the creation of a user using the command-line, enter: $JAVA_HOME/bin/java -Djava.security.properties=$ORACLE_HOME/sqlplus/admin/iplus/provider -jar $ORACLE_HOME/oc4j/j2ee/home/jazn.jar -user "iSQL*Plus DBA/admin" -password admin_password -listusers "iSQL*Plus DBA" The usernames you created are displayed. Grant Users the webDba Role Each user you created above must be granted access to the webDba role. To grant a user access to the webDba role from the JAZN shell, enter: JAZN> grantrole webDba "iSQL*Plus DBA" username To grant a user access to the webDba role from the command-line, enter: $JAVA_HOME/bin/java -Djava.security.properties=$ORACLE_HOME/sqlplus/admin/iplus/provider -jar $ORACLE_HOME/oc4j/j2ee/home/jazn.jar -user "iSQL*Plus DBA/admin" -password admin_password -grantrole webDba "iSQL*Plus DBA" username Remove Users To remove a user using the JAZN shell, enter: JAZN> remuser "iSQL*Plus DBA" username To remove a user using the command-line, enter: $JAVA_HOME/bin/java -Djava.security.properties=$ORACLE_HOME/sqlplus/admin/iplus/provider -jar $ORACLE_HOME/oc4j/j2ee/home/jazn.jar -user "iSQL*Plus DBA/admin" -password admin_password -remuser "iSQL*Plus DBA" username Revoke the webDba Role To revoke a user's webDba role from the JAZN shell, enter: JAZN> revokerole webDba "iSQL*Plus DBA" username To revoke a user's webDba role from the command-line, enter: $JAVA_HOME/bin/java -Djava.security.properties=$ORACLE_HOME/sqlplus/admin/iplus/provider -jar $ORACLE_HOME/oc4j/j2ee/home/jazn.jar -user "iSQL*Plus DBA/admin" -password admin_password -revokerole "iSQL*Plus DBA" username Change User Passwords To change a user's password from the JAZN shell, enter: JAZN> setpasswd "iSQL*Plus DBA" username old_password new_password To change a user's password from the command-line, enter: $JAVA_HOME/bin/java -Djava.security.properties=$ORACLE_HOME/sqlplus/admin/iplus/provider -jar $ORACLE_HOME/oc4j/j2ee/home/jazn.jar -user "iSQL*Plus DBA/admin" -password admin_password -setpasswd "iSQL*Plus DBA" username old_password new_password Test iSQL*Plus DBA Access Test iSQL*Plus DBA access by entering the iSQL*Plus DBA URL in your web browser: " target=_blankhttp://machine_name.domain:5560/isqlplus/dba A dialog is displayed requesting authentication for the iSQL*Plus DBA URL. Log in as the user you created above. You may need to restart iSQL*Plus for the changes to take effect. Help us improve our service. Please email us your comments for this document. . What is a wire protocol ODBC driver? ==================================== A DBMS is written using an application programming interface (API), which is specific to that database. For example, an Oracle 9i database has its own version of the API specification (called Net9), which must run on each client application. Developers write applications compliant to the ODBC specification and use ODBC drivers to access the database. The ODBC driver communicates with the vendor's native API. Then, the native API passes instructions to another vendor-specific low-level API. Finally the wire protocol API communicates with the database. The wire protocol architecture eliminates the need for the database's native API (for example, Net9), so the driver communicates directly to the database through the database's own wire level protocol. This effectively removes an entire communication layer. ================================ 33. ADDM and other 10g features: ================================ ========================= 33.1 Flash_recovery_area: ========================= Note 1: ------- A flash recovery area is a directory, file system, or Automatic Storage Management disk group that serves as the default storage area for files related to recovery. Such files include - Multiplexed copies of the control file and online redo logs - Archived redo logs and flashback logs - RMAN backups - Files created by RESTORE and RECOVER commands Recovery components of the database interact with the flash recovery area to ensure that the database is completely recoverable using files in the flash recovery area. The database manages the disk space in the flash recovery area, and when there is not sufficient disk space to create new files, the database creates more room automatically by deleting the minimum set of files from flash recovery area that are obsolete, backed up to tertiary storage, or redundant. Note 2: ------- Before any Flash Backup and Recovery activity can take place, the Flash Recovery Area must be set up. The Flash Recovery Area is a specific area of disk storage that is set aside exclusively for retention of backup components such as datafile image copies, archived redo logs, and control file autobackup copies. These features include: Unified Backup Files Storage. All backup components can be stored in one consolidated spot. The Flash Recovery Area is managed via Oracle Managed Files (OMF), and it can utilize disk resources managed by Oracle Automated Storage Management (ASM). In addition, the Flash Recovery Area can be configured for use by multiple database instances if so desired. Automated Disk-Based Backup and Recovery. Once the Flash Recovery Area is configured, all backup components (datafile image copies, archived redo logs, and so on) are managed automatically by Oracle. Automatic Deletion of Backup Components. Once backup components have been successfully created, RMAN can be configured to automatically clean up files that are no longer needed (thus reducing risk of insufficient disk space for backups). Disk Cache for Tape Copies. Finally, if your disaster recovery plan involves backing up to alternate media, the Flash Recovery Area can act as a disk cache area for those backup components that are eventually copied to tape. Flashback Logs. The Flash Recovery Area is also used to store and manage flashback logs, which are used during Flashback Backup operations to quickly restore a database to a prior desired state. Sizing the Flash Recovery Area. Oracle recommends that the Flash Recovery Area should be sized large enough to include all files required for backup and recovery. However, if insufficient disk space is available, Oracle recommends that it be sized at least large enough to contain any archived redo logs that have not yet been backed up to alternate media. initialization parameters: DB_RECOVERY_FILE_DEST_SIZE specifies the total size of all files that can be stored in the Flash Recovery Area. Note that Oracle recommends setting this value first. DB_RECOVERY_FILE_DEST specifies the physical disk location where the Flashback Recovery Area will be stored. Oracle recommends that this be a separate location from the database's datafiles, control files, and redo logs. Also, note that if the database is using Oracle's new Automatic Storage Management (ASM) feature, then the shared disk area that ASM manages can be targeted for the Flashback Recovery Area. Examples: ----- -- Listing 2.2: Setting up the Flash Recovery Area - open database ----- -- Be sure to set DB_FILE_RECOVERY_DEST_SIZE first ... ALTER SYSTEM SET db_file_recovery_dest_size = '5G' SCOPE=BOTH SID='*'; -- ... and then set DB_FILE_RECOVERY_DEST and DB_FLASHBACK_RETENTION_TARGET ALTER SYSTEM SET db_file_recovery_dest = 'c:\oracle\fbrdata\zdcdb' SCOPE=BOTH SID='*'; ALTER SYSTEM SET db_flashback_retention_target = 2880; http://download.oracle.com/docs/cd/B19306_01/backup.102/b14192/toc.htm Note 2: ------- Flashback Database Demo An alternative strategy to the demo presented here is to use Recovery Manager RMAN> FLASHBACK DATABASE TO SCN = ; Dependent Objects GV_$FLASHBACK_DATABASE_LOG V_$FLASHBACK_DATABASE_LOG GV_$FLASHBACK_DATABASE_LOGFILE V_$FLASHBACK_DATABASE_LOGFILE GV_$FLASHBACK_DATABASE_STAT V_$FLASHBACK_DATABASE_STAT Syntax 1: SCN FLASHBACK [STANDBY] DATABASE [] TO [BEFORE] SCN Syntax 2: TIMESTAMP FLASHBACK [STANDBY] DATABASE [] TO [BEFORE] TIMESTMP Syntax 3: RESTORE POINT FLASHBACK [STANDBY] DATABASE [] TO [BEFORE] RESTORE POINT Flashback Syntax Elements OFF ALTER DATABASE FLASHBACK OFF alter database flashback off; ON ALTER DATABASE FLASHBACK ON alter database flashback on; Set Retention Target ALTER SYSTEM SET db_flashback_retention_target = ; alter system set DB_FLASHBACK_RETENTION_TARGET = 2880; Start flashback on a tablespace ALTER TABLESPACE FLASHBACK ON; alter tablespace example flashback on; Stop flashback on a tablespace ALTER TABLESPACE FLASHBACK OFF; alter tablespace example flashback off; Initialization Parameters Setting the location of the flashback recovery area db_recovery_file_dest=/oracle/flash_recovery_area Setting the size of the flashback recovery area db_recovery_file_dest_size=2147483648 Setting the retention time for flashback files (in minutes) -- 2 days db_flashback_retention_target=2880 Demo conn / as sysdba SELECT flashback_on, log_mode FROM gv$database; set linesize 121 col name format a30 col value format a30 SELECT name, value FROM gv$parameter WHERE name LIKE '%flashback%'; shutdown immediate; startup mount exclusive; alter database archivelog; alter database flashback on; alter database open; SELECT flashback_on, log_mode FROM gv$database; SELECT name, value FROM gv$parameter WHERE name LIKE '%flashback%'; -- 2 days alter system set DB_FLASHBACK_RETENTION_TARGET=2880; SELECT name, value FROM gv$parameter WHERE name LIKE '%flashback%'; SELECT estimated_flashback_size FROM gv$flashback_database_log; As SYS As UWCLASS SELECT current_scn FROM gv$database; SELECT oldest_flashback_scn, oldest_flashback_time FROM gv$flashback_database_log; create table t ( mycol VARCHAR2(20)) ROWDEPENDENCIES; INSERT INTO t VALUES ('ABC'); INSERT INTO t VALUES ('DEF'); COMMIT; INSERT INTO t VALUES ('GHI'); COMMIT; SELECT ora_rowscn, mycol FROM t; SHUTDOWN immediate; startup mount exclusive; FLASHBACK DATABASE TO SCN 19513917; /* FLASHBACK DATABASE TO TIMESTAMP (SYSDATE-1/24); FLASHBACK DATABASE TO TIMESTAMP timestamp'2002-11-05 14:00:00'; FLASHBACK DATABASE TO TIMESTAMP to_timestamp('2002-11-11 16:00:00', 'YYYY-MM-DD HH24:MI:SS'); */ alter database open; alter database open resetlogs; conn uwclass/uwclass SELECT ora_rowscn, mycol FROM t; SELECT * FROM gv$flashback_database_stat; alter system switch logfile; shutdown immediate; startup mount exclusive; alter database flashback off; alter database noarchivelog; alter database open; SELECT flashback_on, log_mode FROM gv$database; host rman target sys/pwd@orabase RMAN> crosscheck archivelog all; RMAN> delete archivelog all; RMAN> list archivelog all; -- if out of disk space ORA-16014: log 2 sequence# 4163 not archived, no available destinations ORA-00312: online log 2 thread 1: 'c:\oracle\oradata\orabase\redo02.log' -- what happens The error ora-16014 is the real clue for this problem. Once the archive destination becomes full the location also becomes invalid. Normally Oracle does not do a recheck to see if space has been made available. -- then shutdown abort; -- clean up disk space: then startup alter system archive log all to '/oracle/flash_recovery_area/ORABASE/ARCHIVELOG'; ========== 33.2 ADDM: ========== Note 1: ======= Doc ID: Note:250655.1 Content Type: TEXT/PLAIN Subject: How to use the Automatic Database Diagnostic Monitor Creation Date: 09-OCT-2003 Type: BULLETIN Last Revision Date: 10-JUN-2004 Status: PUBLISHED PURPOSE ------- The purpose of this article is to show an introduction on how to use the Automatic Database Diagnostic Monitor feature. The ADDM consists of functionality built into the Oracle kernel to assist in making tuning an Oracle instance less elaborate. SCOPE & APPLICATION ------------------- Audience : Oracle developers and DBAs Use : Using the Automatic Database Diagnostic Monitor feature as a first step in the creation of an autotunable database Level of detail : medium Limitation on use: none USING THE AUTOMATIC DATABASE DIAGNOSTIC MONITOR ----------------------------------------------- Introduction: ------------- The Automatic Database Diagnostic Monitor (hereafter called ADDM) is an integral part of the Oracle RDBMS capable of gathering performance statistics and advising on changes to solve any exitsing performance issues measured. For this it uses the Automatic Workload Repository ( hereafter called AWR), a repository defined in the database to store database wide usage statistics at fixed size intervals (60 minutes). To make use of ADDM, a PL/SQL interface called DBMS_ADVISOR has been implemented. This PL/SQL interface may be called through the supplied $ORACLE_HOME/rdbms/admin/addmrpt.sql script, called directly, or used in combination with the Oracle Enterprise Manager application. Besides this PL/SQL package, a number of views (with names starting with the DBA_ADVISOR_ prefix) allow retrieval of the results of any actions performed with the DBMS_ADVISOR API. The preferred way of accessing ADDM is through the Enterprise Manager interface, as it shows a complete performance overview including recommendations on how to solve bottlenecks on a single screen. When accessing ADDM manually, you should consider using the ADDMRPT.SQL script provided with your Oracle release, as it hides the complexities involved in accessing the DBMS_ADVISOR package. To use ADDM for advising on how to tune the instance and SQL, you need to make sure that the AWR has been populated with at least 2 sets of performance data. When the STATISTICS_LEVEL is set to TYPICAL or ALL the database will automatically schedule the AWR to be populated at 60 minute intervals. When you wish to create performance snapshots outside of the fixed intervals, then you can use the DBMS_WORKLOAD_REPOSITORY package for this, like in: BEGIN DBMS_WORKLOAD_REPOSITORY.CREATE_SNAPSHOT('TYPICAL'); END; / The snapshots need be created before and after the action you wish to examine. E.g. when examining a bad performing query, you need to have performance data snapshots from the timestamps before the query was started and after the query finished. You may also change the frequency of the snapshots and the duration for which they are saved in the AWR. Use the DBMS_WORKLOAD_REPOSITORY package as in the following example: execute DBMS_WORKLOAD_REPOSITORY.MODIFY_SNAPSHOT_SETTINGS(interval=>60,retention=>43200); Example: -------- You can use ADDM through the PL/SQL API and query the various advisory views in SQL*Plus to examine how to solve performance issues. The example is based on the SCOTT account executing the various tasks. To allow SCOTT to both generate AWR snapshots and sumbit ADDM recommendation jobs, he needs to be granted proper access: CONNECT / AS SYSDBA GRANT ADVISOR TO scott; GRANT SELECT_CATALOG_ROLE TO scott; GRANT EXECUTE ON dbms_workload_repository TO scott; Furthermore, the buffer cache size (DB_CACHE_SIZE) has been reduced to 24M. The example presented makes use of a table called BIGEMP, residing in the SCOTT schema. The table (containing about 14 million rows) has been created with: CONNECT scott/tiger CREATE TABLE bigemp AS SELECT * FROM emp; ALTER TABLE bigemp MODIFY (empno NUMBER); DECLARE n NUMBER; BEGIN FOR n IN 1..18 LOOP INSERT INTO bigemp SELECT * FROM bigemp; END LOOP; COMMIT; END; / UPDATE bigemp SET empno = ROWNUM; COMMIT; The next step is to generate a performance data snapshot: EXECUTE dbms_workload_repository.create_snapshot('TYPICAL'); Execute a query on the BIGEMP table to generate some load: SELECT * FROM bigemp WHERE deptno = 10; After this, generate a second performance snapshot: EXECUTE dbms_workload_repository.create_snapshot('TYPICAL'); The easiest way to get the ADDM report is by executing: @?/rdbms/admin/addmrpt Running this script will show which snapshots have been generated, asks for the snapshot IDs to be used for generating the report, and will generate the report containing the ADDM findings. When you do not want to use the script, you need to submit and execute the ADDM task manually. First, query DBA_HIST_SNAPSHOT to see which snapshots have been created. These snapshots will be used by ADDM to generate recommendations: SELECT * FROM dba_hist_snapshot ORDER BY snap_id; SNAP_ID DBID INSTANCE_NUMBER ---------- ---------- --------------- STARTUP_TIME ----------------------------------------------------------------------- BEGIN_INTERVAL_TIME ----------------------------------------------------------------------- END_INTERVAL_TIME ----------------------------------------------------------------------- FLUSH_ELAPSED ----------------------------------------------------------------------- SNAP_LEVEL ERROR_COUNT ---------- ----------- 1 494687018 1 17-NOV-03 09.39.17.000 AM 17-NOV-03 09.39.17.000 AM 17-NOV-03 09.50.21.389 AM +00000 00:00:06.6 1 0 2 494687018 1 17-NOV-03 09.39.17.000 AM 17-NOV-03 09.50.21.389 AM 17-NOV-03 10.29.35.704 AM +00000 00:00:02.3 1 0 3 494687018 1 17-NOV-03 09.39.17.000 AM 17-NOV-03 10.29.35.704 AM 17-NOV-03 10.35.46.878 AM +00000 00:00:02.1 1 0 Mark the 2 snapshot IDs (such as the lowest and highest ones) for use in generating recommendations. Next, you need to submit and execute the ADDM task manually, using a script similar to: DECLARE task_name VARCHAR2(30) := 'SCOTT_ADDM'; task_desc VARCHAR2(30) := 'ADDM Feature Test'; task_id NUMBER; BEGIN (1) dbms_advisor.create_task('ADDM', task_id, task_name, task_desc, null); (2) dbms_advisor.set_task_parameter('SCOTT_ADDM', 'START_SNAPSHOT', 1); dbms_advisor.set_task_parameter('SCOTT_ADDM', 'END_SNAPSHOT', 3); dbms_advisor.set_task_parameter('SCOTT_ADDM', 'INSTANCE', 1); dbms_advisor.set_task_parameter('SCOTT_ADDM', 'DB_ID', 494687018); (3) dbms_advisor.execute_task('SCOTT_ADDM'); END; / Here is the explanation of the steps you need to take to successfully execute an ADDM job: 1) The first step is to create the task. For this, you need to specify the name under which the task will be known in the ADDM task system. Along with the name you can provide a more readable description on what the job should do. The task type must be 'ADDM' in order to have it executed in the ADDM environment. 2) After having defined the ADDM task, you must define the boundaries within which the task needs to be executed. For this you need to set the starting and ending snapshot IDs, instance ID (especially necessary when running in a RAC environment), and database ID for the newly created job. 3) Finally, the task must be executed. When querying DBA_ADVISOR_TASKS you see the just created job: SELECT * FROM dba_advisor_tasks; OWNER TASK_ID TASK_NAME ------------------------------ ---------- ------------------------------ DESCRIPTION ------------------------------------------------------------------------ ADVISOR_NAME CREATED LAST_MODI PARENT_TASK_ID ------------------------------ --------- --------- -------------- PARENT_REC_ID READ_ ------------- ----- SCOTT 5 SCOTT_ADDM ADDM Feature Test ADDM 17-NOV-03 17-NOV-03 0 0 FALSE When the job has successfully completed, examine the recommendations made by ADDM by calling the DBMS_ADVISOR.GET_TASK_REPORT() routine, like in: SET LONG 1000000 PAGESIZE 0 LONGCHUNKSIZE 1000 COLUMN get_clob FORMAT a80 SELECT dbms_advisor.get_task_report('SCOTT_ADDM', 'TEXT', 'TYPICAL') FROM sys.dual; The recommendations supplied should be sufficient to investigate the performance issue, as in: DETAILED ADDM REPORT FOR TASK 'SCOTT_ADDM' WITH ID 5 ---------------------------------------------------- Analysis Period: 17-NOV-2003 from 09:50:21 to 10:35:47 Database ID/Instance: 494687018/1 Snapshot Range: from 1 to 3 Database Time: 4215 seconds Average Database Load: 1.5 active sessions ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ FINDING 1: 65% impact (2734 seconds) ------------------------------------ PL/SQL execution consumed significant database time. RECOMMENDATION 1: SQL Tuning, 65% benefit (2734 seconds) ACTION: Tune the PL/SQL block with SQL_ID fjxa1vp3yhtmr. Refer to the "Tuning PL/SQL Applications" chapter of Oracle's "PL/SQL User's Guide and Reference" RELEVANT OBJECT: SQL statement with SQL_ID fjxa1vp3yhtmr BEGIN EMD_NOTIFICATION.QUEUE_READY(:1, :2, :3); END; FINDING 2: 35% impact (1456 seconds) ------------------------------------ SQL statements consuming significant database time were found. RECOMMENDATION 1: SQL Tuning, 35% benefit (1456 seconds) ACTION: Run SQL Tuning Advisor on the SQL statement with SQL_ID gt9ahqgd5fmm2. RELEVANT OBJECT: SQL statement with SQL_ID gt9ahqgd5fmm2 and PLAN_HASH 547793521 UPDATE bigemp SET empno = ROWNUM FINDING 3: 20% impact (836 seconds) ----------------------------------- The throughput of the I/O subsystem was significantly lower than expected. RECOMMENDATION 1: Host Configuration, 20% benefit (836 seconds) ACTION: Consider increasing the throughput of the I/O subsystem. Oracle's recommended solution is to stripe all data file using the SAME methodology. You might also need to increase the number of disks for better performance. RECOMMENDATION 2: Host Configuration, 14% benefit (584 seconds) ACTION: The performance of file D:\ORACLE\ORADATA\V1010\UNDOTBS01.DBF was significantly worse than other files. If striping all files using the SAME methodology is not possible, consider striping this file over multiple disks. RELEVANT OBJECT: database file "D:\ORACLE\ORADATA\V1010\UNDOTBS01.DBF" SYMPTOMS THAT LED TO THE FINDING: Wait class "User I/O" was consuming significant database time. (34% impact [1450 seconds]) FINDING 4: 11% impact (447 seconds) ----------------------------------- Undo I/O was a significant portion (33%) of the total database I/O. NO RECOMMENDATIONS AVAILABLE SYMPTOMS THAT LED TO THE FINDING: The throughput of the I/O subsystem was significantly lower than expected. (20% impact [836 seconds]) Wait class "User I/O" was consuming significant database time. (34% impact [1450 seconds]) FINDING 5: 9.9% impact (416 seconds) ------------------------------------ Buffer cache writes due to small log files were consuming significant database time. RECOMMENDATION 1: DB Configuration, 9.9% benefit (416 seconds) ACTION: Increase the size of the log files to 796 M to hold at least 20 minutes of redo information. SYMPTOMS THAT LED TO THE FINDING: The throughput of the I/O subsystem was significantly lower than expected. (20% impact [836 seconds]) Wait class "User I/O" was consuming significant database time. (34% impact [1450 seconds]) FINDING 6: 9.2% impact (387 seconds) ------------------------------------ Individual database segments responsible for significant user I/O wait were found. RECOMMENDATION 1: Segment Tuning, 7.2% benefit (304 seconds) ACTION: Run "Segment Advisor" on database object "SCOTT.BIGEMP" with object id 49634. RELEVANT OBJECT: database object with id 49634 ACTION: Investigate application logic involving I/O on database object "SCOTT.BIGEMP" with object id 49634. RELEVANT OBJECT: database object with id 49634 RECOMMENDATION 2: Segment Tuning, 2% benefit (83 seconds) ACTION: Run "Segment Advisor" on database object "SYSMAN.MGMT_METRICS_RAW_PK" with object id 47084. RELEVANT OBJECT: database object with id 47084 ACTION: Investigate application logic involving I/O on database object "SYSMAN.MGMT_METRICS_RAW_PK" with object id 47084. RELEVANT OBJECT: database object with id 47084 SYMPTOMS THAT LED TO THE FINDING: Wait class "User I/O" was consuming significant database time. (34% impact [1450 seconds]) FINDING 7: 8.7% impact (365 seconds) ------------------------------------ Individual SQL statements responsible for significant physical I/O were found. RECOMMENDATION 1: SQL Tuning, 8.7% benefit (365 seconds) ACTION: Run SQL Tuning Advisor on the SQL statement with SQL_ID gt9ahqgd5fmm2. RELEVANT OBJECT: SQL statement with SQL_ID gt9ahqgd5fmm2 and PLAN_HASH 547793521 UPDATE bigemp SET empno = ROWNUM RECOMMENDATION 2: SQL Tuning, 0% benefit (0 seconds) ACTION: Tune the PL/SQL block with SQL_ID fjxa1vp3yhtmr. Refer to the "Tuning PL/SQL Applications" chapter of Oracle's "PL/SQL User's Guide and Reference" RELEVANT OBJECT: SQL statement with SQL_ID fjxa1vp3yhtmr BEGIN EMD_NOTIFICATION.QUEUE_READY(:1, :2, :3); END; SYMPTOMS THAT LED TO THE FINDING: The throughput of the I/O subsystem was significantly lower than expected. (20% impact [836 seconds]) Wait class "User I/O" was consuming significant database time. (34% impact [1450 seconds]) FINDING 8: 8.3% impact (348 seconds) ------------------------------------ Wait class "Configuration" was consuming significant database time. NO RECOMMENDATIONS AVAILABLE ADDITIONAL INFORMATION: Waits for free buffers were not consuming significant database time. Waits for archiver processes were not consuming significant database time. Log file switch operations were not consuming significant database time while waiting for checkpoint completion. Log buffer space waits were not consuming significant database time. High watermark (HW) enqueue waits were not consuming significant database time. Space Transaction (ST) enqueue waits were not consuming significant database time. ITL enqueue waits were not consuming significant database time. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ADDITIONAL INFORMATION ---------------------- An explanation of the terminology used in this report is available when you run the report with the 'ALL' level of detail. The analysis of I/O performance is based on the default assumption that the average read time for one database block is 5000 micro-seconds. Wait class "Administrative" was not consuming significant database time. Wait class "Application" was not consuming significant database time. Wait class "Cluster" was not consuming significant database time. Wait class "Commit" was not consuming significant database time. Wait class "Concurrency" was not consuming significant database time. CPU was not a bottleneck for the instance. Wait class "Network" was not consuming significant database time. Wait class "Scheduler" was not consuming significant database time. Wait class "Other" was not consuming significant database time. ============================= END OF ADDM REPORT ====================== ADDM points out which events cause the performance problems to occur and suggests directions to follow to fix these bottlenecks. The ADDM recommendations show amongst others that the query on BIGEMP needs to be examined; in this case it suggests to run the Segment Advisor to check whether the data segment is fragmented or not; it also advices to check the application logic involved in accessing the BIGEMP table. Furthermore, it shows the system suffers from I/O problems (which is in this example caused by not using SAME and placing all database files on a single disk partition). The findings are sorted descending by impact: the issues causing the greatest performance problems are listed at the top of the report. Solving these issues will result in the greatest performance benefits. Also, in the last section of the report ADDM indicates the areas that are not representing a problem for the performance of the instance In this example the database is rather idle. As such the Enterprise Manager notification job (which runs frequently) is listed at the top. You need not worry about this job at all. Please notice that the output of the last query may differ depending on what took place on your database at the time the ADDM recommendations were generated. RELATED DOCUMENTS ----------------- Oracle10g Database Performance Guide Release 1 (10.1) Oracle10g Database Reference Release 1 (10.1) PL/SQL Packages and Types Reference Release 1 (10.1) Note 2: ======= To determine which segments will benefit from segment shrink, you can invoke Segment Advisor. alter table hr.employees enable row movement; After the Segment Advisor has been invoked to give recommendations, the findings are available in BDA_ADVISOR_FINDINGS and DBA_ADVISOR_RECOMMENDATIONS. variable task_id number; declare name varchar2(100); desc varchar2(500); obj_id number; begin name:=''; desc:='Check HR.EMPLOYEE'; DBMS_ADVISOR.CREATE_TASK('Segment Advisor', :task_id, name, descr, NULL); DBMS_ADVISOR.CREATE_OBJECT(name,'TABLE','HR','EMPLOYEES', NULL,NULL,obj_id); DBMS_ADVISOR.SET_TASK_PARAMETER(name,'RECOMMEND_ALL','TRUE'); DBMS_ADVISOR.EXECUTE_TASK(name); end; PL/SQL procedure successfully completed. print task_id TASK_ID ------- 6 SELECT owner, task_id, task_name, type, message, more_info FROM DBA_ADVISOR_FINDINGS WHERE task_id=6; OWNER TASK_ID TASK_NAME TYPE MESSAGE ----- ------- --------- ---- -------------------------------------------------- RJB 6 TASK_00003 INFORMATION Perform shrink, estimated savings is 107602 bytes. In DBA_ADVISOR_ACTIONS, you can even find the exact SQL statement to shrink the hr.employees segment. alter table hr.employees shrink space; ============================== 34. ASM and RAC in Oracle 10g: ============================== 34.1 ASM ======== ======== Note 1: ======== Automatic Storage Management (ASM) in Oracle Database 10g With ASM, Automatic Storage Management, there is a separate lightweight 10g database involved. This ASM database (+ASM), contains all metadata about the ASM system. It also acts as the interface between the regular database and the filesystems. ASM will provide for presentation and implementation of a special filesystem, on which a number of redundancy/availability and performance features are implemented. In addition to the normal database background processes like CKPT, DBWR, LGWR, SMON, and PMON, an ASM instance uses at least two additional background processes to manage data storage operations. The Rebalancer process, RBAL, coordinates the rebalance activity for ASM disk groups, and the Actual ReBalance processes, ARBn, handle the actual rebalance of data extent movements. There are usually several ARB background processes (ARB0, ARB1, and so forth). Every database instance that uses ASM for file storage, will also need two new processes. The Rebalancer background process (RBAL) handles global opens of all ASM disks in the ASM Disk Groups, while the ASM Bridge process (ASMB) connects as a foreground process into the ASM instance when the regular database instance starts. ASMB facilitates communication between the ASM instance and the regular database, including handling physical file changes like data file creation and deletion. ASMB exchanges messages between both servers for statistics update and instance health validation. These two processes are automatically started by the database instance when a new Oracle file type - for example, a tablespace's datafile -- is created on an ASM disk group. When an ASM instance mounts a disk group, it registers the disk group and connect string with Group Services. The database instance knows the name of the disk group, and can therefore use it to locate connect information for the correct ASM instance. ======== Note 2: ======== Some terminology in RAC: CRS cluster ready services - Clusterware: For Oracle10g on Linux and Windows-based platforms, CRS co-exists with but does not inter-operate with vendor clusterware. You may use vendor clusterware for all UNIX-based operating systems except for Linux. Even though, many of the Unix platforms have their own clusterware products, you need to use the CRS software to provide the HA support services. CRS (cluster ready services) supports services and workload management and helps to maintain the continuous availability of the services. CRS also manages resources such as virtual IP (VIP) address for the node and the global services daemon. Note that the "Voting disks" and the "Oracle Cluster Registry", are regarded as part of the CRS. OCR: The Oracle Cluster Registry (OCR) contains cluster and database configuration information for Real Application Clusters Cluster Ready Services (CRS), including the list of nodes in the cluster database, the CRS application, resource profiles, and the authorizations for the Event Manager (EVM). The OCR can reside in a file on a cluster file system or on a shared raw device. When you install Real Application Clusters, you specify the location of the OCR. OCFS: OCFS is a shared disk cluster filesystem. Version 1 released for Linux is specifically designed to alleviate the need for manag-ing raw devices. It can contain all the oracle datafiles, archive log files and controlfiles. It is however not designed as a general purpose filesystem. OCFS2 is the next generation of the Oracle Cluster File System for Linux. It is an extent based, POSIX compliant file system. Unlike the previous release (OCFS), OCFS2 is a general-purpose file system that can be used for shared Oracle home installations making management of Oracle Real Application Cluster (RAC) installations even easier. Among the new features and benefits are: Node and architecture local files using Context Dependent Symbolic Links (CDSL) Network based pluggable DLM Improved journaling / node recovery using the Linux Kernel "JBD" subsystem Improved performance of meta-data operations (space allocation, locking, etc). Improved data caching / locking (for files such as oracle binaries, libraries, etc) - OCFS1 does NOT support a shared Oracle Home - OCFS2 does support a shared Oracle Home Though ASM appears to be the intended replacement for Oracle Cluster File System (OCFS) for the Real Applications Cluster (RAC). ASM supports Oracle Real Application Clusters (RAC), so there is no need for a separate Cluster LVM or a Cluster File System. So it boils down to: - You use or OCFS2, or RAW, or ASM (preferrably) for your database files. Storage Option Oracle Clusterware Database Recovery area -------------- ------------------ -------- ------------- Automatic Storage Management No Yes Yes Cluster file system (OCFS) Yes Yes Yes Shared raw storage Yes Yes No ======== Note 3: ======== Automatic Storage Management (ASM) simplifies database administration. It eliminates the need for you, as a DBA, to directly manage potentially thousands of Oracle database files. It does this by enabling you to create disk groups, which are comprised of disks and the files that reside on them. You only need to manage a small number of disk groups. In the SQL statements that you use for creating database structures such as tablespaces, redo log and archive log files, and control files, you specify file location in terms of disk groups. Automatic Storage Management then creates and manages the associated underlying files for you. Automatic Storage Management extends the power of Oracle-managed files. With Oracle-managed files, files are created and managed automatically for you, but with Automatic Storage Management you get the additional benefits of features such as mirroring and striping. The primary component of Automatic Storage Management is the disk group. You configure Automatic Storage Management by creating disk groups, which, in your database instance, can then be specified as the default location for files created in the database. Oracle provides SQL statements that create and manage disk groups, their contents, and their metadata. A disk group consists of a grouping of disks that are managed together as a unit. These disks are referred to as ASM disks. Files written on ASM disks are ASM files, whose names are automatically generated by Automatic Storage Management. You can specify user-friendly alias names for ASM files, but you must create a hierarchical directory structure for these alias names. You can affect how Automatic Storage Management places files on disks by specifying failure groups. Failure groups define disks that share components, such that if one fails then other disks sharing the component might also fail. An example of what you might define as a failure group would be a set of SCSI disks sharing the same SCSI controller. Failure groups are used to determine which ASM disks to use for storing redundant data. For example, if two-way mirroring is specified for a file, then redundant copies of file extents must be stored in separate failure groups. If you would take a look at the v$datafile, v$logfile, and v$controlfile of the regular Database, you would see information like in the following example: SQL> select file#, name from v$datafile; 1 +DATA1/rac0/datafile/system.256.1 2 +DATA1/rac0/datafile/undotbs.258.1 3 +DATA1/rac0/datafile/sysaux.257.1 4 +DATA1/rac0/datafile/users.259.1 5 +DATA1/rac0/datafile/example.269.1 SQL> select name from v$controlfile; +DATA1/rac0/controlfile/current.261.3 +DATA1/rac0/controlfile/current.260.3 -- Initialization Parameters (init.ora or SPFILE) for ASM Instances The following initialization parameters relate to an ASM instance. Parameters that start with ASM_ cannot be set in database instances. Name Description INSTANCE_TYPE Must be set to INSTANCE_TYPE = ASM. Note: This is the only required parameter. All other parameters take suitable defaults for most environments. DB_UNIQUE_NAME Unique name for this group of ASM instances within the cluster or on a node. Default: +ASM (Needs to be modified only if trying to run multiple ASM instances on the same node) ASM_POWER_LIMIT The maximum power on an ASM instance for disk rebalancing. Default: 1 Can range from 1 to 11. 1 is the lowest priority. See Also: "Tuning Rebalance Operations" ASM_DISKSTRING Limits the set of disks that Automatic Storage Management considers for discovery. Default: NULL (This default causes ASM to find all of the disks in a platform-specific location to which it has read/write access.). Example: /dev/raw/* ASM_DISKGROUPS Lists the names of disk groups to be mounted by an ASM instance at startup, or when the ALTER DISKGROUP ALL MOUNT statement is used. Default: NULL (If this parameter is not specified, then no disk groups are mounted.) Note: This parameter is dynamic and if you are using a server parameter file (SPFILE), then you should rarely need to manually alter this value. Automatic Storage Management automatically adds a disk group to this parameter when a disk group is successfully mounted, and automatically removes a disk group that is specifically dismounted. However, when using a traditional text initialization parameter file, remember to edit the initialization parameter file to add the name of any disk group that you want automatically mounted at instance startup, and remove the name of any disk group that you no longer want automatically mounted. -- ASM Views: The ASM configuration can be viewed using the V$ASM_% views, which often contain different information depending on whether they are queried from the ASM instance, or a dependant database instance. Viewing ASM Instance Information Via SQL Queries Finally, there are several dynamic and data dictionary views available to view an ASM configuration from within the ASM instance itself: ASM Dynamic Views: FROM ASM Instance Information View Name Description V$ASM_ALIAS Shows every alias for every disk group mounted by the ASM instance V$ASM_CLIENT Shows which database instance(s) are using any ASM disk groups that are being mounted by this ASM instance V$ASM_DISK Lists each disk discovered by the ASM instance, including disks that are not part of any ASM disk group V$ASM_DISKGROUP Describes information about ASM disk groups mounted by the ASM instance V$ASM_FILE Lists each ASM file in every ASM disk group mounted by the ASM instance V$ASM_OPERATION Like its counterpart, V$SESSION_LONGOPS, it shows each long-running ASM operation in the ASM instance V$ASM_TEMPLATE Lists each template present in every ASM disk group mounted by the ASM instance -- Managing disk groups The SQL statements introduced in this section are only available in an ASM instance. You must first start the ASM instance. Creating disk group examples: Example 1: ---------- Creating a Disk Group: Example The following examples assume that the ASM_DISKSTRING is set to '/devices/*'. Assume the following: ASM disk discovery identifies the following disks in directory /devices. /devices/diska1 /devices/diska2 /devices/diska3 /devices/diska4 /devices/diskb1 /devices/diskb2 /devices/diskb3 /devices/diskb4 The disks diska1 - diska4 are on a separate SCSI controller from disks diskb1 - diskb4. The following SQL*Plus session illustrates starting an ASM instance and creating a disk group named dgroup1. % SQLPLUS /NOLOG SQL> CONNECT / AS SYSDBA SQL> CREATE DISKGROUP dgroup1 NORMAL REDUNDANCY 2 FAILGROUP controller1 DISK 3 '/devices/diska1', 4 '/devices/diska2', 5 '/devices/diska3', 6 '/devices/diska4', 7 FAILGROUP controller2 DISK 8 '/devices/diskb1', 9 '/devices/diskb2', 10 '/devices/diskb3', 11 '/devices/diskb4'; In this example, dgroup1 is composed of eight disks that are defined as belonging to either failure group controller1 or controller2. Since NORMAL REDUNDANCY level is specified for the disk group, then Automatic Storage Management provides redundancy for all files created in dgroup1 according to the attributes specified in the disk group templates. For example, in the system default template shown in the table in "Managing Disk Group Templates", normal redundancy for the online redo log files (ONLINELOG template) is two-way mirroring. This means that when one copy of a redo log file extent is written to a disk in failure group controller1, a mirrored copy of the file extent is written to a disk in failure group controller2. You can see that to support normal redundancy level, at least two failure groups must be defined. Since no NAME clauses are provided for any of the disks being included in the disk group, the disks are assigned the names of dgroup1_0001, dgroup1_0002, ..., dgroup1_0008. Example 2: ---------- CREATE DISKGROUP disk_group_1 NORMAL REDUNDANCY FAILGROUP failure_group_1 DISK '/devices/diska1' NAME diska1, '/devices/diska2' NAME diska2, FAILGROUP failure_group_2 DISK '/devices/diskb1' NAME diskb1, '/devices/diskb2' NAME diskb2; Example 3: ---------- At some point in using OUI in installing the software, and creating a database, you will see the following screen: ---------------------------------------------------- |SPECIFY Database File Storage Option | | | | o File system | | Specify Database file location: ######### | | | | o Automatic Storage Management (ASM) | | | | o Raw Devices | | | | Specify Raw Devices mapping file: ########## | ---------------------------------------------------- Suppose that you have on a Linux machine the following raw disk devices: /dev/raw/raw1 8GB /dev/raw/raw2 8GB /dev/raw/raw3 6GB /dev/raw/raw4 6GB /dev/raw/raw5 6GB /dev/raw/raw6 6GB Then you can choose ASM in the upper screen, and see the following screen, where you can create the initial diskgroup and assign disks to it: ----------------------------------------------------- | Configure Automatic Storage Management | | | | Disk Group Name: data1 | | | | Redundancy | | o High o Normal o External | | | | Add member Disks | | |-------------------------------- | | | select Disk Path | | | |[#] /dev/raw/raw1 | | | |[#] /dev/raw/raw2 | | | |[ ] /dev/raw/raw3 | | | |[ ] /dev/raw/raw4 | | | -------------------------------- | | | ----------------------------------------------------- -- Mounting and Dismounting Disk Groups Disk groups that are specified in the ASM_DISKGROUPS initialization parameter are mounted automatically at ASM instance startup. This makes them available to all database instances running on the same node as Automatic Storage Management. The disk groups are dismounted at ASM instance shutdown. Automatic Storage Management also automatically mounts a disk group when you initially create it, and dismounts a disk group if you drop it. There may be times that you want to mount or dismount disk groups manually. For these actions use the ALTER DISKGROUP ... MOUNT or ALTER DISKGROUP ... DISMOUNT statement. You can mount or dismount disk groups by name, or specify ALL. If you try to dismount a disk group that contains open files, the statement will fail, unless you also specify the FORCE clause. Example The following statement dismounts all disk groups that are currently mounted to the ASM instance: ALTER DISKGROUP ALL DISMOUNT; The following statement mounts disk group dgroup1: ALTER DISKGROUP dgroup1 MOUNT; ======== Note 4: ======== -- Installing Oracle ASMLib for Linux: ASMLib is a support library for the Automatic Storage Management feature of Oracle Database 10g. This document is a set of tips for installing the Linux specific ASM library and its assocated driver. This library is provide to enable ASM I/O to Linux disks without the limitations of the standard Unix I/O API. The steps below are steps that the system administrator must follow. The ASMLib software is available from the Oracle Technology Network. Go to ASMLib download page and follow the link for your platform. You will see 4-6 packages for your Linux platform. -The oracleasmlib package provides the actual ASM library. -The oracleasm-support package provides the utilities used to get the ASM driver up and running. Both of these packages need to be installed. -The remaining packages provide the kernel driver for the ASM library. Each package provides the driver for a different kernel. You must install the appropriate package for the kernel you are running. Use the "uname -r command to determine the version of the kernel. The oracleasm kerel driver package will have that version string in its name. For example, if you were running Red Hat Enterprise Linux 4 AS, and the kernel you were using was the 2.6.9-5.0.5.ELsmp kernel, you would choose the oracleasm-2.6.9-5.0.5-ELsmp package. So, for example, to install these packages on RHEL4 on an Intel x86 machine, you might use the command: rpm -Uvh oracleasm-support-2.0.0-1.i386.rpm \ oracleasm-lib-2.0.0-1.i386.rpm \ oracleasm-2.6.9-5.0.5-ELsmp-2.0.0-1.i686.rpm Once the command completes, ASMLib is now installed on the system. -- Configuring ASMLib: Now that the ASMLib software is installed, a few steps have to be taken by the system administrator to make the ASM driver available. The ASM driver needs to be loaded, and the driver filesystem needs to be mounted. This is taken care of by the initialization script, "/etc/init.d/oracleasm". Run the "/etc/init.d/oracleasm" script with the "configure" option. It will ask for the user and group that default to owning the ASM driver access point. If the database was running as the 'oracle' user and the 'dba' group, the output would look like this: [root@ca-test1 /]# /etc/init.d/oracleasm configure Configuring the Oracle ASM library driver. This will configure the on-boot properties of the Oracle ASM library driver. The following questions will determine whether the driver is loaded on boot and what permissions it will have. The current values will be shown in brackets ('[]'). Hitting without typing an answer will keep that current value. Ctrl-C will abort. Default user to own the driver interface []: oracle Default group to own the driver interface []: dba Start Oracle ASM library driver on boot (y/n) [n]: y Fix permissions of Oracle ASM disks on boot (y/n) [y]: y Writing Oracle ASM library driver configuration [ OK ] Creating /dev/oracleasm mount point [ OK ] Loading module "oracleasm" [ OK ] Mounting ASMlib driver filesystem [ OK ] Scanning system for ASM disks [ OK ] This should load the oracleasm.o driver module and mount the ASM driver filesystem. By selecting enabled = 'y' during the configuration, the system will always load the module and mount the filesystem on boot. The automatic start can be enabled or disabled with the 'enable' and 'disable' options to /etc/init.d/oracleasm: [root@ca-test1 /]# /etc/init.d/oracleasm disable Writing Oracle ASM library driver configuration [ OK ] Unmounting ASMlib driver filesystem [ OK ] Unloading module "oracleasm" [ OK ] [root@ca-test1 /]# /etc/init.d/oracleasm enable Writing Oracle ASM library driver configuration [ OK ] Loading module "oracleasm" [ OK ] Mounting ASMlib driver filesystem [ OK ] Scanning system for ASM disks [ OK ] -- Making Disks Available to ASMLib: The system administrator has one last task. Every disk that ASMLib is going to be accessing needs to be made available. This is accomplished by creating an ASM disk. The /etc/init.d/oracleasm script is again used for this task: [root@ca-test1 /]# /etc/init.d/oracleasm createdisk VOL1 /dev/sdg1 Creating Oracle ASM disk "VOL1" [ OK ] Disk names are ASCII capital letters, numbers, and underscores. They must start with a letter. Disks that are no longer used by ASM can be unmarked as well: [root@ca-test1 /]# /etc/init.d/oracleasm deletedisk VOL1 Deleting Oracle ASM disk "VOL1" [ OK ] Any operating system disk can be queried to see if it is used by ASM: [root@ca-test1 /]# /etc/init.d/oracleasm querydisk /dev/sdg1 Checking if device "/dev/sdg1" is an Oracle ASM disk [ OK ] [root@ca-test1 /]# /etc/init.d/oracleasm querydisk /dev/sdh1 Checking if device "/dev/sdh1" is an Oracle ASM disk [FAILED] Existing disks can be listed and queried: [root@ca-test1 /]# /etc/init.d/oracleasm listdisks VOL1 VOL2 VOL3 [root@ca-test1 /]# /etc/init.d/oracleasm querydisk VOL1 Checking for ASM disk "VOL1" [ OK ] When a disk is added to a RAC setup, the other nodes need to be notified about it. Run the 'createdisk' command on one node, and then run 'scandisks' on every other node: [root@ca-test1 /]# /etc/init.d/oracleasm scandisks Scanning system for ASM disks [ OK ] -- Discovery Strings for Linux ASMLib: ASMLib uses discovery strings to determine what disks ASM is asking for. The generic Linux ASMLib uses glob strings. The string must be prefixed with "ORCL:". Disks are specified by name. A disk created with the name "VOL1" can be discovered in ASM via the discovery string "ORCL:VOL1". Similarly, all disks that start with the string "VOL" can be queried with the discovery string "ORCL:VOL*". Disks cannot be discovered with path names in the discovery string. If the prefix is missing, the generic Linux ASMLib will ignore the discovery string completely, expecting that it is intended for a different ASMLib. The only exception is the empty string (""), which is considered a full wildcard. This is precisely equivalent to the discovery string "ORCL:*". NOTE: Once you mark your disks with Linux ASMLib, Oracle Database 10g R1 (10.1) OUI will not be able to discover your disks. It is recommended that you complete a Software Only install and then use DBCA to create your database (or use the custom install). ======== Note 5: ======== Automatic Storage Management (ASM) is a new feature that has be introduced in Oracle 10g to simplify the storage of Oracle datafiles, controlfiles and logfiles. - Overview of Automatic Storage Management (ASM) - Initialization Parameters and ASM Instance Creation - Startup and Shutdown of ASM Instances - Administering ASM Disk Groups - Disks - Templates - Directories - Aliases - Files - Checking Metadata - ASM Filenames - ASM Views - SQL and ASM - Migrating to ASM Using RMAN Overview of Automatic Storage Management (ASM) Automatic Storage Management (ASM) simplifies administration of Oracle related files by allowing the administrator to reference disk groups rather than individual disks and files, which are managed by ASM. The ASM functionality is an extention of the Oracle Managed Files (OMF) functionality that also includes striping and mirroring to provide balanced and secure storage. The new ASM functionality can be used in combination with existing raw and cooked file systems, along with OMF and manually managed files. The ASM functionality is controlled by an ASM instance. This is not a full database instance, just the memory structures and as such is very small and lightweight. The main components of ASM are disk groups, each of which comprise of several physical disks that are controlled as a single unit. The physical disks are known as ASM disks, while the files that reside on the disks are know as ASM files. The locations and names for the files are controlled by ASM, but user-friendly aliases and directory structures can be defined for ease of reference. The level of redundancy and the granularity of the striping can be controlled using templates. Default templates are provided for each file type stored by ASM, but additional templates can be defined as needed. Failure groups are defined within a disk group to support the required level of redundancy. For two-way mirroring you would expect a disk group to contain two failure groups so individual files are written to two locations. In summary ASM provides the following functionality: Manages groups of disks, called disk groups. Manages disk redundancy within a disk group. Provides near-optimal I/O balancing without any manual tuning. Enables management of database objects without specifying mount points and filenames. Supports large files. Initialization Parameters and ASM Instance Creation The init.ora / spfile initialization parameters that are of specific interest for an ASM instance are: INSTANCE_TYPE - Set to ASM or RDBMS depending on the instance type. The default is RDBMS. DB_UNIQUE_NAME - Specifies a globally unique name for the database. This defaults to +ASM but must be altered if you intend to run multiple ASM instances. ASM_POWER_LIMIT - The maximum power for a rebalancing operation on an ASM instance. The valid values range from 1 to 11, with 1 being the default. The higher the limit the more resources are allocated resulting in faster rebalancing operations. This value is also used as the default when the POWER clause is omitted from a rebalance operation. ASM_DISKGROUPS - The list of disk groups that should be mounted by an ASM instance during instance startup, or by the ALTER DISKGROUP ALL MOUNT statement. ASM configuration changes are automatically reflected in this parameter. ASM_DISKSTRING - Specifies a value that can be used to limit the disks considered for discovery. Altering the default value may improve the speed of disk group mount time and the speed of adding a disk to a disk group. Changing the parameter to a value which prevents the discovery of already mounted disks results in an error. The default value is NULL allowing all suitable disks to be considered. Incorrect usage of parameters in ASM or RDBMS instances result in ORA-15021 errors. To create an ASM instance first create a file called init+ASM.ora in the /tmp directory containing the following information. INSTANCE_TYPE=ASM Next, using SQL*Plus connect to the ide instance. export ORACLE_SID=+ASM sqlplus / as sysdba Create an spfile using the contents of the init+ASM.ora file. SQL> CREATE SPFILE FROM PFILE='/tmp/init+ASM.ora'; File created. Finally, start the instance with the NOMOUNT option. SQL> startup nomount ASM instance started Total System Global Area 125829120 bytes Fixed Size 1301456 bytes Variable Size 124527664 bytes Database Buffers 0 bytes Redo Buffers 0 bytes SQL> The ASM instance is now ready to use for creating and mounting disk groups. To shutdown the ASM instance issue the following command. SQL> shutdown ASM instance shutdown SQL> Once an ASM instance is present disk groups can be used for the following parameters in database instances (INSTANCE_TYPE=RDBMS) to allow ASM file creation: DB_CREATE_FILE_DEST DB_CREATE_ONLINE_LOG_DEST_n DB_RECOVERY_FILE_DEST CONTROL_FILES LOG_ARCHIVE_DEST_n LOG_ARCHIVE_DEST STANDBY_ARCHIVE_DEST Startup and Shutdown of ASM Instances ASM instance are started and stopped in a similar way to normal database instances. The options for the STARTUP command are: FORCE - Performs a SHUTDOWN ABORT before restarting the ASM instance. MOUNT - Starts the ASM instance and mounts the disk groups specified by the ASM_DISKGROUPS parameter. NOMOUNT - Starts the ASM instance without mounting any disk groups. OPEN - This is not a valid option for an ASM instance. The options for the SHUTDOWN command are: NORMAL - The ASM instance waits for all connected ASM instances and SQL sessions to exit then shuts down. IMMEDIATE - The ASM instance waits for any SQL transactions to complete then shuts down. It doesn't wait for sessions to exit. TRANSACTIONAL - Same as IMMEDIATE. ABORT - The ASM instance shuts down instantly. Aministering ASM Disk Groups Disk groups are created using the CREATE DISKGROUP statement. This statement allows you to specify the level of redundancy: NORMAL REDUNDANCY - Two-way mirroring, requiring two failure groups. HIGH REDUNDANCY - Three-way mirroring, requiring three failure groups. EXTERNAL REDUNDANCY - No mirroring for disks that are already protected using hardware mirroring or RAID. In addition failure groups and preferred names for disks can be defined. If the NAME clause is omitted the disks are given a system generated name like "disk_group_1_0001". The FORCE option can be used to move a disk from another disk group into this one. CREATE DISKGROUP disk_group_1 NORMAL REDUNDANCY FAILGROUP failure_group_1 DISK '/devices/diska1' NAME diska1, '/devices/diska2' NAME diska2, FAILGROUP failure_group_2 DISK '/devices/diskb1' NAME diskb1, '/devices/diskb2' NAME diskb2; Disk groups can be deleted using the DROP DISKGROUP statement. DROP DISKGROUP disk_group_1 INCLUDING CONTENTS; Disks can be added or removed from disk groups using the ALTER DISKGROUP statement. Remember that the wildcard "*" can be used to reference disks so long as the resulting string does not match a disk already used by an existing disk group. -- Add disks. ALTER DISKGROUP disk_group_1 ADD DISK '/devices/disk*3', '/devices/disk*4'; -- Drop a disk. ALTER DISKGROUP disk_group_1 DROP DISK diska2; Disks can be resized using the RESIZE clause of the ALTER DISKGROUP statement. The statement can be used to resize individual disks, all disks in a failure group or all disks in the disk group. If the SIZE clause is omitted the disks are resized to the size of the disk returned by the OS. -- Resize a specific disk. ALTER DISKGROUP disk_group_1 RESIZE DISK diska1 SIZE 100G; -- Resize all disks in a failure group. ALTER DISKGROUP disk_group_1 RESIZE DISKS IN FAILGROUP failure_group_1 SIZE 100G; -- Resize all disks in a disk group. ALTER DISKGROUP disk_group_1 RESIZE ALL SIZE 100G;The UNDROP DISKS clause of the ALTER DISKGROUP statement allows pending disk drops to be undone. It will not revert drops that have completed, or disk drops associated with the dropping of a disk group. ALTER DISKGROUP disk_group_1 UNDROP DISKS; Disk groups can be rebalanced manually using the REBALANCE clause of the ALTER DISKGROUP statement. If the POWER clause is omitted the ASM_POWER_LIMIT parameter value is used. Rebalancing is only needed when the speed of the automatic rebalancing is not appropriate. ALTER DISKGROUP disk_group_1 REBALANCE POWER 5; Disk groups are mounted at ASM instance startup and unmounted at ASM instance shutdown. Manual mounting and dismounting can be accomplished using the ALTER DISKGROUP statement as seen below. ALTER DISKGROUP ALL DISMOUNT; ALTER DISKGROUP ALL MOUNT; ALTER DISKGROUP disk_group_1 DISMOUNT; ALTER DISKGROUP disk_group_1 MOUNT; Templates Templates are named groups of attributes that can be applied to the files within a disk group. The following example show how templates can be created, altered and dropped. -- Create a new template. ALTER DISKGROUP disk_group_1 ADD TEMPLATE my_template ATTRIBUTES (MIRROR FINE); -- Modify template. ALTER DISKGROUP disk_group_1 ALTER TEMPLATE my_template ATTRIBUTES (COARSE); -- Drop template. ALTER DISKGROUP disk_group_1 DROP TEMPLATE my_template;Available attributes include: UNPROTECTED - No mirroring or striping regardless of the redundancy setting. MIRROR - Two-way mirroring for normal redundancy and three-way mirroring for high redundancy. This attribute cannot be set for external redundancy. COARSE - Specifies lower granuality for striping. This attribute cannot be set for external redundancy. FINE - Specifies higher granularity for striping. This attribute cannot be set for external redundancy. Directories A directory heirarchy can be defined using the ALTER DISKGROUP statement to support ASM file aliasing. The following examples show how ASM directories can be created, modified and deleted. -- Create a directory. ALTER DISKGROUP disk_group_1 ADD DIRECTORY '+disk_group_1/my_dir'; -- Rename a directory. ALTER DISKGROUP disk_group_1 RENAME DIRECTORY '+disk_group_1/my_dir' TO '+disk_group_1/my_dir_2'; -- Delete a directory and all its contents. ALTER DISKGROUP disk_group_1 DROP DIRECTORY '+disk_group_1/my_dir_2' FORCE;Aliases Aliases allow you to reference ASM files using user-friendly names, rather than the fully qualified ASM filenames. -- Create an alias using the fully qualified filename. ALTER DISKGROUP disk_group_1 ADD ALIAS '+disk_group_1/my_dir/my_file.dbf' FOR '+disk_group_1/mydb/datafile/my_ts.342.3'; -- Create an alias using the numeric form filename. ALTER DISKGROUP disk_group_1 ADD ALIAS '+disk_group_1/my_dir/my_file.dbf' FOR '+disk_group_1.342.3'; -- Rename an alias. ALTER DISKGROUP disk_group_1 RENAME ALIAS '+disk_group_1/my_dir/my_file.dbf' TO '+disk_group_1/my_dir/my_file2.dbf'; -- Delete an alias. ALTER DISKGROUP disk_group_1 DELETE ALIAS '+disk_group_1/my_dir/my_file.dbf'; Attempting to drop a system alias results in an error. Files Files are not deleted automatically if they are created using aliases, as they are not Oracle Managed Files (OMF), or if a recovery is done to a point-in-time before the file was created. For these circumstances it is necessary to manually delete the files, as shown below. -- Drop file using an alias. ALTER DISKGROUP disk_group_1 DROP FILE '+disk_group_1/my_dir/my_file.dbf'; -- Drop file using a numeric form filename. ALTER DISKGROUP disk_group_1 DROP FILE '+disk_group_1.342.3'; -- Drop file using a fully qualified filename. ALTER DISKGROUP disk_group_1 DROP FILE '+disk_group_1/mydb/datafile/my_ts.342.3'; Checking Metadata The internal consistency of disk group metadata can be checked in a number of ways using the CHECK clause of the ALTER DISKGROUP statement. -- Check metadata for a specific file. ALTER DISKGROUP disk_group_1 CHECK FILE '+disk_group_1/my_dir/my_file.dbf' -- Check metadata for a specific failure group in the disk group. ALTER DISKGROUP disk_group_1 CHECK FAILGROUP failure_group_1; -- Check metadata for a specific disk in the disk group. ALTER DISKGROUP disk_group_1 CHECK DISK diska1; -- Check metadata for all disks in the disk group. ALTER DISKGROUP disk_group_1 CHECK ALL; ASM Views The ASM configuration can be viewed using the V$ASM_% views, which often contain different information depending on whether they are queried from the ASM instance, or a dependant database instance. Viewing ASM Instance Information Via SQL Queries Finally, there are several dynamic and data dictionary views available to view an ASM configuration from within the ASM instance itself: -- ASM Dynamic Views: FROM ASM Instance Information View Name Description V$ASM_ALIAS Shows every alias for every disk group mounted by the ASM instance V$ASM_CLIENT Shows which database instance(s) are using any ASM disk groups that are being mounted by this ASM instance V$ASM_DISK Lists each disk discovered by the ASM instance, including disks that are not part of any ASM disk group V$ASM_DISKGROUP Describes information about ASM disk groups mounted by the ASM instance V$ASM_FILE Lists each ASM file in every ASM disk group mounted by the ASM instance V$ASM_OPERATION Like its counterpart, V$SESSION_LONGOPS, it shows each long-running ASM operation in the ASM instance V$ASM_TEMPLATE Lists each template present in every ASM disk group mounted by the ASM instance I was also able to query the following dynamic views against my database instance to view the related ASM storage components of that instance: -- ASM Dynamic Views: FROM Database Instance Information View Name Description V$ASM_DISKGROUP Shows one row per each ASM disk group that's mounted by the local ASM instance V$ASM_DISK Displays one row per each disk in each ASM disk group that are in use by the database instance V$ASM_CLIENT Lists one row per each ASM instance for which the database instance has any open ASM files ASM Filenames There are several ways to reference ASM file. Some forms are used during creation and some for referencing ASM files. The forms for file creation are incomplete, relying on ASM to create the fully qualified name, which can be retrieved from the supporting views. The forms of the ASM filenames are summarised below. Filename Type Format Fully Qualified ASM Filename +dgroup/dbname/file_type/file_type_tag.file.incarnation Numeric ASM Filename +dgroup.file.incarnation Alias ASM Filenames +dgroup/directory/filename Alias ASM Filename with Template +dgroup(template)/alias Incomplete ASM Filename +dgroup Incomplete ASM Filename with Template +dgroup(template) SQL and ASM ASM filenames can be used in place of conventional filenames for most Oracle file types, including controlfiles, datafiles, logfiles etc. For example, the following command creates a new tablespace with a datafile in the disk_group_1 disk group. CREATE TABLESPACE my_ts DATAFILE '+disk_group_1' SIZE 100M AUTOEXTEND ON;Migrating to ASM Using RMAN The following method shows how a primary database can be migrated to ASM from a disk based backup: Disable change tracking (only available in Enterprise Edition) if it is currently being used. SQL> ALTER DATABASE DISABLE BLOCK CHANGE TRACKING;Shutdown the database. SQL> SHUTDOWN IMMEDIATEModify the parameter file of the target database as follows: Set the DB_CREATE_FILE_DEST and DB_CREATE_ONLINE_LOG_DEST_n parameters to the relevant ASM disk groups. Remove the CONTROL_FILES parameter from the spfile so the control files will be moved to the DB_CREATE_* destination and the spfile gets updated automatically. If you are using a pfile the CONTROL_FILES parameter must be set to the appropriate ASM files or aliases. Start the database in nomount mode. RMAN> STARTUP NOMOUNTRestore the controlfile into the new location from the old location. RMAN> RESTORE CONTROLFILE FROM 'old_control_file_name';Mount the database. RMAN> ALTER DATABASE MOUNT;Copy the database into the ASM disk group. RMAN> BACKUP AS COPY DATABASE FORMAT '+disk_group';Switch all datafile to the new ASM location. RMAN> SWITCH DATABASE TO COPY;Open the database. RMAN> ALTER DATABASE OPEN;Create new redo logs in ASM and delete the old ones. Enable change tracking if it was being used. SQL> ALTER DATABASE ENABLE BLOCK CHANGE TRACKING;Form more information see: Using Automatic Storage Management Migrating a Database into ASM Hope this helps. Regards Tim... Note 6: ======= Good example !!!! How to Use Oracle10g release 2 ASM on Linux: [root@danaly etc]# fdisk /dev/cciss/c0d0 The number of cylinders for this disk is set to 8854. There is nothing wrong with that, but this is larger than 1024, and could in certain setups cause problems with: 1) software that runs at boot time (e.g., old versions of LILO) 2) booting and partitioning software from other OSs (e.g., DOS FDISK, OS/2 FDISK) Command (m for help): p Disk /dev/cciss/c0d0: 72.8 GB, 72833679360 bytes 255 heads, 63 sectors/track, 8854 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Device Boot Start End Blocks Id System /dev/cciss/c0d0p1 * 1 33 265041 83 Linux /dev/cciss/c0d0p2 34 555 4192965 82 Linux swap /dev/cciss/c0d0p3 556 686 1052257+ 83 Linux /dev/cciss/c0d0p4 687 8854 65609460 5 Extended /dev/cciss/c0d0p5 687 1730 8385898+ 83 Linux /dev/cciss/c0d0p6 1731 2774 8385898+ 83 Linux /dev/cciss/c0d0p7 2775 3818 8385898+ 83 Linux /dev/cciss/c0d0p8 3819 4601 6289416 83 Linux Command (m for help): n First cylinder (4602-8854, default 4602): Using default value 4602 Last cylinder or +size or +sizeM or +sizeK (4602-8854, default 8854): +20000M Command (m for help): n First cylinder (7035-8854, default 7035): Using default value 7035 Last cylinder or +size or +sizeM or +sizeK (7035-8854, default 8854): +3000M Command (m for help): n First cylinder (7401-8854, default 7401): Using default value 7401 Last cylinder or +size or +sizeM or +sizeK (7401-8854, default 8854): +3000M Command (m for help): p Disk /dev/cciss/c0d0: 72.8 GB, 72833679360 bytes 255 heads, 63 sectors/track, 8854 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Device Boot Start End Blocks Id System /dev/cciss/c0d0p1 * 1 33 265041 83 Linux /dev/cciss/c0d0p2 34 555 4192965 82 Linux swap /dev/cciss/c0d0p3 556 686 1052257+ 83 Linux /dev/cciss/c0d0p4 687 8854 65609460 5 Extended /dev/cciss/c0d0p5 687 1730 8385898+ 83 Linux /dev/cciss/c0d0p6 1731 2774 8385898+ 83 Linux /dev/cciss/c0d0p7 2775 3818 8385898+ 83 Linux /dev/cciss/c0d0p8 3819 4601 6289416 83 Linux /dev/cciss/c0d0p9 4602 7034 19543041 83 Linux /dev/cciss/c0d0p10 7035 7400 2939863+ 83 Linux /dev/cciss/c0d0p11 7401 7766 2939863+ 83 Linux Command (m for help): w The partition table has been altered! Calling ioctl() to re-read partition table. WARNING: Re-reading the partition table failed with error 16: Device or resource busy. The kernel still uses the old table. The new table will be used at the next reboot. Syncing disks. [root@danaly data1]# /etc/init.d/oracleasm createdisk VOL5 /dev/cciss/c0d0p10 Marking disk "/dev/cciss/c0d0p10" as an ASM disk: [ OK ] [root@danaly data1]# /etc/init.d/oracleasm createdisk VOL6 /dev/cciss/c0d0p11 Marking disk "/dev/cciss/c0d0p11" as an ASM disk: [ OK ] [root@danaly data1]# /etc/init.d/oracleasm listdisks VOL1 VOL2 VOL3 VOL4 VOL5 VOL6 (THE FOLLOWING QUERIES ARE ISSUED FROM THE ASM INSTANCE.) [oracle@danaly ~]$ export ORACLE_SID=+ASM [oracle@danaly ~]$ sqlplus "/ as sysdba" SQL*Plus: Release 10.2.0.1.0 - Production on Sun Sep 3 00:28:09 2006 Copyright (c) 1982, 2005, Oracle. All rights reserved. Connected to an idle instance. SQL> startup ASM instance started Total System Global Area 83886080 bytes Fixed Size 1217836 bytes Variable Size 57502420 bytes ASM Cache 25165824 bytes ASM diskgroups mounted SQL> select group_number,disk_number,mode_status from v$asm_disk; GROUP_NUMBER DISK_NUMBER MODE_STATUS ------------ ----------- -------------- 0 4 ONLINE 0 5 ONLINE 1 0 ONLINE 1 1 ONLINE 1 2 ONLINE 1 3 ONLINE 6 rows selected. SQL> select group_number,disk_number,mode_status,name from v$asm_disk; GROUP_NUMBER DISK_NUMBER MODE_STATUS NAME ------------ ----------- -------------- --------------------------------- 0 4 ONLINE 0 5 ONLINE 1 0 ONLINE VOL1 1 1 ONLINE VOL2 1 2 ONLINE VOL3 1 3 ONLINE VOL4 6 rows selected. SQL> create diskgroup orag2 external redundancy disk 'ORCL:VOL5'; Diskgroup created. SQL> select group_number,disk_number,mode_status,name from v$asm_disk; GROUP_NUMBER DISK_NUMBER MODE_STATUS NAME ------------ ----------- -------------- ------------------------------------- 0 5 ONLINE 1 0 ONLINE VOL1 1 1 ONLINE VOL2 1 2 ONLINE VOL3 1 3 ONLINE VOL4 2 0 ONLINE VOL5 6 rows selected. (THE FOLLOWING QUERIES ARE ISSUED FROM THE DATABASE INSTANCE.) [oracle@danaly ~]$ export ORACLE_SID=danaly [oracle@danaly ~]$ sqlplus "/ as sysdba" SQL*Plus: Release 10.2.0.1.0 - Production on Sun Sep 3 00:47:04 2006 Copyright (c) 1982, 2005, Oracle. All rights reserved. Connected to an idle instance. SQL> startup ORACLE instance started. Total System Global Area 943718400 bytes Fixed Size 1222744 bytes Variable Size 281020328 bytes Database Buffers 654311424 bytes Redo Buffers 7163904 bytes Database mounted. Database opened. SQL> select name from v$datafile; NAME -------------------------------------------------------------------------------- +ORADG/danaly/datafile/system.264.600016955 +ORADG/danaly/datafile/undotbs1.265.600016969 +ORADG/danaly/datafile/sysaux.266.600016977 +ORADG/danaly/datafile/users.268.600016987 SQL> create tablespace eygle datafile '+ORAG2' ; Tablespace created. SQL> select name from v$datafile; NAME --------------------------------------------------------------------------------- +ORADG/danaly/datafile/system.264.600016955 +ORADG/danaly/datafile/undotbs1.265.600016969 +ORADG/danaly/datafile/sysaux.266.600016977 +ORADG/danaly/datafile/users.268.600016987 +ORAG2/danaly/datafile/eygle.256.600137647 oracle@danaly log]$ export ORACLE_SID=+ASM [oracle@danaly log]$ sqlplus "/ as sysdba" SQL*Plus: Release 10.2.0.1.0 - Production on Sun Sep 3 01:36:37 2006 Copyright (c) 1982, 2005, Oracle. All rights reserved. Connected to: Oracle Database 10g Enterprise Edition Release 10.2.0.1.0 - Production With the Partitioning, Oracle Label Security, OLAP and Data Mining Scoring Engine options SQL> alter diskgroup orag2 add disk 'ORCL:VOL6'; Diskgroup altered. ============ Note 7: OMF ============ Using Oracle-managed files simplifies the administration of an Oracle database. Oracle-managed files eliminate the need for you, the DBA, to directly manage the operating system files comprising an Oracle database. You specify operations in terms of database objects rather than filenames. Oracle internally uses standard file system interfaces to create and delete files as needed for the following database structures: Tablespaces Online redo log files Control files The following initialization parameters init.ora/spfile.ora allow the database server to use the Oracle Managed Files feature: - DB_CREATE_FILE_DEST Defines the location of the default file system directory where Oracle creates datafiles or tempfiles when no file specification is given in the creation operation. Also used as the default file system directory for online redo log and control files if DB_CREATE_ONLINE_LOG_DEST_n is not specified. - DB_CREATE_ONLINE_LOG_DEST_n Defines the location of the default file system directory for online redo log files and control file creation when no file specification is given in the creation operation. You can use this initialization parameter multiple times, where n specifies a multiplexed copy of the online redo log or control file. You can specify up to five multiplexed copies Example: DB_CREATE_FILE_DEST = '/u01/oradata/payroll' DB_CREATE_ONLINE_LOG_DEST_1 = '/u02/oradata/payroll' DB_CREATE_ONLINE_LOG_DEST_2 = '/u03/oradata/payroll' 34.2 RAC 10g: ============= =========================================== Note 1: High Level Overview Oracle 10g RAC =========================================== - RAC Architecture Overview Let's begin with a brief overview of RAC architecture. A cluster is a set of 2 or more machines (nodes) that share or coordinate resources to perform the same task. A RAC database is 2 or more instances running on a set of clustered nodes, with all instances accessing a shared set of database files. Depending on the O/S platform, a RAC database may be deployed on a cluster that uses vendor clusterware plus Oracle's own clusterware (Cluster Ready Services), or on a cluster that solely uses Oracle's own clusterware. Thus, every RAC sits on a cluster that is running Cluster Ready Services. srvctl is the primary tool DBAs use to configure CRS for their RAC database and processes. - Cluster Ready Services and the OCR Cluster Ready Services, or CRS, is a new feature for 10g RAC. Essentially, it is Oracle's own clusterware. On most platforms, Oracle supports vendor clusterware; in these cases, CRS interoperates with the vendor clusterware, providing high availability support and service and workload management. On Linux and Windows clusters, CRS serves as the sole clusterware. In all cases, CRS provides a standard cluster interface that is consistent across all platforms. CRS consists of four processes (crsd, occsd, evmd, and evmlogger) and two disks: the Oracle Cluster Registry (OCR), and the voting disk. CRS manages the following resources: . The ASM instances on each node . Databases . The instances on each node . Oracle Services on each node . The cluster nodes themselves, including the following processes, or "nodeapps": . VIP . GSD . The listener . The ONS daemon CRS stores information about these resources in the OCR. If the information in the OCR for one of these resources becomes damaged or inconsistent, then CRS is no longer able to manage that resource. Fortunately, the OCR automatically backs itself up regularly and frequently. 10g RAC (10.2) uses, or depends on,: - Oracle Clusterware (10.2), formerly referred to as CRS "Cluster Ready Services" (10.1). - Oracle's optional Cluster File System OCFS (This is optional), or use ASM and RAW. - Oracle Database extensions RAC is "scale out" technology: just add commodity nodes to the system. The key component is "cache fusion". Data are transferred from one node to another via very fast interconnects. Essential to 10g RAC is a "Shared Cache" technology. Automatic Workload Repository (AWR) plays a role also. The Fast Application Notification (FAN) mechanism that is part of RAC, publishes events that describe the current service level being provided by each instance, to AWR. The load balancing advisory information is then used to determine the best instance to serve the new request. . With RAC, ALL Instances of ALL nodes in a cluster, access a SINGLE database. . But every instance has it's own UNDO tablespace, and REDO logs. The Oracle Clusterware comprise several background processes that facilitate cluster operations. The Cluster Synchronization Service CSS, Event Management EVM, and Oracle Cluster components communicate with other cluster components layers in the other instances within the same cluster database environment. Questions per implementation arise in the following points: . Storage . Computer Systems/Storage-Interconnect . Database . Application Server . Public and Private networks . Application Control & Display On the Storage level, it can be said that 10g RAC supports - Automatic Storage Management (ASM) - Oracle Cluster File System (OCFS) - ??? Network File System (NFS) - limited (only theoretical actually) - Disk raw partitions - Third party cluster file systems For application control and tools, it can be said that 10g RAC supports - OEM Grid Control http://hostname:5500/em OEM Database Control http://hostname:1158/em - "svrctl" is a command line interface to manage the cluster configuration, for example, starting and stopping all nodes in one command. - Cluster Verification Utility (cluvfy) can be used for an installation and sanity check. Failure in Client connections: Depending on the Net configuration, type of connection, type of transaction etc.., Oracle Net services provides a feature called "Transparant Application Failover", or TAF, which can fail over a client session to another backup connection. About HA and DR: - RAC is HA , High Availability, that will keep things Up and Running in one site. - Data Guard is DR, Disaster Recovery, and is able to mirror one site to another remote site. =========================================================== Note 2: 10g RAC processes, services, daemons and start stop =========================================================== CRS consists of four processes (crsd, occsd, evmd, and evmlogger) and two disks: the Oracle Cluster Registry (OCR), and the voting disk. On most platforms, you may see the following processes: oprocd the Process Monitor Daemon crsd the CRS Daemon occsd Oracle Cluster Synchronization Service Daemon evmd Event Volume Manager Daemon To start and stop CRS when the machine starts or shutdown, on unix there are rc scripts in place. You can also, as root, manually start, stop, enable or disable the services with: /etc/init.d/init.crs start /etc/init.d/init.crs stop /etc/init.d/init.crs enable /etc/init.d/init.crs disable Or with # crsctl start crs # crsctl stop crs # crsctl enable crs # crsctl disable crs ============================================== Note 3: Installation notes 10g RAC on Windows ============================================== See the next note for installation on Linux 3.1 Before you install: ----------------------- Each node in a cluster requires the following: > One private internet protocol (IP) address for each node to serve as the private interconnect. The following must be true for each private IP address: -It must be separate from the public network -It must be accessible on the same network interface on each node -It must have a unique address on each node The private interconnect is used for inter-node communication by both Oracle Clusterware and RAC. If the private address is available from a network name server (DNS), then you can use that name. Otherwise, the private IP address must be available in each node's C:\WINNT\system32\drivers\etc\hosts file. > One public IP address for each node, to be used as the Virtual IP (VIP) address for client connections and for connection failover. The name associated with the VIP must be different from the default host name. This VIP must be associated with the same interface name on every node that is part of your cluster. In addition, the IP addresses that you use for all of the nodes that are part of a cluster must be from the same subnet. > One public fixed hostname address for each node, typically assigned by the system administrator during operating system installation. If you have a DNS, then register both the fixed IP and the VIP address with DNS. If you do not have DNS, then you must make sure that the public IP and VIP addresses for all nodes are in each node's host file. For example, with a two node cluster where each node has one public and one private interface, you might have the configuration shown in the following table for your network interfaces, where the hosts file is %SystemRoot%\system32\drivers\etc\hosts: Node Interface Name Type IP Address Registered In rac1 rac1 Public 143.46.43.100 DNS (if available, else the hosts file) rac1 rac1-vip Virtual 143.46.43.104 DNS (if available, else the hosts file) rac1 rac1-priv Private 10.0.0.1 Hosts file rac2 rac2 Public 143.46.43.101 DNS (if available, else the hosts file) rac2 rac2-vip Virtual 143.46.43.105 DNS (if available, else the hosts file) rac2 rac2-priv Private 10.0.0.2 Hosts file The virtual IP addresses are assigned to the listener process. To enable VIP failover, the configuration shown in the preceding table defines the public and VIP addresses of both nodes on the same subnet, 143.46.43. When a node or interconnect fails, then the associated VIP is relocated to the surviving instance, enabling fast notification of the failure to the clients connecting through that VIP. If the application and client are configured with transparent application failover options, then the client is reconnected to the surviving instance. To disable Windows Media Sensing for TCP/IP, you must set the value of the DisableDHCPMediaSense parameter to 1 on each node. Disable Media Sensing by completing the following steps on each node of your cluster: Use Registry Editor (Regedt32.exe) to view the following key in the registry: HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Tcpip\Parameters Add the following registry value: Value Name: DisableDHCPMediaSense Data Type: REG_DWORD -Boolean Value: 1 - External shared disks for storing Oracle Clusterware and database files. The disk configuration options available to you are described in Chapter 3, "Storage Pre-Installation Tasks". Review these options before you decide which storage option to use in your RAC environment. However, note that when Database Configuration Assistant (DBCA) configures automatic disk backup, it uses a database recovery area which must be shared. The database files and recovery files do not necessarily have to be located on the same type of storage. Determine the storage option for your system and configure the shared disk. Oracle recommends that you use Automatic Storage Management (ASM) and Oracle Managed Files (OMF), or a cluster file system. If you use ASM or a cluster file system, then you can also take advantage of OMF and other Oracle Database 10g storage features. If you use RAC on Oracle Database 10g Standard Edition, then you must use ASM. If you use ASM, then Oracle recommends that you install ASM in a separate home from the Oracle Clusterware home and the Oracle home. Oracle Database 10g Real Application Clusters installation is a two-phase installation. In phase one, use Oracle Universal Installer (OUI) to install Oracle Clusterware. In phase two, install the database software using OUI. When you install Oracle Clusterware or RAC, OUI copies the Oracle software onto the node from which you are running it. If your Oracle home is not on a cluster file system, then OUI propagates the software onto the other nodes that you have selected to be part of your OUI installation session. - Shared Storage for Database Recovery Area When you configure a database recovery area in a RAC environment, the database recovery area must be on shared storage. When Database Configuration Assistant (DBCA) configures automatic disk backup, it uses a database recovery area that must be shared. If the database files are stored on a cluster file system, then the recovery area can also be shared through the cluster file system. If the database files are stored on an Automatic Storage Management (ASM) disk group, then the recovery area can also be shared through ASM. If the database files are stored on raw devices, then you must use either a cluster file system or ASM for the recovery area. Note: ASM disk groups are always valid recovery areas, as are cluster file systems. Recovery area files do not have to be in the same location where datafiles are stored. For instance, you can store datafiles on raw devices, but use ASM for the recovery area. Data files are not placed on NTFS partitions, because they cannot be shared. Data files can be placed on Oracle Cluster File System (OCFS), on raw disks using ASM, or on raw disks. - Oracle Clusterware You must provide OUI with the names of the nodes on which you want to install Oracle Clusterware. The Oracle Clusterware home can be either shared by all nodes, or private to each node, depending on your responses when you run OUI. The home that you select for Oracle Clusterware must be different from the RAC-enabled Oracle home. Versions of cluster manager previous to Oracle Database 10g were sometimes referred to as "Cluster Manager". In Oracle Database 10g, this function is performed by a Oracle Clusterware component known as Cluster Synchronization Services (CSS). The OracleCSService, OracleCRService, and OracleEVMService replace the service known previous to Oracle Database 10g as OracleCMService9i. 3.2 cluvfy or runcluvfy.bat: ---------------------------- Once you have installed Oracle Clusterware, you can use CVU by entering cluvfy commands on the command line. To use CVU before you install Oracle Clusterware, you must run the commands using a command file available on the Oracle Clusterware installation media. Use the following syntax to run a CVU command run from the installation media, where media is the location of the Oracle Clusterware installation media and options is a list of one or more CVU command options: media\clusterware\cluvfy\runcluvfy.bat options The following code example is of a CVU help command, run from a staged copy of the Oracle Clusterware directory downloaded from OTN into a directory called stage on your C: drive: C:\stage\clusterware\cluvfy> runcluvfy.bat comp nodereach -n node1,node2 -verbose For a quick test, you can run the following CVU command that you would normally use after you have completed the basic hardware and software configuration: prompt> media\clusterware\cluvfy\runcluvfy.bat stage –post hwos –n node_list Use the location of your Oracle Clusterware installation media for the media value and a list of the nodes, separated by commas, in your cluster for node_list. Expect to see many errors if you run this command before you or your system administrator complete the cluster pre-installation steps. On Oracle Real Application Clusters systems, each member node of the cluster must have user equivalency for the Administrative privileges account that installs the database. This means that the administrative privileges user account and password must be the same on all nodes. - Checking the Hardware and Operating System Setup with CVU You can use two different CVU commands to check your hardware and operating system configuration. The first is a general check of the configuration, and the second specifically checks for the components required to install Oracle Clusterware. The syntax of the more general CVU command is: cluvfy stage –post hwos –n node_list [-verbose] where node_list is the names of the nodes in your cluster, separated by commas. However, because you have not yet installed Oracle Clusterware, you must execute the CVU command from the installation media using a command like the one following. In this example, the command checks the hardware and operating system of a two-node cluster with nodes named node1 and node2, using a staged copy of the installation media in a directory called stage on the C: drive: C:\stage\clusterware\cluvfy> runcluvfy.bat stage –post hwos –n node1,node2 -verbose You can omit the -verbose keyword if you do not wish to see detailed results listed as CVU performs each individual test. The following example is a command, without the -verbose keyword, to check for the readiness of the cluster for installing Oracle Clusterware: C:\stage\clusterware\cluvfy> runcluvfy.bat comp sys -n node1,node2 -p crs - Checking the Network Setup Enter a command using the following syntax to verify node connectivity between all of the nodes for which your cluster is configured: cluvfy comp nodecon -n node_list [-verbose] - Verifying Cluster Privileges Before running Oracle Universal Installer, from the node where you intend to run the Installer, verify that you have administrative privileges on the other nodes. To do this, enter the following command for each node that is a part of the cluster: net use \\node_name\C$ where node_name is the node name. If your installation will access drives in addition to the C: drive, repeat this command for every node in the cluster, substituting the drive letter for each drive you plan to use. For the installation to be successful, you must use the same user name and password on each node in a cluster or use a domain user name. If you use a domain user name, then log on under a domain with a user name and password to which you have explicitly granted local administrative privileges on all nodes. 3.3 Shared disk considerations: ------------------------------- Preliminary Shared Disk Preparation Complete the following steps to prepare shared disks for storage: -- Disabling Write Caching You must disable write caching on all disks that will be used to share data between nodes in your cluster. To disable write caching, perform these steps: Click Start, then click Settings, then Control Panel, then Administrative Tools, then Computer Management, then Device Manager, and then Disk drives Expand the Disk drives and double-click the first drive listed Under the Disk Properties tab for the selected drive, uncheck the option that enables the write cache Double-click each of the other drives listed in the Disk drives hive and disable the write cache as described in the previous step Caution: Any disks that you use to store files, including database files, that will be shared between nodes, must have write caching disabled. -- Enabling Automounting for Windows 2003 If you are using Windows 2003, then you must enable disk automounting, depending on the Oracle products you are installing and on other conditions. You must enable automounting when using: Raw partitions for Oracle Real Application Clusters (RAC) Cluster file system for Oracle Real Application Clusters Oracle Clusterware Raw partitions for a single-node database installation Logical drives for Automatic Storage Management (ASM) To enable automounting: Enter the following commands at a command prompt: c:\> diskpart DISKPART> automount enable Automatic mounting of new volumes enabled. Type exit to end the diskpart session Repeat steps 1 and 2 for each node in the cluster. 3.4 Reviewing Storage Options for Oracle Clusterware, Database, and Recovery Files: ----------------------------------------------------------------------------------- This section describes supported options for storing Oracle Clusterware files, Oracle Database software, and database files. -- Overview of Oracle Clusterware Storage Options Note that Oracle Clusterware files include the Oracle Cluster Registry (OCR) and the Oracle Clusterware voting disk. There are two ways to store Oracle Clusterware files: 1. Oracle Cluster File System (OCFS): The cluster file system Oracle provides for the Windows and Linux communities. If you intend to store Oracle Clusterware files on OCFS, then you must ensure that OCFS volume sizes are at least 500 MB each. 2. Raw storage: Raw logical volumes or raw partitions are created and managed by Microsoft Windows disk management tools or by tools provided by third party vendors. Note that you must provide disk space for one mirrored Oracle Cluster Registry (OCR) file, and two mirrored voting disk files. -- Overview of Oracle Database and Recovery File Options There are three ways to store Oracle Database and recovery files on shared disks: 1. Automatic Storage Management (database files only): Automatic Storage Management (ASM) is an integrated, high-performance database file system and disk manager for Oracle files. Because ASM requires an Oracle Database instance, it cannot contain Oracle software, but you can use ASM to manage database and recovery files. 2. Oracle Cluster File System (OCFS): Note that if you intend to use OCFS for your database files, then you should create partitions large enough for the database files when you create partitions for Oracle Clusterware Note: If you want to have a shared Oracle home directory for all nodes, then you must use OCFS. 3. Raw storage: Note that you cannot use raw storage to store Oracle database recovery files. The storage option that you choose for recovery files can be the same as or different to the option you choose for the database files. Storage Option Oracle Clusterware Database Recovery area -------------- ------------------ -------- ------------- Automatic Storage Management No Yes Yes Cluster file system (OCFS) Yes Yes Yes Shared raw storage Yes Yes No -- Checking for Available Shared Storage with CVU To check for all shared file systems available across all nodes on the cluster, use the following CVU command: cluvfy comp ssa -n node_list Remember to use the full path name and the runcluvfy.bat command on the installation media and include the list of nodes in your cluster, separated by commas, for the node_list. The following example is for a system with two nodes, node1 and node2, and the installation media on drive F: F:\clusterware\cluvfy> runcluvfy.bat comp ssa -n node1,node2 If you want to check the shared accessibility of a specific shared storage type to specific nodes in your cluster, then use the following command syntax: cluvfy comp ssa -n node_list -s storageID_list In the preceding syntax, the variable node_list is the list of nodes you want to check, separated by commas, and the variable storageID_list is the list of storage device IDs for the storage devices managed by the file system type that you want to check. ===================================== Note 4: Installation on Redhat Linux ===================================== 4.2 Prepare your nodes: ----------------------- 4.2.1 Scetch of a 2-node Linux cluster 192.168.2.0 ------------------------------------------ public network | | | | ------------ ------------- |InstanceA |Private network |InstanceB | | |Ethernet | | | |--------------------| | | |192.168.1.0 | | | | | | | |____________ | | | | ----- -|--- | | | |--|PWR| |PWR|----| | | | ----- ----- | | | | |_______________| | | | | | ------------ ------------- | SCSI bus or Fible Channel | ------------------ -------------- Interconnect | | | | Fig 4.1 ----------- |Shared | - has Single DB on: ASM or OCFS or RAW |Disk | - has OCR and Voting disk on: OCFS or RAW (not ASM) |Storage | - has Recovery area on: ASM or OCFS (not RAW) ----------- 4.2.2 Storage Options Storage Oracle Clusterware Database Recovery area -------------- ------------------ -------- ------------- Automatic Storage Management No Yes Yes Cluster file system (OCFS) Yes Yes Yes Shared raw storage Yes Yes No In the following, we will do an example installation on 3 nodes. 4.2.3 Install Redhat on all nodes with all options. 4.2.4 create oracle user and groups dba, oinstall on all nodes. Make sure they all have the same UID and GUI. 4.2.5 Make sure the user oracle has an appropriate .profile or .bash_profile 4.2.6 Every node needs a private network connection and a public network connection (at least two networkcards). 4.2.7 Linux kernel parameters: Most out of the box kernel parameters (of RHELS 3,4,5) are set correctly for Oracle except a few. You should have the following minimal configuration: net.ipv4.ip_local_port_range 1024 65000 kernel.sem 250 32000 100 128 kernel.shmmni 4096 kernel.shmall 2097152 kernel.shmmax 2147483648 fs.file-max 65536 You can check the most important parameters using the following command: # /sbin/sysctl -a | egrep 'sem|shm|file-max|ip_local' net.ipv4.ip_local_port_range = 1024 65000 kernel.sem = 250 32000 100 128 kernel.shmmni = 4096 kernel.shmall = 2097152 kernel.shmmax = 2147483648 fs.file-max = 65536 If some value should be changed, you can change the "/etc/sysctl.conf" file and run the "/sbin/sysctl -p" command to change the value immediately. Every time the system boots, the init program runs the /etc/rc.d/rc.sysinit script. This script contains a command to execute sysctl using /etc/sysctl.conf to dictate the values passed to the kernel. Any values added to /etc/sysctl.conf will take effect each time the system boots. 4.2.8 make sure ssh and scp are working on all nodes without asking for a password. Use shh-keygen to arrange that. 4.2.9 Example "/etc/host" on the nodes: Suppose you have the following 3 hosts, with their associated public and private names: public private oc1 poc1 oc2 poc2 oc3 poc3 Then this could be a valid host file on the nodes: 127.0.0.1 localhost.localdomain localhost 192.168.2.99 rhes30 192.168.2.166 oltp 192.168.2.167 mw 192.168.2.101 oc1 #public1 192.168.1.101 poc1 #private1 192.168.2.176 voc1 #virtual1 192.168.2.102 oc2 #public2 192.168.1.102 poc2 #private2 192.168.2.177 voc2 #virtual2 192.168.2.103 oc3 #public3 192.168.1.103 poc3 #private3 192.168.2.178 voc3 #virtual3 4.2.10 Example disk devices On all nodes, the shared disk devices should be accessible through the same devices names. Raw Device Name Physical Device Name Purpose /dev/raw/raw1 /dev/sda1 ASM Disk 1: +DATA1 /dev/raw/raw2 /dev/sdb1 ASM Disk 1: +DATA1 /dev/raw/raw3 /dev/sdc1 ASM Disk 2: +RECOV1 /dev/raw/raw4 /dev/sdd1 ASM Disk 2: +RECOV1 /dev/raw/raw5 /dev/sde1 OCR Disk (on RAW device) /dev/raw/raw6 /dev/sdf1 Voting Disk (on RAW device) 4.3 CRS installation: --------------------- 4.3.1 First install CRS in its own home directory First install CRS in its own home directory, e.g. CRS10gHome, apart from the Oracle home dir. As Oracle user: ./runInstaller --------------------------------------------------- | | Screen 1 |Specify File LOcations | | | |Source | |Path: /install/crs10g/Disk1/stage/products.xml | | | |Destination | |Name: CRS10gHome | |Path: /u01/app/oracle/product/10.1.0/CRS10gHome | | | --------------------------------------------------- --------------------------------------------------- | | Screen 2 |Cluster Configuration | | | |Cluster Name: lec1 | | | | Public Node Name Private Node Name | | --------------------------------------------- | | |oc1 | p0c1 | | | |-------------------------------------------- | | |oc2 | p0c2 | | | |-------------------------------------------- | | |oc3 | poc3 | | | |-------------------------------------------- | --------------------------------------------------- In the next screen, you specify which of your networks is to be used as the public interface (to connect to the public network) and which will be used for the private interconnect to support cache fushion and the cluster heartbeat. --------------------------------------------------- | | Screen 3 |Private Interconnect Enforcement | | | | | | | | Interface Name Subnet Interface type | | --------------------------------------------- | | |eth0 |192.168.2.0 |Public | | | |-------------------------------------------- | | |eth1 |192.168.1.0 |Private | | | |-------------------------------------------- | | | --------------------------------------------------- In the next screen, you specify /dev/raw/raw5 as the raw disk for the Oracle Cluster Registry. --------------------------------------------------- | | Screen 4 |Oracle Cluster Registry | | | |Specify OCR Location: /dev/raw/raw5 | | | --------------------------------------------------- In a similar fashion you specify the location of the Voting Disk. --------------------------------------------------- | | Screen 5 |Voting Disk | | | |Specify Voting Disk: /dev/raw/raw6 | | | --------------------------------------------------- You now have to execute the /u01/app/oracle/orainventory/orainstRoot.sh script on all Cluster Nodes as the root user. After this, you can continue with the other window, and see an "Install Summary" screen. Now you click "Install" and the installation begins. Apart from the node you work on, the software will also be copied to the other nodes as well. After the installation is complete, you are once again prompted to run a script as root on each node of the Cluster. This is the script "/u01/app/oracle/product/10.1.0/CRS10gHome/root.sh". -- The olsnodes command. After finishing the CSR installation, you can verify that the installation completed successfully by running on any node the following command: # cd /u01/app/oracle/product/10.1.0/CRS10gHome/bin # olsnodes -n oc1 1 oc2 2 oc3 3 4.4 Database software installation: ----------------------------------- You can install the database software into the same directory in each node. With OCFS2, you might do one install in a common shared directory for all nodes. Because CSR is already running, the OUI detects that, and because its cluster aware, it provides you with the options to install a clustered implementation. You start the installation by running ./runInstaller as the oracle user on one node. For most part, it looks the same as a single-instance installation. After the file location screen, that is source and destination, you will see this screen: --------------------------------------------------- | | |Specify Hardware Cluster Installation Mode | | | | o Cluster installation mode | | | | Node name | | --------------------------------------------- | | | [] oc1 | | | | [] oc2 | | | | [] oc3 | | | --------------------------------------------- | | | | o Local installation (non cluster) | | | |-------------------------------------------------| Most of the time, you will do a "software only" installation, and create the database later with the DBCA. For the first node only, after some time, the Virtual IP Configuration Assistant, VIPCA, will start. Here you can configure the Virtual IP adresses you will use for application failover and the Enterprise Manager Agent. Here you will select the Virtual IP's for all nodes. VIPCA only needs to run once per Cluster. 4.5 Creating the RAC database with DBCA: ---------------------------------------- Launching the DBCA for installing a RAC database is much the same as launching DBCA for a single instance. If DBCA detects cluster software installed, it gives you the option to install a RAC database or a single instance. as oracle user: % dbca & --------------------------------------------------- | | |Welcome to the database configuration assistant | | | | | | | | o Oracle Real Application Cluster database | | | | o Oracle single instance database | | | |-------------------------------------------------| After selecting RAC, the next screen gives you the option to select nodes: --------------------------------------------------- | | |Select the nodes on which you want to create | |the cluster database. The local node oc1 will | |always be used whether or not it is selected. | | | | Node name | | --------------------------------------------- | | | [] oc1 | | | | [] oc2 | | | | [] oc3 | | | --------------------------------------------- | | | | | |-------------------------------------------------| In the next screens, you can choose the type of database (oltp, dw etc..), and all other items, just like a single instance install. At a cetain point, you can choose to use ASM diskgroups, flash-recovery area etc.. =========================================== Note 5. RAC tools an utilities. =========================================== Example 1: removing and adding a failed node -------------------------------------------- Suppose, using above example, that instance rac3 on node oc3, fails. Suppose that you need to repair the node (e.g. harddisk crash). -- Remove the instance: % srvctl remove instance -d rac -i rac3 Remove instance rac3 for the database rac (y/n)? y -- Remove the node from the cluster: # cd /u01/app/oracle/product/10.1.0/CRS10gHome/bin # ./olsnode -n oc1 1 oc2 2 oc3 3 # cd ../install # ./rootdeletenode.sh oc3,3 # cd ../bin # ./olsnode -n oc1 1 oc2 2 # Suppose that you have repared host oc3. We now want to add it back into the cluster. Host oc3 has the OS newly installed, and its /etc/host file is just like it is on the other nodes. -- Add the node at the clusterware layer: From oc1 or oc2, go to the $CRS_Home/oui/bin directory, and run # ./addNode.sh A graphical screen pops up, and you are able to add oc3 to the cluster. Al CRS files are copied to the new node. To start the services on the new node, you are then prompted to run "rootaddnode.sh" on the active node and "root.sh" on the new node. # ./rootaddnode.sh # ssh oc3 # cd /u01/app/oracle product/10.1.0/CRS10gHome # ./root.sh -- Install the Oracle software on the new node: Example 2: showing all nodes from a node ---------------------------------------- # lsnodes -v # cd /u01/app/oracle/product/10.1.0/CRS10gHome/bin # ./olsnodes -n oc1 1 oc2 2 oc3 3 Example 3: using svrctl ----------------------- The Server Control SVRCTL utility is installed on each node by default. You can use SRVCTL to start and stop the database and instances, manage configuration information, and to move or remove instances and services. Some SVRCTL operations store configuration information in the OCR. SVRCTL performs other operations, such as starting and stopping instances, by sending request to the Oracle Clusterware process CSRD, which then starts or stops the Oracle Clusterware resources. srvctl must be run from the $ORACLE_HOME of the RAC you are administering. The basic format of a srvctl command is srvctl [options] where command is one of enable|disable|start|stop|relocate|status|add|remove|modify|getenv|setenv|unsetenv|config and the target, or object, can be a -database, -instance, -service, -ASM instance, or the -nodeapps. -- Example 1: To view help: % svrctl -h % svrctl command -h -- Example 2: To see the SRVCTL version number, enter % svrctl -V -- Example 3. Bring up the MYSID1 instance of the MYSID database. % srvctl start instance -d MYSID -i MYSID1 -- Example 4. Stop the MYSID database: all its instances and all its services, on all nodes. % srvctl stop database -d MYSID The following command mounts all of the non-running instances, using the default connection information: % srvctl start database -d orcl -o mount -- Example 5. Stop the nodeapps on the myserver node. NB: Instances and services also stop. % srvctl stop nodeapps -n myserver -- Example 6. Add the MYSID3 instance, which runs on the myserver node, to the MYSID clustered database. % srvctl add instance -d MYSID -i MYSID3 -n myserver -- Example 7. Add a new node, the mynewserver node, to a cluster. % srvctl add nodeapps -n mynewserver -o $ORACLE_HOME -A 149.181.201.1/255.255.255.0/eth1 (The -A flag precedes an address specification.) -- Example 8. To change the VIP (virtual IP) on a RAC node, use the command % srvctl modify nodeapps -A new_address -- Example 9. Status of components . Find out whether the nodeapps on mynewserver are up. % srvctl status nodeapps -n mynewserver VIP is running on node: mynewserver GSD is running on node: mynewserver Listener is not running on node: mynewserver ONS daemon is running on node: mynewserver . Find out whether the ASM is running: % srvctl status asm -n docrac1 ASM instance +ASM1 is running on node docrac1. . Find status of cluster database % srvctl status database -d EOPP Instance EOPP1 is running on node dbq0201 Instance EOPP2 is running on node dbq0102 % srvctl config database -d EOPP dbq0201 EOPP1 /ora/product/10.2.0/db dbq0102 EOPP2 /ora/product/10.2.0/db % srvctl config service -d EOPP opp.et.supp PREF: EOPP1 AVAIL: EOPP2 opp.et.grid PREF: EOPP1 AVAIL: EOPP2 -- Example 10. The following command and output show the expected configuration for a three node database called ORCL. % srvctl config database -d ORCL server01 ORCL1 /u01/app/oracle/product/10.1.0/db_1 server02 ORCL2 /u01/app/oracle/product/10.1.0/db_1 server03 ORCL3 /u01/app/oracle/product/10.1.0/db_1 -- Example 11. Disable the ASM instance on myserver for maintenance. % srvctl disable asm -n myserver -- Example 12. Debugging srvctl Debugging srvctl in 10g couldn't be easier. Simply set the SRVM_TRACE environment variable. % export SRVM_TRACE=true -- Example 13. Question Version 10G RAC Q: how to add a listener to the nodeapps using the srvctl command ?? or even if it can be added using srvctl ?? A: just edit listener.ora on all concerned nodes and add entries ( the usual way). srvctl will automatically make use of it. For example % srvctl start database -d SAMPLE will start database SAMPLE and its associated listener LSNR_SAMPLE. -- Example 14. Adding services. % srvctl add database -d ORCL -o /u01/app/oracle/product/10.1.0/db_1 % srvctl add instance -d ORCL -i ORCL1 -n server01 % srvctl add instance -d ORCL -i ORCL2 -n server02 % srvctl add instance -d ORCL -i ORCL3 -n server03 -- Example 15. Administering ASM Instances with SRVCTL in RAC You can use SRVCTL to add, remove, enable, and disable an ASM instance as described in the following procedure: Use the following to add configuration information about an existing ASM instance: % srvctl add asm -n node_name -i asm_instance_name -o oracle_home Use the following to remove an ASM instance: % srvctl remove asm -n node_name [-i asm_instance_name] -- Example 16. Stop multiple instances. The following command provides its own connection information to shut down the two instances orcl3 and orcl4 using the IMMEDIATE option: % srvctl stop instance -d orcl -i "orcl3,orcl4" -o immediate -c "sysback/oracle as sysoper" -- Example 17. Showing policies. Clusterware can automatically start your RAC database when the system restarts. You can use Automatic or Manual "policies", to control whether clusterware restarts RAC. To display the current policy: % srvctl config database -d database_name -a To change to another policy: % srvctl modify database -d database_name -y policy_name -- Example 18. % srvctl start service -d DITOB -- More examples % srvctl remove instance -d rac -i rac3 % srvctl disable instance -d orcl -i orcl2 % srvctl enable instance -d orcl -i orcl2 Example 4: crsctl ----------------- Use CRSCTL to Control Your Clusterware Oracle Clusterware enables servers in an Oracle database Real Application Cluster to coordinate simultaneous workload on the same database files. The crsctl command provides administrators many useful capabilities. For example, with crsctl, you can check Clusterware health disable/enable Oracle Clusterware startup on boot, find information on the voting disk and check the Clusterware version, and more. 1. Do you want to check the health of the Clusterware? # crsctl check crs CSS appears healthy CRS appears healthy EVM appears healthy 2. Do you want to reboot a node for maintenance without Clusterware coming up on boot? ## Disable clusterware on machine2 bootup: # crsctl disable crs ## Stop the database then stop clusterware processes: # srvctl stop instance –d db –i db2 # crsctl stop crs # reboot ## Enable clusterware on machine bootup: # crsctl enable crs # crsctl start crs # srvctl start instance –d db –i db2 3. Do you wonder where your voting disk is? # crsctl query css votedisk 0. 0 /dev/raw/raw2 4. Do you need to find out what clusterware version is running on a server? # crsctl query crs softwareversion CRS software version on node [db2] is [10.2.0.2.0] 5. Adding and Removing Voting Disks You can dynamically add and remove voting disks after installing Oracle RAC. Do this using the following commands where path is the fully qualified path for the additional voting disk. Run the following command as the root user to add a voting disk: # crsctl add css votedisk path Run the following command as the root user to remove a voting disk: # crsctl delete css votedisk path Example 5: cluvfy ----------------- The Cluster Verification Utility pre or post validates an Oracle Clusterware environment or configuration. We found the CVU utility to be very useful for checking a cluster server environment for RAC. The CVU can check shared storage, interconnects, server systems and user permissions. The Universal Installer runs the verification utility at the end of the cluster ware install. The utility can also be run from the command line with parameters and options to validate components. For example, a script that verifies a cluster using cluvfy is named runcluvfy.sh and is located on the /clusterware/cluvfy directory in the installation area. This script unpacks the utility, sets environment variables and executes the verification command. This command verifies that the hosts atlanta1, atlanta2 and atlanta3 are ready for a clustered database install of release 2. ./runcluvfy.sh stage -pre dbinst -n atlanta1,atlanta2,atlanta3 -r 10gR2 -osdba dba –verbose The results of the command above check user and group equivalence across machines, connectivity, interface settings, system requirements like memory, disk space and kernel settings and versions, required Linux package existence and so on. Any problems are reported as errors, all successful checks are marked as passed. Many other aspects of the cluster can be verified with this utility for Release 2 or Release 1. Some more examples: -- Checking for Available Shared Storage with CVU To check for all shared file systems available across all nodes on the cluster, use the following CVU command: % cluvfy comp ssa -n node_list Remember to use the full path name and the runcluvfy.bat command on the installation media and include the list of nodes in your cluster, separated by commas, for the node_list. The following example is for a system with two nodes, node1 and node2, and the installation media on drive F: % runcluvfy.bat comp ssa -n node1,node2 If you want to check the shared accessibility of a specific shared storage type to specific nodes in your cluster, then use the following command syntax: % cluvfy comp ssa -n node_list -s storageID_list In the preceding syntax, the variable node_list is the list of nodes you want to check, separated by commas, and the variable storageID_list is the list of storage device IDs for the storage devices managed by the file system type that you want to check. ================================= Note 6: Example tnsnames.ora in RAC ================================= Example 1: ---------- tnsnames.ora File TEST = (DESCRIPTION = (LOAD_BALANCE = ON) (ADDRESS_LIST = (ADDRESS = (PROTOCOL = TCP)(HOST = testlinux1)(PORT = 1521)) (ADDRESS = (PROTOCOL = TCP)(HOST = testlinux2)(PORT = 1521))) (CONNECT_DATA = (SERVICE_NAME = TEST)))) TEST1 = (DESCRIPTION = (ADDRESS_LIST = (LOAD_BALANCE = ON) (ADDRESS = (PROTOCOL = TCP)(HOST = testlinux1)(PORT = 1521))) (CONNECT_DATA = (SERVICE_NAME = TEST)(INSTANCE_NAME = TEST1))) TEST2 = (DESCRIPTION = (ADDRESS_LIST = (LOAD_BALANCE = ON) (ADDRESS = (PROTOCOL = TCP)(HOST = testlinux2)(PORT = 1521))) (CONNECT_DATA = (SERVICE_NAME = TEST)(INSTANCE_NAME = TEST2))) EXTPROC_CONNECTION_DATA = (DESCRIPTION = (ADDRESS_LIST = (ADDRESS = (PROTOCOL = IPC)(KEY = EXTPROC))) (CONNECT_DATA = (SID=PLSExtProc)(PRESENTATION = RO))) LISTENERS_TEST = (ADDRESS = (PROTOCOL = TCP)(HOST = testlinux1)(PORT = 1521)) (ADDRESS = (PROTOCOL = TCP)(HOST = testlinux2)(PORT = 1521)) Example 2: ---------- Connect-Time Failover From the clients end, when your connection fails at one node or service, you can then do a look up from your tnsnames.ora file and go on seeking a connection with the other available node. Take this example of our 4-node VMware ESX 3.x Oracle Linux Servers: FOKERAC = (DESCRIPTION = (ADDRESS = (PROTOCOL = TCP)(HOST = nick01.wolga.com)(PORT = 1521)) (ADDRESS = (PROTOCOL = TCP)(HOST = nick02.wolga. com)(PORT = 1521)) (ADDRESS = (PROTOCOL = TCP)(HOST = brian01.wolga. com)(PORT = 1521)) (ADDRESS = (PROTOCOL = TCP)(HOST = brian02.wolga. com)(PORT = 1521)) (CONNECT_DATA = (SERVICE_NAME = fokerac) ) ) Here the first address in the list is tried at the client’s end. Should the connection to nick01.wolga.nl fail, then the next address, nick02.wolga.nl, will be tried. This phenomenon is called connection-time failover. You could very well have a 32-node RAC cluster monitoring the galactic system at NASA and thus have all those nodes typed in your tnsnames.ora file. Moreover, these entries do not necessarily have to be part of the RAC cluster. So it is possible that you are using Streams, Log Shipping or Advanced Replication to maintain your HA (High Availability) model. These technologies facilitate continued processing of the database by such a HA (High Availability) model in a non-RAC environment. In a RAC environment we know (and expect) the data to be the same across all nodes since there is only one database. Example 3: ---------- TAF (Transparent Application Failover) Transparent Application Failover actually refers to a failover that occurs when a node or instance is unavailable due to an outage or other reason that prohibits a connection to be established on that node. This can be set to on with the following parameter FAILOVER. Setting it to ON will activate the TAF. It is turned on by default unless you set it to OFF to disable it. Now, when you turn it on you have two types of connections available by the means of the FAILOVER_MODE parameter. The type can be session, which is default or select. When the type is SESSION, if the instance fails, then the user is automatically connected to the next available node without the user’s manual intervention. The SQL statements need to be carried out again on the next node. However, when you set the TYPE to SELECT, then if you are connected and are in the middle of your query, then your query will be restarted after you have been failed over to the next available node. Take this example of our tnsnames.ora file, (go to the section beginning with CONNECT_DATA): (CONNECT_DATA = (SERVER = DEDICATED) (SERVICE_NAME = fokerac.wolga.com) (FAILOVER_MODE = (TYPE = SELECT) (METHOD = BASIC) (RETRIES = 180) (DELAY = 5) ) ) ============================================== Note 7: Notes about Backup and Restore of RAC ============================================== 7.1 Backing up Voting Disk: --------------------------- Run the following command to backup a voting disk. Perform this operation on every voting disk as needed where 'voting_disk_name' is the name of the active voting disk, and 'backup_file_name' is the name of the file to which you want to backup the voting disk contents: # dd if=voting_disk_name of=backup_file_name When you use the dd command for making backups of the voting disk, the backup can be performed while the Cluster Ready Services (CRS) process is active; you do not need to stop the crsd.bin process before taking a backup of the voting disk. -- Adding and Removing Voting Disks You can dynamically add and remove voting disks after installing Oracle RAC. Do this using the following commands where path is the fully qualified path for the additional voting disk. Run the following command as the root user to add a voting disk: # crsctl add css votedisk path Run the following command as the root user to remove a voting disk: # crsctl delete css votedisk path 7.2 Recovering Voting Disk: --------------------------- Run the following command to recover a voting disk where 'backup_file_name' is the name of the voting disk backupfile, and 'voting_disk_name' is the name of the active voting disk: # dd if=backup_file_name of=voting_disk_name 7.3 Backup and Recovery OCR: ---------------------------- Oracle Clusterware automatically creates OCR backups every 4 hours. At any one time, Oracle Clusterware always retains the latest 3 backup copies of the OCR that are 4 hours old, 1 day old, and 1 week old. You cannot customize the backup frequencies or the number of files that Oracle Clusterware retains. You can use any backup software to copy the automatically generated backup files at least once daily to a different device from where the primary OCR file resides. The default location for generating backups on Red Hat Linux systems is "CRS_home/cdata/cluster_name" where cluster_name is the name of your cluster and CRS_home is the home directory of your Oracle Clusterware installation. -- Viewing Available OCR Backups To find the most recent backup of the OCR, on any node in the cluster, use the following command: # ocrconfig -showbackup -- Backing Up the OCR Because of the importance of OCR information, Oracle recommends that you use the ocrconfig tool to make copies of the automatically created backup files at least once a day. In addition to using the automatically created OCR backup files, you should also export the OCR contents to a file before and after making significant configuration changes, such as adding or deleting nodes from your environment, modifying Oracle Clusterware resources, or creating a database. Exporting the OCR contents to a file lets you restore the OCR if your configuration changes cause errors. For example, if you have unresolvable configuration problems, or if you are unable to restart your cluster database after such changes, then you can restore your configuration by importing the saved OCR content from the valid configuration. To export the contents of the OCR to a file, use the following command, where backup_file_name is the name of the OCR backup file you want to create: # ocrconfig -export backup_file_name -- Recovering the OCR This section describes two methods for recovering the OCR. The first method uses automatically generated OCR file copies and the second method uses manually created OCR export files. In event of a failure, before you attempt to restore the OCR, ensure that the OCR is unavailable. Run the following command to check the status of the OCR: # ocrcheck If this command does not display the message 'Device/File integrity check succeeded' for at least one copy of the OCR, then both the primary OCR and the OCR mirror have failed. You must restore the OCR from a backup. -- Restoring the Oracle Cluster Registry from Automatically Generated OCR Backups When restoring the OCR from automatically generated backups, you first have to determine which backup file you will use for the recovery. To restore the OCR from an automatically generated backup on a Red Hat Linux system: Identify the available OCR backups using the ocrconfig command: # ocrconfig -showbackup Note: You must be logged in as the root user to run the ocrconfig command. Review the contents of the backup using the following ocrdump command, where file_name is the name of the OCR backup file: $ ocrdump -backupfile file_name As the root user, stop Oracle Clusterware on all the nodes in your Oracle RAC cluster by executing the following command: # crsctl stop crs Repeat this command on each node in your Oracle RAC cluster. As the root user, restore the OCR by applying an OCR backup file that you identified in step 1 using the following command, where file_name is the name of the OCR that you want to restore. Make sure that the OCR devices that you specify in the OCR configuration exist, and that these OCR devices are valid before running this command. # ocrconfig -restore file_name As the root user, restart Oracle Clusterware on all the nodes in your cluster by restarting each node, or by running the following command: # crsctl start crs Repeat this command on each node in your Oracle RAC cluster. Use the Cluster Verify Utility (CVU) to verify the OCR integrity. Run the following command, where the -n all argument retrieves a list of all the cluster nodes that are configured as part of your cluster: $ cluvfy comp ocr -n all [-verbose] -- Recovering the OCR from an OCR Export File Using the ocrconfig -export command enables you to restore the OCR using the -import option if your configuration changes cause errors. To restore the previous configuration stored in the OCR from an OCR export file: Place the OCR export file that you created previously with the ocrconfig -export command in an accessible directory on disk. As the root user, stop Oracle Clusterware on all the nodes in your Oracle RAC cluster by executing the following command: # crsctl stop crs Repeat this command on each node in your Oracle RAC cluster. As the root user, restore the OCR data by importing the contents of the OCR export file using the following command, where file_name is the name of the OCR export file: # ocrconfig -import file_name As the root user, restart Oracle Clusterware on all the nodes in your cluster by restarting each node, or by running the following command: # crsctl start crs Repeat this command on each node in your Oracle RAC cluster. Use the CVU to verify the OCR integrity. Run the following command, where the -n all argument retrieves a list of all the cluster nodes that are configured as part of your cluster: $ cluvfy comp ocr -n all [-verbose] 7.4 RMAN snapshot controlfile: ------------------------------ RMAN> SHOW SNAPSHOT CONTROLFILE NAME; RMAN> CONFIGURE SNAPSHOT CONTROLFILE NAME TO 'ORACLE_HOME/dbf/scf/snap_prod.cf'; ================================= Note 8: Noticable items in 10g RAC ================================= 8.1 SPFILE: ----------- If an initialization parameter applies to all instances, use *. notation, otherwise prefix the parameter with the name of the instance. For example: *.OPEN_CURSORS=500 prod1.OPEN_CURSORS=1000 8.2 Start and stop of RAC: -------------------------- 8.2.1 Stopping RAC: ------------------- #### NOTE 1: #### > Stop Oracle Clusterware or Cluster Ready Services Processes If you are modifying an Oracle Clusterware or Oracle Cluster Ready Services (CRS) installation, then shut down the following Oracle Database 10g services. Note: You must perform these steps in the order listed. Shut down any processes in the Oracle home on each node that might be accessing a database; for example, shut down Oracle Enterprise Manager Database Control. Note: Before you shut down any processes that are monitored by Enterprise Manager Grid Control, set a blackout in Grid Control for the processes that you intend to shut down. This is necessary so that the availability records for these processes indicate that the shutdown was planned downtime, rather than an unplanned system outage. Shut down all Oracle RAC instances on all nodes. To shut down all Oracle RAC instances for a database, enter the following command, where db_name is the name of the database: $ oracle_home/bin/srvctl stop database -d db_name Shut down all ASM instances on all nodes. To shut down an ASM instance, enter the following command, where node is the name of the node where the ASM instance is running: $ oracle_home/bin/srvctl stop asm -n node Stop all node applications on all nodes. To stop node applications running on a node, enter the following command, where node is the name of the node where the applications are running $ oracle_home/bin/srvctl stop nodeapps -n node Log in as the root user, and shut down the Oracle Clusterware or CRS process by entering the following command on all nodes: # CRS_home/bin/crsctl stop crs #### END NOTE 1 #### #### NOTE 2: #### To stop process in an existing Oracle Real Application Clusters Database, where you want to shut down the entire database, complete the following steps. -- Shut Down Oracle Real Application Clusters Databases Shut down any existing Oracle Database instances on each node, with normal or immediate priority. If Automatic Storage Management (ASM) is running, then shut down all databases that use ASM, and then shut down the ASM instance on each node of the cluster. Note: -- Stop All Oracle Processes Stop all listener and other processes running in the Oracle home directories where you want to modify the database software. Note: If you shut down ASM instances, then you must first shut down all database instances that use ASM, even if these databases run from different Oracle homes. -- Stop Oracle Clusterware or Cluster Ready Services Processes If you are modifying an Oracle Clusterware or Oracle Cluster Ready Services (CRS) installation, then shut down the following Oracle Database 10g services. Note: You must perform these steps in the order listed. Shut down any processes in the Oracle home on each node that might be accessing a database; for example, shut down Oracle Enterprise Manager Database Control. Note: Before you shut down any processes that are monitored by Enterprise Manager Grid Control, set a blackout in Grid Control for the processes that you intend to shut down. This is necessary so that the availability records for these processes indicate that the shutdown was planned downtime, rather than an unplanned system outage. Shut down all Oracle RAC instances on all nodes. To shut down all Oracle RAC instances for a database, enter the following command, where db_name is the name of the database: $ oracle_home/bin/srvctl stop database -d db_name Shut down all ASM instances on all nodes. To shut down an ASM instance, enter the following command, where node is the name of the node where the ASM instance is running: $ oracle_home/bin/srvctl stop asm -n node Stop all node applications on all nodes. To stop node applications running on a node, enter the following command, where node is the name of the node where the applications are running $ oracle_home/bin/srvctl stop nodeapps -n node Log in as the root user, and shut down the Oracle Clusterware or CRS process by entering the following command on all nodes: # CRS_home/bin/crsctl stop crs #### END NOTE 2 #### Notes about Starting up: ------------------------ crsd : Cluster Ready Services Daemon (CRSD) occsd : Oracle Cluster Synchronization Server Daemon (OCSSD), the CCS. evmd : Event Manager Daemon (EVMD). evmlogger The CRSD manages the HA functionality by starting, stopping, and failing over the application resources and maintaining the profiles and current states in the Oracle Cluster Registry (OCR) whereas the OCSSD manages the participating nodes in the cluster by using the voting disk. The OCSSD also protects against the data corruption potentially caused by "split brain" syndrome by forcing a machine to reboot. >Linux: # cat /etc/inittab | grep crs h3:35:respawn:/etc/init.d/init.crsd run > /dev/null 2>&1 /dev/null 2>&1 /dev/null 2>&1 ls -al *init* init.crs init.crsd init.cssd init.evmd # cat /etc/inittab .. .. h1:35:respawn:/etc/init.d/init.evmd run > /dev/null 2>&1 /dev/null 2>&1 /dev/null 2>&1 calls crsd correct order for stopping: Reverse order of startup. crsd should be shutdown before cssd and evmd. evmd should be shutdown before cssd. init.crs stop: init.crsd init.evmd init.cssd init.crs start init.cssd autostart|manualstart ------------------------------------------- links: http://dmx0201.nl.eu.abnamro.com:7900/wi https://dmp0101.nl.eu.abnamro.com:1159/em ------------------------------------------- ============================ 35. ORACLE STREAMS AND CDC: ============================ 35.1 Data replication, Heterogeneous Services, Gateway. Streams: ================================================================ To connect Oracle to a non Oracle database: There are a couple of answers a) http://www.oracle.com/gateways/ is the most complete. distributed query, distributed transactions -- 100% functionality. Lets you treat DB2 as if it were an Oracle instance for all intents and purposes. b) generic connectivity. If you have ODBC on the SERVER (oracle server) and can use that to connect to DB2, you can use generic connectivity. Less functional then a) http://asktom.oracle.com/pls/asktom/f?p=100:11:::::P11_QUESTION_ID:4406709207206 c) lastly, you can get their type4 (thin) jdbc (all java) drivers and load them into Oracle. Then, you can write a java stored procedure in Oracle that accesses DB2 over jdbc. 35.2 Information on CDC: ======================== Change Data Capture can capture and publish committed change data in either of the following modes: -- Synchronous Triggers on the source database allow change data to be captured immediately, as each SQL statement that performs a data manipulation language (DML) operation (INSERT, UPDATE, or DELETE) is made. In this mode, change data is captured as part of the transaction modifying the source table. Synchronous Change Data Capture is available with Oracle Standard Edition and Enterprise Edition. -- Asynchronous By taking advantage of the data sent to the redo log files, change data is captured after a SQL statement that performs a DML operation is committed. In this mode, change data is not captured as part of the transaction that is modifying the source table, and therefore has no effect on that transaction. Asynchronous Change Data Capture is available with Oracle Enterprise Edition only. There are three modes of asynchronous Change Data Capture: HotLog, Distributed HotLog, and AutoLog. Asynchronous Change Data Capture is built on, and provides a relational interface to, Oracle Streams. See Oracle Streams Concepts and Administration for information on Oracle Streams. - Change tables With any CDC mode, change tables are involved. A given change table contains the change data resulting from DML operations performed on a given source table. A change table consists of two things: the change data itself, which is stored in a database table, ; and the system metadata necessary to maintain the change table, which includes control columns. The publisher specifies the source columns that are to be included in the change table. Typically, for a change table to contain useful data, the publisher needs to include the primary key column in the change table along with any other columns of interest to subscribers. For example, suppose subscribers are interested in changes that occur to the UNIT_COST and the UNIT_PRICE columns in the sh.costs table. If the publisher does not include the PROD_ID column in the change table, subscribers will know only that the unit cost and unit price of some products have changed, but will be unable to determine for which products these changes have occurred. There are optional and required control columns. The required control columns are always included in a change table; the optional ones are included if specified by the publisher when creating the change table. Control columns are managed by Change Data Capture. - Interface Change Data Capture includes the DBMS_CDC_PUBLISH and DBMS_CDC_SUBSCRIBE packages, which provide easy-to-use publish and subscribe interfaces. - Publish and Subscribe Model Most Change Data Capture systems have one person who captures and publishes change data; this person is the publisher. There can be multiple applications or individuals that access the change data; these applications and individuals are the subscribers. Change Data Capture provides PL/SQL packages to accomplish the publish and subscribe tasks. -- TASKS: => These are the main tasks performed by the publisher: . Determines the source databases and tables from which the subscribers are interested in viewing change data, and the mode (synchronous or one of the asynchronous modes) in which to capture the change data. . Uses the Oracle-supplied package, DBMS_CDC_PUBLISH, to set up the system to capture change data from the source tables of interest. . Allows subscribers to have controlled access to the change data in the change tables by using the SQL GRANT and REVOKE statements to grant and revoke the SELECT privilege on change tables for users and roles. (Keep in mind, however, that subscribers use views, not change tables directly, to access change data.) => These are the main tasks performed by the subscriber: The subscribers are consumers of the published change data. A subscriber performs the following tasks: > Uses the Oracle supplied package, DBMS_CDC_SUBSCRIBE, to: . Create subscriptions A subscription controls access to the change data from one or more source tables of interest within a single change set. A subscription contains one or more subscriber views. A subscriber view is a view that specifies the change data from a specific publication in a subscription. The subscriber is restricted to seeing change data that the publisher has published and has granted the subscriber access to use. See "Subscribing to Change Data" for more information on choosing a method for specifying a subscriber view. . Notify Change Data Capture when ready to receive a set of change data A subscription window defines the time range of rows in a publication that the subscriber can currently see in subscriber views. The oldest row in the window is called the low boundary; the newest row in the window is called the high boundary. Each subscription has its own subscription window that applies to all of its subscriber views. . Notify Change Data Capture when finished with a set of change data > Uses SELECT statements to retrieve change data from the subscriber views. -- Other items: Source Database MODE CHANGE SOURCE Represented Associated Change Set ----------------------------------------------------------------------------------------------------- Synchronous Predefined SYNC_SOURCE Local Predefined SYNC_SET and publisher-defined Async HotLog Predefined HOTLOG_SOURCE Local Publisher-defined Async Distr HotLog Publisher-defined Remote Publisher-defined. Change sets must all be on the same staging database Async AutoLog online Publisher-defined Remote Publisher-defined. There can only be one change set in an AutoLog online change source Asynchronous AutoLog archive Publisher-defined Remote Publisher-defined -- Views intended for Publisher or Subscriber: CHANGE_SOURCES Describes existing change sources. CHANGE_PROPAGATIONS Describes the Oracle Streams propagation associated with a given Distributed HotLog change source on the source database. This view is populated on the source database for 10.2 change sources or on the staging database for 9.2 or 10.1 change sources. CHANGE_PROPAGATION_SETS Describes the Oracle Streams propagation associated with a given Distributed HotLog change set on the staging database. This view is populated on the source database for 10.2 change sources or on the staging database for 9.2 or 10.1 change sources. CHANGE_SETS Describes existing change sets. CHANGE_TABLES Describes existing change tables. DBA_SOURCE_TABLES Describes all published source tables in the database. DBA_PUBLISHED_COLUMNS Describes all published columns of source tables in the database. DBA_SUBSCRIPTIONS Describes all subscriptions. DBA_SUBSCRIBED_TABLES Describes all source tables to which any subscriber has subscribed. DBA_SUBSCRIBED_COLUMNS Describes the columns of source tables to which any subscriber has subscribed. ALL_SOURCE_TABLES Describes all published source tables accessible to the current user. USER_SOURCE_TABLES Describes all published source tables owned by the current user. ALL_PUBLISHED_COLUMNS Describes all published columns of source tables accessible to the current user. USER_PUBLISHED_COLUMNS Describes all published columns of source tables owned by the current user. ALL_SUBSCRIPTIONS Describes all subscriptions accessible to the current user. USER_SUBSCRIPTIONS Describes all the subscriptions owned by the current user. ALL_SUBSCRIBED_TABLES Describes the source tables to which any subscription accessible to the current user has subscribed. USER_SUBSCRIBED_TABLES Describes the source tables to which the current user has subscribed. ALL_SUBSCRIBED_COLUMNS Describes the columns of source tables to which any subscription accessible to the current user has subscribed. USER_SUBSCRIBED_COLUMNS Describes the columns of source tables to which the current user has subscribed. -- Adjusting Initialization Parameter Values When Oracle Streams Values Change Asynchronous Change Data Capture uses an Oracle Streams configuration for each change set. This Streams configuration consists of a Streams capture process and a Streams apply process, with an accompanying queue and queue table. Each Streams configuration uses additional processes, parallel execution servers, and memory. For details about the Streams architecture, see Oracle Streams Concepts and Administration. Oracle Streams capture and apply processes each have a parallelism parameter that is used to improve performance. When a publisher first creates a change set, its capture parallelism value and apply parallelism value are each 1. If desired, a publisher can increase one or both of these values using Streams interfaces. If Oracle Streams capture parallelism and apply parallelism values are increased after change sets are created, the DBA (or DBAs in the case of the Distributed HotLog mode) must adjust initialization parameter values accordingly. How these adjustments are made vary slightly, depending on the mode of Change Data Capture being employed, as described in the following sections. -- Adjustments for HotLog and AutoLog Change Data Capture For HotLog and AutoLog change data capture, adjustments to initialization parameters are made on the staging database. Examples below demonstrate how to obtain the current capture parallelism and apply parallelism values for change set CHICAGO_DAILY. By default, each parallelism value is 1, so the amount by which a given parallelism value has been increased is the returned value minus 1. Example 1 Obtaining the Oracle Streams Capture Parallelism Value for a Change Set SELECT cp.value FROM DBA_CAPTURE_PARAMETERS cp, CHANGE_SETS cset WHERE cset.SET_NAME = 'CHICAGO_DAILY' AND cset.CAPTURE_NAME = cp.CAPTURE_NAME AND cp.PARAMETER = 'PARALLELISM'; Example 2 Obtaining the Oracle Streams Apply Parallelism Value for a Change Set SELECT ap.value FROM DBA_APPLY_PARAMETERS ap, CHANGE_SETS cset WHERE cset.SET_NAME = 'CHICAGO_DAILY' AND cset.APPLY_NAME = ap.APPLY_NAME AND ap.parameter = 'PARALLELISM'; The staging database DBA must adjust the staging database initialization parameters as described in the following list to accommodate the parallel execution servers and other processes and memory required for Change Data Capture: PARALLEL_MAX_SERVERS For each change set for which Oracle Streams capture or apply parallelism values were increased, increase the value of this parameter by the increased Streams parallelism value. For example, if the statement in Example 1 returns a value of 2, and the statement in Example 2 returns a value of 3, then the staging database DBA should increase the value of the PARALLEL_MAX_SERVERS parameter by (2-1) + (3-1), or 3 for the CHICAGO_DAILY change set. If the Streams capture or apply parallelism values have increased for other change sets, increases for those change sets must also be made. PROCESSES For each change set for which Oracle Streams capture or apply parallelism values were changed, increase the value of this parameter by the sum of increased Streams parallelism values. See the previous list item, PARALLEL_MAX_SERVERS, for an example. STREAMS_POOL_SIZE For each change set for which Oracle Streams capture or apply parallelism values were changed, increase the value of this parameter by (10MB * (the increased capture parallelism value)) + (1MB * increased apply parallelism value). For example, if the statement in Example 1 returns a value of 2, and the statement in Example 2 returns a value of 3, then the staging database DBA should increase the value of the STREAMS_POOL_SIZE parameter by (10 MB * (2-1) + 1MB * (3-1)), or 12MB for the CHICAGO_DAILY change set. If the Oracle Streams capture or apply parallelism values have increased for other change sets, increases for those change sets must also be made. See Oracle Streams Concepts and Administration for more information on Streams capture parallelism and apply parallelism values. See Oracle Database Reference for more information about database initialization parameters. Note 3: Oracle 10.2 Sync CDC Example: ===================================== CDC Mode : Synchroneous CDC Source table : hr.cdc_demo Changing table : cdcadmin.cdc_demo_ct table with added change data added by means of handler function: hr.salary_history conn / as sysdba -- *NIX only define _editor=vi -- validate database parameters archive log list -- Archive Mode show parameter aq_tm_processes -- min 3 show parameter compatible -- must be 10.1.0 or above show parameter global_names -- must be TRUE show parameter job_queue_processes -- min 2 recommended 4-6 show parameter open_links -- not less than the default 4 show parameter shared_pool_size -- must be 0 or at least 200MB show parameter streams_pool_size -- min. 480MB (10MB/capture 1MB/apply) show parameter undo_retention -- min. 3600 (1 hr.) (900) -- Examples of altering initialization parameters alter system set aq_tm_processes=3 scope=BOTH; alter system set compatible='10.2.0.1.0' scope=SPFILE; alter system set global_names=TRUE scope=BOTH; alter system set job_queue_processes=6 scope=BOTH; alter system set open_links=4 scope=SPFILE; alter system set streams_pool_size=200M scope=BOTH; -- very slow if making smaller alter system set undo_retention=3600 scope=BOTH; alter system set processes=150 scope=SPFILE; /* JOB_QUEUE_PROCESSES (current value) + 2 PARALLEL_MAX_SERVERS (current value) + (5 * (the number of change sets planned)) PROCESSES (current value) + (7 * (the number of change sets planned)) SESSIONS (current value) + (2 * (the number of change sets planned)) */ -- Retest parameter after modification shutdown immediate; startup mount; alter database archivelog; -- important alter database force logging; -- one option among several alter database add supplemental log data; alter database open; -- validate archivelogging archive log list alter system switch logfile; archive log list -- validate force and supplemental logging SELECT supplemental_log_data_min, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all, force_logging FROM gv$database; SELECT force_logging FROM dba_tablespaces; desc dba_hist_streams_apply_sum SELECT apply_name, reader_total_messages_dequeued, reader_lag, server_total_messages_applied FROM dba_hist_streams_apply_sum; -- examine CDC related data dictionary objects SELECT table_name FROM dba_tables WHERE owner = 'SYS' AND table_name LIKE 'CDC%$'; desc cdc_system$ SELECT * FROM cdc_system$; Setup As SYS - Create Streams Administrators conn / as sysdba SELECT * FROM dba_streams_administrator; CREATE USER cdcadmin IDENTIFIED BY cdcadmin DEFAULT TABLESPACE users TEMPORARY TABLESPACE temp QUOTA 0 ON system QUOTA 10M ON sysaux QUOTA 20M ON users; -- system privs GRANT create session TO cdcadmin; GRANT create table TO cdcadmin; GRANT create sequence TO cdcadmin; GRANT create procedure TO cdcadmin; GRANT dba TO cdcadmin; -- role privs GRANT execute_catalog_role TO cdcadmin; GRANT select_catalog_role TO cdcadmin; -- object privileges GRANT execute ON dbms_cdc_publish TO cdcadmin; GRANT execute ON dbms_cdc_subscribe TO cdcadmin; -- do also to HR -- streams specific priv execute dbms_streams_auth.grant_admin_privilege('CDCADMIN'); SELECT account_status, created FROM dba_users WHERE username = 'CDCADMIN'; SELECT * FROM dba_sys_privs WHERE grantee = 'CDCADMIN'; SELECT username FROM dba_users u, streams$_privileged_user s WHERE u.user_id = s.user#; SELECT * FROM dba_streams_administrator; Prepare Schema Tables for CDC Replication conn / as sysdba alter user hr account unlock identified by hr; connect hr/hr desc employees SELECT * FROM employees; -- create CDC demo table CREATE TABLE cdc_demo AS SELECT * FROM employees; ALTER TABLE cdc_demo ADD CONSTRAINT pk_cdc_demo PRIMARY KEY (employee_id) USING INDEX PCTFREE 0; -- a second way to implement supplemental logging ALTER TABLE cdc_demo ADD SUPPLEMENTAL LOG DATA (ALL) COLUMNS; -- table to track salary history changes originating in cdc_demo -- CREATE TABLE salary_history ( employee_id NUMBER(6), first_name VARCHAR2(20), last_name VARCHAR2(25), old_salary NUMBER(8,2), new_salary NUMBER(8,2), pct_change NUMBER(4,2), action_date DATE); SELECT table_name FROM user_tables; Instantiate Source Table conn cdcadmin/cdcadmin desc dba_capture_prepared_tables SELECT table_name, scn, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all FROM dba_capture_prepared_tables; dbms_capture_adm.prepare_table_instantiation( table_name IN VARCHAR2, supplemental_logging IN VARCHAR2 DEFAULT 'keys'); Note: This procedure performs the synchronization necessary for instantiating the table at another database. This procedure records the lowest SCN of the table for instantiation. SCNs subsequent to the lowest SCN for an object can be used for instantiating the object. exec dbms_capture_adm.prepare_table_instantiation('HR.CDC_DEMO'); SELECT table_name, scn, supplemental_log_data_pk PK, supplemental_log_data_ui UI, supplemental_log_data_fk FK, supplemental_log_data_all "ALL" FROM dba_capture_prepared_tables; Create Synchronous Change Set conn cdcadmin/cdcadmin col object_name format a30 SELECT object_name, object_type FROM user_objects ORDER BY 2,1; dbms_cdc_publish.create_change_set( change_set_name IN VARCHAR2, description IN VARCHAR2 DEFAULT NULL, change_source_name IN VARCHAR2, stop_on_ddl IN CHAR DEFAULT 'N', begin_date IN DATE DEFAULT NULL, end_date IN DATE DEFAULT NULL); -- this may take a minute or two exec dbms_cdc_publish.create_change_set('CDC_DEMO_SET', 'Synchronous Demo Set', 'SYNC_SOURCE'); SELECT object_name, object_type FROM user_objects ORDER BY 2,1; conn / as sysdba desc cdc_change_sets$ set linesize 121 col set_name format a20 col capture_name format a20 col queue_name format a20 col queue_table_name format a20 SELECT set_name, capture_name, queue_name, queue_table_name FROM cdc_change_sets$; SELECT set_name, change_source_name, capture_enabled, stop_on_ddl, publisher FROM change_sets; Create Change Table conn cdcadmin/cdcadmin dbms_cdc_publish.create_change_table( owner IN VARCHAR2, change_table_name IN VARCHAR2, change_set_name IN VARCHAR2, source_schema IN VARCHAR2, source_table IN VARCHAR2, column_type_list IN VARCHAR2, capture_values IN VARCHAR2, -- BOTH, NEW, OLD rs_id IN CHAR, row_id IN CHAR, user_id IN CHAR, timestamp IN CHAR, object_id IN CHAR, source_colmap IN CHAR, target_colmap IN CHAR, options_string IN VARCHAR2); BEGIN dbms_cdc_publish.create_change_table('CDCADMIN', 'CDC_DEMO_CT', 'CDC_DEMO_SET', 'HR', 'CDC_DEMO', 'EMPLOYEE_ID NUMBER(6), FIRST_NAME VARCHAR2(20), LAST_NAME VARCHAR2(25), SALARY NUMBER', 'BOTH', 'Y', 'Y', 'Y', 'N', 'N', 'Y', 'Y', ' TABLESPACE USERS pctfree 0 pctused 99'); END; / GRANT select ON cdc_demo_ct TO hr; conn / as sysdba SELECT set_name, change_source_name, queue_name, queue_table_name FROM cdc_change_sets$; desc cdc_change_tables$ SELECT change_set_name, source_schema_name, source_table_name FROM cdc_change_tables$; conn cdcadmin/cdcadmin SELECT object_name, object_type FROM user_objects ORDER BY 2,1; col high_value format a15 SELECT table_name, composite, partition_name, high_value FROM user_tab_partitions; Create Subscription conn hr/hr dbms_cdc_subscribe.create_subscription( change_set_name IN VARCHAR2, description IN VARCHAR2, subscription_name IN VARCHAR2); exec dbms_cdc_subscribe.create_subscription('CDC_DEMO_SET', 'Sync Capture Demo Set', 'CDC_DEMO_SUB'); conn / as sysdba set linesize 121 col description format a30 col subscription_name format a20 col username format a10 SELECT subscription_name, handle, set_name, username, earliest_scn, description FROM cdc_subscribers$; Subscribe to and Activate Subscription conn hr/hr dbms_cdc_subscribe.subscribe( subscription_name IN VARCHAR2, source_schema IN VARCHAR2, source_table IN VARCHAR2, column_list IN VARCHAR2, subscriber_view IN VARCHAR2); BEGIN dbms_cdc_subscribe.subscribe('CDC_DEMO_SUB', 'HR', 'CDC_DEMO', 'EMPLOYEE_ID, FIRST_NAME, LAST_NAME, SALARY', 'CDC_DEMO_SUB_VIEW'); END; / desc user_subscriptions SELECT set_name, subscription_name, status FROM user_subscriptions; SELECT set_name, subscription_name, status FROM dba_subscriptions; dbms_cdc_subscribe.activate_subscription( subscription_name IN VARCHAR2); exec dbms_cdc_subscribe.activate_subscription('CDC_DEMO_SUB'); SELECT set_name, subscription_name, status FROM user_subscriptions; Create Procedure To Populate Salary History Table conn hr/hr /* Create a stored procedure to populate the new HR.SALARY_HISTORY table. The procedure extends the subscription window of the CDC_DEMP_SUB subscription to get the most recent set of source table changes. It uses the subscriber's DEMO_SUB_VIEW view to scan the changes and insert them into the SALARY_HISTORY table. It then purges the subscription window to indicate that it is finished with that set of changes. */ CREATE OR REPLACE PROCEDURE update_salary_history IS CURSOR cur IS SELECT * FROM ( SELECT 'I' opt, cscn$, rsid$, employee_id, first_name, last_name, 0 old_salary, salary new_salary, commit_timestamp$ FROM cdc_demo_sub_view WHERE operation$ = 'I ' UNION ALL SELECT 'D' opt, cscn$, rsid$, employee_id, first_name, last_name, salary old_salary, 0 new_salary, commit_timestamp$ FROM cdc_demo_sub_view WHERE operation$ = 'D ' UNION ALL SELECT 'U' opt , v1.cscn$, v1.rsid$, v1.employee_id, v1.first_name, v1.last_name, v1.salary old_salary, v2.salary new_salaryi, v1.commit_timestamp$ FROM cdc_demo_sub_view v1, cdc_demo_sub_view v2 WHERE v1.operation$ = 'UU' and v2.operation$ = 'UN' AND v1.cscn$ = v2.cscn$ AND v1.rsid$ = v2.rsid$ AND ABS(v1.salary - v2.salary) > 0) ORDER BY cscn$, rsid$; percent NUMBER; BEGIN --Step 1 Get the change (extend the window). dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); FOR rec IN cur LOOP IF rec.opt = 'I' THEN INSERT INTO salary_history (employee_id, first_name, last_name, old_salary, new_salary, pct_change, action_date) VALUES (rec.employee_id, rec.first_name, rec.last_name, 0, rec.new_salary, NULL, rec.commit_timestamp$); END IF; IF rec.opt = 'D' THEN INSERT INTO salary_history (employee_id, first_name, last_name, old_salary, new_salary, pct_change, action_date) VALUES (rec.employee_id, rec.first_name, rec.last_name, rec.old_salary, 0, NULL, rec.commit_timestamp$); END IF; IF rec.opt = 'U' THEN percent := (rec.new_salary - rec.old_salary) / rec.old_salary * 100; INSERT INTO salary_history (employee_id, first_name, last_name, old_salary, new_salary, pct_change, action_date) VALUES (rec.employee_id, rec.first_name, rec.last_name, rec.old_salary, rec.new_salary, percent, rec.commit_timestamp$); END IF; END LOOP; COMMIT; --Step 3 Purge the window of consumed data dbms_cdc_subscribe.purge_window('CDC_DEMO_SUB'); END update_salary_history; / desc dba_hist_streams_apply_sum SELECT apply_name, reader_total_messages_dequeued, reader_lag, server_total_messages_applied FROM dba_hist_streams_apply_sum; DML On Source Table conn hr/hr SELECT employee_id, first_name, last_name, salary FROM cdc_demo ORDER BY 1 DESC; SELECT employee_id, first_name, last_name, salary FROM cdc_demo_sub_view; SELECT * FROM salary_history; UPDATE cdc_demo SET salary = salary+1 WHERE employee_id = 100; COMMIT; SELECT employee_id,first_name,last_name,salary FROM cdc_demo_sub_view; exec update_salary_history; SELECT employee_id,first_name,last_name,salary FROM cdc_demo_sub_view; SELECT * FROM salary_history; -- Capture Cleanup conn hr/hr exec dbms_cdc_subscribe.drop_subscription('CDC_DEMO_SUB'); conn / as sysdba -- reverse prepare table instantiation exec dbms_capture_adm.abort_table_instantiation('HR.CDC_DEMO'); -- drop the change table exec dbms_cdc_publish.drop_change_table('CDCADMIN', 'CDC_DEMO_CT', 'Y'); exec dbms_cdc_publish.drop_change_table('TDBA_CDC','EMPCANCELLEDHOTELBKNG', 'Y'); -- drop the change set exec dbms_cdc_publish.drop_change_set('CDC_DEMO_SET'); conn hr/hr drop table salary_history purge; drop table cdc_demo purge; drop procedure update_salary_history; conn / as sysdba drop user cdcadmin; Note 4: Oracle 10.2 ASync Hotlog CDC Example: ============================================= conn / as sysdba -- *NIX only define _editor=vi -- validate database parameters archive log list -- Archive Mode show parameter aq_tm_processes -- min 3 show parameter compatible -- must be 10.1.0 or above show parameter global_names -- must be TRUE show parameter job_queue_processes -- min 2 recommended 4-6 show parameter open_links -- not less than the default 4 show parameter shared_pool_size -- must be 0 or at least 200MB show parameter streams_pool_size -- min. 480MB (10MB/capture 1MB/apply) show parameter undo_retention -- min. 3600 (1 hr.) (900) -- Examples of altering initialization parameters alter system set aq_tm_processes=3 scope=BOTH; alter system set compatible='10.2.0.1.0' scope=SPFILE; alter system set global_names=TRUE scope=BOTH; alter system set job_queue_processes=6 scope=BOTH; alter system set open_links=4 scope=SPFILE; alter system set streams_pool_size=200M scope=BOTH; -- very slow if making smaller alter system set undo_retention=3600 scope=BOTH; /* JOB_QUEUE_PROCESSES (current value) + 2 PARALLEL_MAX_SERVERS (current value) + (5 * (the number of change sets planned)) PROCESSES (current value) + (7 * (the number of change sets planned)) SESSIONS (current value) + (2 * (the number of change sets planned)) */ -- Retest parameter after modification shutdown immediate; startup mount; alter database archivelog; -- important alter database force logging; -- one option among several alter database add supplemental log data; alter database open; -- validate archivelogging archive log list alter system switch logfile; archive log list -- validate force and supplemental logging SELECT supplemental_log_data_min, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all, force_logging FROM gv$database; SELECT force_logging FROM dba_tablespaces; -- examine existing queues desc dba_queues set linesize 121 col owner format a6 col queue_table format a25 col user_comment format a31 SELECT owner, name, queue_table, queue_type, user_comment FROM dba_queues ORDER BY 1,4,2; -- examine existing streams desc dba_hist_streams_capture SELECT capture_name, total_messages_captured, total_messages_enqueued FROM dba_hist_streams_capture; desc dba_hist_streams_apply_sum SELECT apply_name, reader_total_messages_dequeued, reader_lag, server_total_messages_applied FROM dba_hist_streams_apply_sum; -- examine CDC related data dictionary objects SELECT table_name FROM dba_tables WHERE owner = 'SYS' AND table_name LIKE 'CDC%$'; desc cdc_system$ SELECT * FROM cdc_system$; Setup As SYS - Create Streams Administrators conn / as sysdba SELECT * FROM dba_streams_administrator; CREATE USER cdcadmin IDENTIFIED BY cdcadmin DEFAULT TABLESPACE users TEMPORARY TABLESPACE temp QUOTA 0 ON system QUOTA 10M ON sysaux QUOTA 20M ON users; -- system privs GRANT create session TO cdcadmin; GRANT create table TO cdcadmin; GRANT create sequence TO cdcadmin; GRANT create procedure TO cdcadmin; GRANT dba TO cdcadmin; -- role privs GRANT execute_catalog_role TO cdcadmin; GRANT select_catalog_role TO cdcadmin; -- object privileges GRANT execute ON dbms_cdc_publish TO cdcadmin; GRANT execute ON dbms_cdc_subscribe TO cdcadmin; -- required for this demo but not by CDC GRANT execute ON dbms_lock TO cdcadmin; -- streams specific priv execute dbms_streams_auth.grant_admin_privilege('CDCADMIN'); SELECT account_status, created FROM dba_users WHERE username = 'CDCADMIN'; SELECT * FROM dba_sys_privs WHERE grantee = 'CDCADMIN'; SELECT username FROM dba_users u, streams$_privileged_user s WHERE u.user_id = s.user#; SELECT * FROM dba_streams_administrator; Prepare Schema Tables for CDC Replication conn / as sysdba alter user hr account unlock identified by hr; connect hr/hr desc employees SELECT * FROM employees; -- create CDC demo table CREATE TABLE cdc_demo AS SELECT * FROM employees; -- a second way to implement supplemental logging ALTER TABLE cdc_demo ADD SUPPLEMENTAL LOG DATA (ALL) COLUMNS; -- table to track salary history changes originating in cdc_demo CREATE TABLE salary_history ( employee_id NUMBER(6) NOT NULL, job_id VARCHAR2(10) NOT NULL, department_id NUMBER(4), old_salary NUMBER(8,2), new_salary NUMBER(8,2), percent_change NUMBER(4,2), salary_action_date DATE); SELECT table_name FROM user_tables; Instantiate Source Table conn / as sysdba desc dba_capture_prepared_tables SELECT table_name, scn, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all FROM dba_capture_prepared_tables; dbms_capture_adm.prepare_table_instantiation( table_name IN VARCHAR2, supplemental_logging IN VARCHAR2 DEFAULT 'keys'); Note: This procedure performs the synchronization necessary for instantiating the table at another database. This procedure records the lowest SCN of the table for instantiation. SCNs subsequent to the lowest SCN for an object can be used for instantiating the object. exec dbms_capture_adm.prepare_table_instantiation(table_name => 'HR.CDC_DEMO'); SELECT table_name, scn, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all FROM dba_capture_prepared_tables; Create Asynchronous HotLog Change Set conn cdcadmin/cdcadmin col object_name format a30 SELECT object_name, object_type FROM user_objects ORDER BY 2,1; dbms_cdc_publish.create_change_set( change_set_name IN VARCHAR2, description IN VARCHAR2 DEFAULT NULL, change_source_name IN VARCHAR2, stop_on_ddl IN CHAR DEFAULT 'N', begin_date IN DATE DEFAULT NULL, end_date IN DATE DEFAULT NULL); -- this may take awhile don't be impatient exec dbms_cdc_publish.create_change_set('CDC_DEMO_SET', 'CDC Demo 2 Change Set', 'HOTLOG_SOURCE', 'Y', NULL, NULL); -- here is why SELECT object_name, object_type FROM user_objects ORDER BY 2,1; SELECT table_name, tablespace_name, iot_type FROM user_tables; col high_value format a15 SELECT table_name, composite, partition_name, high_value FROM user_tab_partitions; conn / as sysdba desc cdc_change_sets$ set linesize 121 col set_name format a20 col capture_name format a20 col queue_name format a20 col queue_table_name format a20 SELECT set_name, capture_name, queue_name, queue_table_name FROM cdc_change_sets$; SELECT set_name, change_source_name, capture_enabled, stop_on_ddl, publisher FROM change_sets; SELECT process_type, name FROM streams$_process_params; Create Change Table conn cdcadmin/cdcadmin dbms_cdc_publish.create_change_table( owner IN VARCHAR2, change_table_name IN VARCHAR2, change_set_name IN VARCHAR2, source_schema IN VARCHAR2, source_table IN VARCHAR2, column_type_list IN VARCHAR2, capture_values IN VARCHAR2, -- BOTH, NEW, OLD rs_id IN CHAR, row_id IN CHAR, user_id IN CHAR, timestamp IN CHAR, object_id IN CHAR, source_colmap IN CHAR, target_colmap IN CHAR, options_string IN VARCHAR2); BEGIN dbms_cdc_publish.create_change_table('CDCADMIN', 'CDC_DEMO_CT', 'CDC_DEMO_SET', 'HR', 'CDC_DEMO', 'EMPLOYEE_ID NUMBER(6), FIRST_NAME VARCHAR2(20), LAST_NAME VARCHAR2(25), EMAIL VARCHAR2(25), PHONE_NUMBER VARCHAR2(20), HIRE_DATE DATE, JOB_ID VARCHAR2(10), SALARY NUMBER, COMMISSION_PCT NUMBER, MANAGER_ID NUMBER, DEPARTMENT_ID NUMBER', 'BOTH', 'N', 'N', 'N', 'N', 'N', 'N', 'Y', NULL); END; / exec dbms_cdc_publish.alter_change_table('CDCADMIN', 'CDC_DEMO_CT', rs_id=>'Y'); GRANT select ON cdc_demo_ct TO hr; conn / as sysdba SELECT set_name, change_source_name, queue_name, queue_table_name FROM cdc_change_sets$; desc cdc_change_tables$ SELECT change_set_name, source_schema_name, source_table_name FROM cdc_change_tables$; Enable Capture conn / as sysdba SELECT set_name, change_source_name, capture_enabled FROM cdc_change_sets$; conn cdcadmin/cdcadmin dbms_cdc_publish.alter_change_set( change_set_name IN VARCHAR2, description IN VARCHAR2 DEFAULT NULL, remove_description IN CHAR DEFAULT 'N', enable_capture IN CHAR DEFAULT NULL, recover_after_error IN CHAR DEFAULT NULL, remove_ddl IN CHAR DEFAULT NULL, stop_on_ddl IN CHAR DEFAULT NULL); exec dbms_cdc_publish.alter_change_set(change_set_name=>'CDC_DEMO_SET', enable_capture=> 'Y'); conn / as sysdba SELECT set_name, change_source_name, capture_enabled FROM cdc_change_sets$; Create Subscription conn hr/hr dbms_cdc_subscribe.create_subscription( change_set_name IN VARCHAR2, description IN VARCHAR2, subscription_name IN VARCHAR2); exec dbms_cdc_subscribe.create_subscription('CDC_DEMO_SET', 'cdc_demo subx', 'CDC_DEMO_SUB'); conn / as sysdba set linesize 121 col description format a30 col subscription_name format a20 col username format a10 SELECT subscription_name, handle, set_name, username, earliest_scn, description FROM cdc_subscribers$; Subscribe to and Activate Subscription conn hr/hr dbms_cdc_subscribe.subscribe( subscription_name IN VARCHAR2, source_schema IN VARCHAR2, source_table IN VARCHAR2, column_list IN VARCHAR2, subscriber_view IN VARCHAR2); BEGIN dbms_cdc_subscribe.subscribe('CDC_DEMO_SUB', 'HR', 'CDC_DEMO', 'EMPLOYEE_ID, FIRST_NAME, LAST_NAME, EMAIL, PHONE_NUMBER, HIRE_DATE, JOB_ID, SALARY, COMMISSION_PCT, MANAGER_ID, DEPARTMENT_ID', 'CDC_DEMO_SUB_VIEW'); END; / desc user_subscriptions SELECT set_name, subscription_name, status FROM user_subscriptions; dbms_cdc_subscribe.activate_subscription( subscription_name IN VARCHAR2); exec dbms_cdc_subscribe.activate_subscription('CDC_DEMO_SUB'); SELECT set_name, subscription_name, status FROM user_subscriptions; Create Procedure To Populate Salary History Table conn hr/hr /* Create a stored procedure to populate the new HR.SALARY_HISTORY table. The procedure extends the subscription window of the CDC_DEMP_SUB subscription to get the most recent set of source table changes. It uses the subscriber's DEMO_SUB_VIEW view to scan the changes and insert them into the SALARY_HISTORY table. It then purges the subscription window to indicate that it is finished with that set of changes. */ CREATE OR REPLACE PROCEDURE update_salary_history IS CURSOR cur IS SELECT * FROM ( SELECT 'I' opt, cscn$, rsid$, employee_id, job_id, department_id, 0 old_salary, salary new_salary, commit_timestamp$ FROM cdc_demo_sub_view WHERE operation$ = 'I ' UNION ALL SELECT 'D' opt, cscn$, rsid$, employee_id, job_id, department_id, salary old_salary, 0 new_salary, commit_timestamp$ FROM cdc_demo_sub_view WHERE operation$ = 'D ' UNION ALL SELECT 'U' opt , v1.cscn$, v1.rsid$, v1.employee_id, v1.job_id, v1.department_id, v1.salary old_salary, v2.salary new_salaryi, v1.commit_timestamp$ FROM cdc_demo_sub_view v1, cdc_demo_sub_view v2 WHERE v1.operation$ = 'UO' and v2.operation$ = 'UN' AND v1.cscn$ = v2.cscn$ AND v1.rsid$ = v2.rsid$ AND ABS(v1.salary - v2.salary) > 0) ORDER BY cscn$, rsid$; percent NUMBER; BEGIN -- Get the next set of changes to the HR.CDC_DEMO source table dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); -- Process each change FOR rec IN cur LOOP IF rec.opt = 'I' THEN INSERT INTO salary_history VALUES (rec.employee_id, rec.job_id, rec.department_id, 0, rec.new_salary, NULL, rec.commit_timestamp$); END IF; IF rec.opt = 'D' THEN INSERT INTO salary_history VALUES (rec.employee_id, rec.job_id, rec.department_id, rec.old_salary, 0, NULL, rec.commit_timestamp$); END IF; IF rec.opt = 'U' THEN percent := (rec.new_salary - rec.old_salary) / rec.old_salary * 100; INSERT INTO salary_history VALUES (rec.employee_id, rec.job_id, rec.department_id, rec.old_salary, rec.new_salary, percent, rec.commit_timestamp$); END IF; END LOOP; -- Indicate subscriber is finished with this set of changes dbms_cdc_subscribe.purge_window('CDC_DEMO_SUB'); END update_salary_history; / Create Procedure To Wait For Changes /* Create function CDCADMIN.WAIT_FOR_CHANGES to enable this demo to run predictably. The asynchronous nature of CDC HotLog mode means that there is a delay for source table changes to appear in the CDC change table and the subscriber view. By default this procedure waits up to 3 minutes for the change table and 1 additional minute for the subscriber view. This can be adjusted if it is insufficient. The caller must specify the name of the change table and the number of rows expected to be in the change table. The caller may also optionally specify a different number of seconds to wait for changes to appear in the change table. */ conn cdcadmin/cdcadmin CREATE OR REPLACE FUNCTION wait_for_changes ( rowcount NUMBER, -- number of rows to wait for maxwait_seconds NUMBER := 300) -- maximum time to wait, in seconds RETURN VARCHAR2 AUTHID CURRENT_USER AS numrows NUMBER := 0; -- number of rows in change table slept NUMBER := 0; -- total time slept sleep_time NUMBER := 3; -- number of seconds to sleep return_msg VARCHAR2(100); -- informational message keep_waiting BOOLEAN := TRUE; -- whether to keep waiting BEGIN WHILE keep_waiting LOOP SELECT COUNT(*) INTO numrows FROM CDC_DEMO_CT; -- Got expected number of rows IF numrows >= rowcount THEN keep_waiting := FALSE; return_msg := 'Change table contains at least ' || TO_CHAR(rowcount) || ' rows'; EXIT; -- Reached maximum number of seconds to wait ELSIF slept > maxwait_seconds THEN return_msg := ' - Timed out while waiting for the change table to reach ' || TO_CHAR(rowcount) || ' rows'; EXIT; END IF; dbms_lock.sleep(sleep_time); slept := slept+sleep_time; END LOOP; -- additional wait time for changes to become available to subscriber view dbms_lock.sleep(60); RETURN return_msg; END wait_for_changes; / Preparation for DML -- In a separate terminal window cd $ORACLE_BASE/admin/ORCL/bdump tail -f alertorcl.log -- tailing the alert log allows us to watch log miner at work -- open a SQL*Plus session as SYS desc gv$streams_capture set linesize 121 col state format a20 SELECT capture_name, logminer_id, state, total_messages_captured FROM gv$streams_capture; -- open a SQL*Plus session as SYS desc gv$streams_apply_reader set linesize 121 col state format a20 SELECT apply_name, state, total_messages_dequeued FROM gv$streams_apply_reader; DML On Source Table conn hr/hr UPDATE cdc_demo SET salary = salary + 500 WHERE job_id = 'SH_CLERK'; UPDATE cdc_demo SET salary = salary + 1000 WHERE job_id = 'ST_CLERK'; UPDATE cdc_demo SET salary = salary + 1500 WHERE job_id = 'PU_CLERK'; COMMIT; INSERT INTO cdc_demo (employee_id, first_name, last_name, email, phone_number, hire_date, job_id, salary, commission_pct, manager_id, department_id) VALUES (207, 'Mary', 'Lee', 'MLEE', '310.234.4590', TO_DATE('10-JAN-2003'), 'SH_CLERK', 4000, NULL, 121, 50); INSERT INTO cdc_demo (employee_id, first_name, last_name, email, phone_number, hire_date, job_id, salary, commission_pct, manager_id, department_id) VALUES (208, 'Karen', 'Prince', 'KPRINCE', '345.444.6756', TO_DATE('10-NOV-2003'), 'SH_CLERK', 3000, NULL, 111, 50); INSERT INTO cdc_demo (employee_id, first_name, last_name, email, phone_number, hire_date, job_id, salary, commission_pct, manager_id, department_id) VALUES (209, 'Frank', 'Gate', 'FGATE', '451.445.5678', TO_DATE('13-NOV-2003'), 'IT_PROG', 8000, NULL, 101, 50); INSERT INTO cdc_demo (employee_id, first_name, last_name, email, phone_number, hire_date, job_id, salary, commission_pct, manager_id, department_id) VALUES (210, 'Paul', 'Jeep', 'PJEEP', '607.345.1112', TO_DATE('28-MAY-2003'), 'IT_PROG', 8000, NULL, 101, 50); COMMIT; Validate Capture -- Expecting 94 rows to appear in the change table CDCADMIN.CDC_DEMO_CT. This first -- capture may take a few minutes. Later captures should be substantially faster. conn cdcadmin/cdcadmin SELECT wait_for_changes(94, 180) message FROM dual; Another Test conn hr/hr /* The wait_for_changes function having indicated the changes have been populated apply the changes to the salary_history table */ exec update_salary_history; SELECT employee_id, job_id, department_id, old_salary, new_salary, percent_change FROM salary_history ORDER BY 1, 4, 5; delete from cdc_demo where first_name = 'Mary' and last_name = 'Lee'; delete from cdc_demo where first_name = 'Karen' and last_name = 'Prince'; delete from cdc_demo where first_name = 'Frank' and last_name = 'Gate'; delete from cdc_demo where first_name = 'Paul' and last_name = 'Jeep'; COMMIT; update cdc_demo set salary = salary + 5000 where job_id = 'AD_VP'; update cdc_demo set salary = salary - 1000 where job_id = 'ST_MAN'; update cdc_demo set salary = salary - 500 where job_id = 'FI_ACCOUNT'; COMMIT; -- Expecting 122 rows to appear in the change table CDCADMIN.CDC_DEMO_CT. -- (94 rows from the first set of DMLs and 28 from the second set) conn cdcadmin/cdcadmin SELECT wait_for_changes(122, 180) message from dual; conn hr/hr exec update_salary_history SELECT employee_id, job_id, department_id, old_salary, new_salary, percent_change FROM salary_history order by 1, 4, 5; Capture Cleanup conn hr/hr exec dbms_cdc_subscribe.drop_subscription('CDC_DEMO_SUB'); conn / as sysdba -- reverse prepare table instantiation exec dbms_capture_adm.abort_table_instantiation('HR.CDC_DEMO'); -- drop the change table exec dbms_cdc_publish.drop_change_table('CDCADMIN', 'CDC_DEMO_CT', 'Y'); -- drop the change set exec dbms_cdc_publish.drop_change_set('CDC_DEMO_SET'); conn cdcadmin/cdcadmin drop function wait_for_changes; SELECT COUNT(*) FROM user_objects; conn hr/hr drop table salary_history purge; drop table cdc_demo purge; drop procedure update_salary_history; conn / as sysdba drop user cdcadmin; Note 5: Oracle 9.2 CDC Example: =============================== -- Change table example code -- by Jon Emmons -- www.lifeaftercoffee.com -- NOTE: This code is provided for educational purposes only! Use at your -- own risk. I have only used this code on Oracle 9.2 Enterprise Edition. -- Due to the way variables are handled, this should be run one command at -- a time, but must be run all in the same SQLPlus session. -- Connect as a priveleged user conn system -- Create scott if he doesn't already exist CREATE user scott IDENTIFIED BY tiger DEFAULT tablespace users TEMPORARY tablespace temp quota unlimited ON users; -- Grant scott appropriate priveleges GRANT connect TO scott; GRANT execute_catalog_role TO scott; GRANT select_catalog_role TO scott; GRANT CREATE TRIGGER TO scott; -- Connect up as scott conn scott/tiger -- Create Table CREATE TABLE scott.classes ( class_id NUMBER, class_title VARCHAR2(30), class_instructor VARCHAR2(30), class_term_code VARCHAR2(6), class_credits NUMBER, CONSTRAINT PK_classes PRIMARY KEY (class_id ) ); -- Load some data INSERT INTO classes VALUES (100, 'Reading', 'Jon', '200510', 3); INSERT INTO classes VALUES (101, 'Writing', 'Stacey', '200510', 4); INSERT INTO classes VALUES (102, 'Arithmetic', 'Laurianne', '200530', 3); commit; -- Confirm current data SELECT * FROM classes; -- Set up the change table exec dbms_logmnr_cdc_publish.create_change_table - ('scott', 'classes_ct', 'SYNC_SET', 'scott', 'classes', - 'class_id NUMBER, - class_title VARCHAR2(30), - class_instructor VARCHAR2(30), - class_term_code VARCHAR2(6), - class_credits NUMBER', - 'BOTH', 'Y', 'N', 'N', 'Y', 'N', 'Y', 'N', NULL); -- Subscribe to the change table variable subhandle NUMBER; execute dbms_logmnr_cdc_subscribe.get_subscription_handle - (CHANGE_SET => 'SYNC_SET', - DESCRIPTION => 'Changes to classes table', - SUBSCRIPTION_HANDLE => :subhandle); execute dbms_logmnr_cdc_subscribe.subscribe - (subscription_handle => :subhandle, - source_schema => 'scott', - source_table => 'classes', - column_list => 'class_id, class_title, class_instructor, class_term_code, class_credits'); execute dbms_logmnr_cdc_subscribe.activate_subscription - (SUBSCRIPTION_HANDLE => :subhandle); -- Now modify the table in a few different ways UPDATE classes SET class_title='Math' WHERE class_id=102; INSERT INTO classes VALUES (103, 'Computers', 'Ken', '200510', 1); INSERT INTO classes VALUES (104, 'Racketball', 'Matt', '200530', 2); UPDATE classes SET class_credits=3 WHERE class_id=103; DELETE FROM classes WHERE class_title='Reading'; commit; -- Confirm current data SELECT * FROM classes; -- Now lets check out the change table variable viewname varchar2(40) execute dbms_logmnr_cdc_subscribe.extend_window - (subscription_handle => :subhandle); execute dbms_logmnr_cdc_subscribe.prepare_subscriber_view - (SUBSCRIPTION_HANDLE => :subhandle, - SOURCE_SCHEMA => 'scott', - SOURCE_TABLE => 'classes', - VIEW_NAME => :viewname); print viewname -- This little trick will move the bind variable :viewname into the -- substitution variable named subscribed_view COLUMN myview new_value subscribed_view noprint SELECT :viewname myview FROM dual; -- Examine the actual change data. You could also look at the table in a -- browser such as TOAD for easier viewing. SELECT * FROM &subscribed_view; -- Close the subscriber view execute dbms_logmnr_cdc_subscribe.drop_subscriber_view - (SUBSCRIPTION_HANDLE => :subhandle, - SOURCE_SCHEMA => 'scott', - SOURCE_TABLE => 'classes'); -- Purge the window execute dbms_logmnr_cdc_subscribe.purge_window - (subscription_handle => :subhandle); -- If done altogether, end the subscription execute dbms_logmnr_cdc_subscribe.drop_subscription - (subscription_handle => :subhandle); -- drop the change table exec dbms_logmnr_cdc_publish.drop_change_table('scott', 'classes_ct', 'N'); -- Delete the table DROP TABLE scott.classes; Note 6: ======= DBMS_CDC_PUBLISH: In previous releases, this package was named DBMS_LOGMNR_CDC_PUBLISH. Beginning with release 10g, the LOGMNR string has been removed from the name, resulting in the name DBMS_CDC_PUBLISH. Although both variants of the name are still supported, the variant with the LOGMNR string has been deprecated and may not be supported in a future release The DBMS_CDC_PUBLISH package is used by a publisher to set up an Oracle Change Data Capture system to capture and publish change data from one or more Oracle relational source tables. Change Data Capture captures and publishes only committed data. Oracle Change Data Capture identifies new data that has been added to, updated in, or removed from relational tables, and publishes the change data in a form that is usable by subscribers. Typically, a Change Data Capture system has one publisher who captures and publishes changes for any number of Oracle relational source tables. The publisher then provides subscribers (applications or individuals) with access to the published data. Note 7: ======= Oracle Tips by Burleson Oracle 10g Create the change tables The dbms_cdc_publish.create_change_table procedure is used by the publisher user on the staging database to create change tables. The publisher user creates one or more change tables for each source table to be published, specifies which columns should be included, and specifies the combination of before and after images of the change data to capture. To have more control over the physical properties and tablespace properties of the change tables, the publisher can set the options_string field of the dbms_cdc_publish.create_change_table procedure. The options_string field can contain any option available on the CREATE TABLE statement. The following script creates a change table on the staging database that captures changes made to a source table in the source database. The example uses the sample table pl.project_history. BEGIN DBMS_CDC_PUBLISH.CREATE_CHANGE_TABLE( owner => 'cdcproj', change_table_name => 'PROJ_HIST_CT', change_set_name => 'PROJECT_DAILY', source_schema => 'PL', source_table => 'PROJ_HISTORY', column_type_list => 'EMPLOYEE_ID NUMBER(6),START_DATE DATE, END_DATE DATE, PROJ_ID VARCHAR2(10), DEPARTMENT_ID NUMBER(4)', capture_values => 'both', rs_id => 'y', row_id => 'n', user_id => 'n', timestamp => 'n', object_id => 'n', source_colmap => 'n', target_colmap => 'y', options_string => NULL); END; / PL/SQL procedure successfully completed. This example statement creates a change table named proj_hist_ct, within change set project_daily. The column_type_list parameter is used to identify the columns captured by the change table. Remember that the source_schema and source_table parameters identify the schema and source table that reside in the source database, not the staging database. Note 8: Example using streams (1) ================================= http://blogs.ittoolbox.com/oracle/guide/archives/oracle-streams-configuration-change-data-capture-13501 I have been playing with Oracle Streams again lately. My goal is to capture changes in 10g and send them to a 9i database. Below is the short list for setting up Change Data Capture using Oracle Streams. These steps are mostly from the docs with a few tweaks I have added. This entry only covers setting up the local capture and apply. I'll add the propagation to 9i later this week or next weekend. First the set up: we will use the HR account's Employee table. We'll capture all changes to the Employee table and insert them into an audit table. I'm not necessarily saying this is the way you should audit your database but it makes a nice example. I'll also add a monitoring piece to capture process. I want to be able to see exactly what is being captured when it is being captured. You will need to have sysdba access to follow along with me. Your database must also be in archivelog mode. The changes are picked up from the redo log. So, away we go! The first step is to create out streams administrator. I will follow the guidelines from the oracle docs exactly for this: - Connect as sysdba: sqlplus / as sysdba - Create the streams tablespace (change the name and/or location to suit): create tablespace streams_tbs datafile 'c:\temp\stream_tbs.dbf' size 25M reuse autoextend on maxsize unlimited; - Create our streams administrator: create user strmadmin identified by strmadmin default tablespace streams_tbs quota unlimited on streams_tbs; I haven't quite figured out why, but we need to grant our administrator DBA privs. I think this is a bad thing. There is probably a work around where I could do some direct grants instead but I haven't had time to track those down. grant dba to strmadmin; We also want to grant streams admin privs to the user. BEGIN SYS.DBMS_STREAMS_AUTH.GRANT_ADMIN_PRIVILEGE( grantee => 'strmadmin', grant_privileges => true); END; / -The next steps we'll run as the HR user. conn hr/hr - Grant all access to the employee table to the streams admin: grant all on hr.employee to strmadmin; - We also need to create the employee_audit table. Note that I am adding three columns in this table that do not exist in the employee table. CREATE TABLE employee( employee_id NUMBER(6), first_name VARCHAR2(20), last_name VARCHAR2(25), email VARCHAR2(25), phone_number VARCHAR2(20), hire_date DATE, job_id VARCHAR2(10), salary NUMBER(8,2), commission_pct NUMBER(2,2), manager_id NUMBER(6), department_id NUMBER(4)); ALTER TABLE employee add constraint pk_employee_id PRIMARY KEY; ALTER TABLE employee ADD CONSTRAINT pk_employee_id PRIMARY KEY (employee_id) INSERT INTO hr.employee VALUES(206, 'Albert', 'Sel','avds@antapex.org',NULL, '07-JUN-94', 'AC_ACCOUNT', 777, NULL, NULL, 110); COMMIT; CREATE TABLE employee_audit( employee_id NUMBER(6), first_name VARCHAR2(20), last_name VARCHAR2(25), email VARCHAR2(25), phone_number VARCHAR2(20), hire_date DATE, job_id VARCHAR2(10), salary NUMBER(8,2), commission_pct NUMBER(2,2), manager_id NUMBER(6), department_id NUMBER(4), upd_date DATE, user_name VARCHAR2(30), action VARCHAR2(30)); ALTER TABLE employee_audit ADD CONSTRAINT pk_employee_audit_id PRIMARY KEY (employee_id) - Grant all access to the audit table to the streams admin user: grant all on hr.employee_audit to strmadmin; - We connect as the streams admin user: conn strmadmin/strmadmin We can create a logging table. You would NOT want to do this in a high-volume production system. I am doing this to illustrate user defined monitoring and show how you can get inside the capture process. CREATE TABLE streams_monitor ( date_and_time TIMESTAMP(6) DEFAULT systimestamp, txt_msg CLOB ); - Here we create the queue. Unlike AQ, where you have to create a separate table, this step creates the queue and the underlying ANYDATA table. BEGIN DBMS_STREAMS_ADM.SET_UP_QUEUE( queue_table => 'strmadmin.streams_queue_table', queue_name => 'strmadmin.streams_queue'); END; / - This just defines that we want to capture DML and not DDL. BEGIN DBMS_STREAMS_ADM.ADD_TABLE_RULES( table_name => 'hr.employee', streams_type => 'capture', streams_name => 'capture_emp', queue_name => 'strmadmin.streams_queue', include_dml => true, include_ddl => false, inclusion_rule => true); END; / | Possible errors on that statement: | | ERROR at line 1: | ORA-32593: database supplemental logging attributes in flux | ORA-06512: at "SYS.DBMS_STREAMS_ADM", line 372 | ORA-06512: at "SYS.DBMS_STREAMS_ADM", line 312 | ORA-06512: at line 2 | | Oracle Error :: ORA-32593 | database supplemental logging attributes in flux| | | Cause | there is another process actively modifying the database wide supplemental logging attributes. | | Action | Retry the DDL or the LogMiner dictionary build that raised this error. | | Restaring the database worked for me. - Tell the capture process that we want to know who made the change: BEGIN DBMS_CAPTURE_ADM.INCLUDE_EXTRA_ATTRIBUTE( capture_name => 'capture_emp', attribute_name => 'username', include => true); END; / - We also need to tell Oracle where to start our capture. Change the source_database_name to match your database. DECLARE iscn NUMBER; -- Variable to hold instantiation SCN value BEGIN iscn := DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER(); DBMS_APPLY_ADM.SET_TABLE_INSTANTIATION_SCN( source_object_name => 'hr.employee', source_database_name => 'test10g', instantiation_scn => iscn); END; / Note: To get the latest SCN from a database: SQL> select DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() from dual; DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() ----------------------------------------- 8854917 And the fun part! This is where we define our capture procedure. I'm taking this right from the docs but I'm adding a couple steps. The follwing will be a userdefined procedure, on what to do additionally when changes occurs. CREATE OR REPLACE PROCEDURE emp_dml_handler(in_any IN ANYDATA) IS lcr SYS.LCR$_ROW_RECORD; rc PLS_INTEGER; command VARCHAR2(30); old_values SYS.LCR$_ROW_LIST; BEGIN -- Access the LCR rc := in_any.GETOBJECT(lcr); -- Get the object command type command := lcr.GET_COMMAND_TYPE(); -- I am inserting the XML equivalent of the LCR into the monitoring table. insert into streams_monitor (txt_msg) values (command || DBMS_STREAMS.CONVERT_LCR_TO_XML(in_any) ); -- Set the command_type in the row LCR to INSERT lcr.SET_COMMAND_TYPE('INSERT'); -- Set the object_name in the row LCR to EMP_DEL lcr.SET_OBJECT_NAME('EMPLOYEE_AUDIT'); -- Set the new values to the old values for update and delete IF command IN ('DELETE', 'UPDATE') THEN -- Get the old values in the row LCR old_values := lcr.GET_VALUES('old'); -- Set the old values in the row LCR to the new values in the row LCR lcr.SET_VALUES('new', old_values); -- Set the old values in the row LCR to NULL lcr.SET_VALUES('old', NULL); END IF; -- Add a SYSDATE for upd_date lcr.ADD_COLUMN('new', 'UPD_DATE', ANYDATA.ConvertDate(SYSDATE)); -- Add a user column lcr.ADD_COLUMN('new', 'user_name', lcr.GET_EXTRA_ATTRIBUTE('USERNAME') ); -- Add an action column lcr.ADD_COLUMN('new', 'ACTION', ANYDATA.ConvertVarChar2(command)); -- Make the changes lcr.EXECUTE(true); commit; END; / - Create the DML handlers: BEGIN DBMS_APPLY_ADM.SET_DML_HANDLER( object_name => 'hr.employee', object_type => 'TABLE', operation_name => 'INSERT', error_handler => false, user_procedure => 'strmadmin.emp_dml_handler', apply_database_link => NULL, apply_name => NULL); END; / BEGIN DBMS_APPLY_ADM.SET_DML_HANDLER( object_name => 'hr.employee', object_type => 'TABLE', operation_name => 'UPDATE', error_handler => false, user_procedure => 'strmadmin.emp_dml_handler', apply_database_link => NULL, apply_name => NULL); END; / BEGIN DBMS_APPLY_ADM.SET_DML_HANDLER( object_name => 'hr.employee', object_type => 'TABLE', operation_name => 'DELETE', error_handler => false, user_procedure => 'strmadmin.emp_dml_handler', apply_database_link => NULL, apply_name => NULL); END; / - Create the apply rule. This tells streams, yet again, that we in fact do want to capture changes. The second calls tells streams where to put the info. Change the source_database_name to match your database. DECLARE emp_rule_name_dml VARCHAR2(30); emp_rule_name_ddl VARCHAR2(30); BEGIN DBMS_STREAMS_ADM.ADD_TABLE_RULES( table_name => 'hr.employee', streams_type => 'apply', streams_name => 'apply_emp', queue_name => 'strmadmin.streams_queue', include_dml => true, include_ddl => false, source_database => 'test10g', dml_rule_name => emp_rule_name_dml, ddl_rule_name => emp_rule_name_ddl); DBMS_APPLY_ADM.SET_ENQUEUE_DESTINATION( rule_name => emp_rule_name_dml, destination_queue_name => 'strmadmin.streams_queue'); END; / We don't want to stop applying changes when there is an error, so: BEGIN DBMS_APPLY_ADM.SET_PARAMETER( apply_name => 'apply_emp', parameter => 'disable_on_error', value => 'n'); END; / - Turn on the apply process: BEGIN DBMS_APPLY_ADM.START_APPLY( apply_name => 'apply_emp'); END; / - Turn on the capture process: BEGIN DBMS_CAPTURE_ADM.START_CAPTURE( capture_name => 'capture_emp'); END; / - Connect as HR and make some changes to Employees. sqlplus hr/hr INSERT INTO hr.employee VALUES(207, 'JOHN', 'SMITH','JSMITH@MYCOMPANY.COM',NULL, '07-JUN-94', 'AC_ACCOUNT', 777, NULL, NULL, 110); COMMIT; INSERT INTO hr.employee VALUES(208, 'Piet', 'Pietersen','JSMITH@MYCOMPANY.COM',NULL, '07-JUN-94', 'AC_ACCOUNT', 777, NULL, NULL, 110); COMMIT; INSERT INTO hr.employee VALUES(209, 'Piet', 'Pietersen','JSMITH@MYCOMPANY.COM',NULL, '07-JUN-94', 'AC_ACCOUNT', 777, NULL, NULL, 110); COMMIT; UPDATE hr.employee SET salary=5999 WHERE employee_id=206; COMMIT; DELETE FROM hr.employees WHERE employee_id=207; COMMIT; It takes a few seconds for the data to make it to the logs and then back into the system to be appled. Run this query until you see data (remembering that it is not instantaneous): SELECT employee_id, first_name, last_name, upd_Date, action FROM hr.employee_audit ORDER BY employee_id; Then you can log back into the streams admin account: sqlplus strmadmin/strmadmin View the XML LCR that we inserted during the capture process: set long 9999 set pagesize 0 select * from streams_monitor; That's it! It's really not that much work to capture and apply changes. Of course, it's a little bit more work to cross database instances, but it's not that much. Keep an eye out for a future entry where I do just that. One of the things that amazes me is how little code is required to accomplish this. The less code I have to write, the less code I have to maintain. Thank care, LewisC Note 9: Streams example (2) =========================== The entry builds directly on my last entry, Oracle Streams Configuration: Change Data Capture. This entry will show you how to propagate the changes you captured in that entry to a 9i database. NOTE #1: I would recommend that you run the commands and make sure the last entry works for you before trying the code in this entry. That way you will need to debug as few moving parts as possible. NOTE #2: I have run this code windows to windows, windows to linux, linux to solaris and solaris to solaris. The only time I had any problem at all was solaris to solaris. If you run into problems with propagation running but not sending data, shutdown the source database and restart it. That worked for me. NOTE #3: I have run this code 10g to 10g and 10g to 9i. It works without change between them. NOTE #4: If you are not sure of the exact name of your database (including domain), use global_name, i.e. select * from global_name; NOTE #5: Streams is not available with XE. Download and install EE. If you have 1 GB or more of RAM on your PC, you can download EE and use the DBCA to run two database instances. You do not physically need two machines to get this to work. NOTE #6: I promise this is the last note. Merry Christmas and/or Happy Holidays! Now for the fun part. As I mentioned above, you need two instances for this. I called my first instance ORCL (how creative!) and I called my second instance SECOND. It works for me! ORCL will be my source instance and SECOND will be my target instance. You should already have the CDC code from the last article running in ORCL. ORCL must be in archivelog mode to run CDC. SECOND does not need archivelog mode. Having two databases running on a single PC in archivelog mode can really beat up a poor IDE drive. You already created your streams admin user in ORCL so now do the same thing in SECOND. The code below is mostly the same code that you ran on ORCL. I made a few minor changes in case you are running both instances on a single PC: sqlplus / as sysdba create tablespace streams_second_tbs datafile 'c:\temp\stream_2_tbs.dbf' size 25M reuse autoextend on maxsize unlimited; create user strmadmin identified by strmadmin default tablespace streams_second_tbs quota unlimited on streams_second_tbs; grant dba to strmadmin; Connect as strmadmin. You need to create an AQ table, AQ queue and then start the queue. That's what the code below does. BEGIN DBMS_AQADM.CREATE_QUEUE_TABLE( queue_table => 'lrc_emp_t', queue_payload_type => 'sys.anydata', multiple_consumers => TRUE, compatible => '8.1'); DBMS_AQADM.CREATE_QUEUE( queue_name => 'lrc_emp_q', queue_table => 'lrc_emp_t'); DBMS_AQADM.START_QUEUE ( queue_name => 'lrc_emp_q'); END; / You also need to create a database link. You have to have one from ORCL to SECOND but for debugging, I like a link in both. So, while you're in SECOND, create a link: CREATE DATABASE LINK orcl.world CONNECT TO strmadmin IDENTIFIED BY strmadmin USING 'orcl.world'; Log into ORCL as strmadmin and run the exact same command there. Most of the setup for this is exactly the same between the two instances. Create your link on this side also. CREATE DATABASE LINK second.world CONNECT TO strmadmin IDENTIFIED BY strmadmin USING 'second.world'; Ok, now we have running queues in ORCL and SECOND. While you are logged into ORCL, you will create a propagation schedule. You DO NOT need to run this in SECOND. BEGIN DBMS_STREAMS_ADM.ADD_TABLE_PROPAGATION_RULES( table_name => 'hr.employees', streams_name => 'orcl_2_second', source_queue_name => 'strmadmin.lrc_emp_q', destination_queue_name => 'strmadmin.lrc_emp_q@second.world', include_dml => true, include_ddl => FALSE, source_database => 'orcl.world'); END; / This tells the database to take the data in the local lrc_emp_q and send it to the named destination queue. We're almost done with the propagation now. We just need to change the code we wrote in the last article in our DML handler. Go back and review that code now. We are going to modify the EMP_DML_HANDLER so that we get an enqueue block just above the execute statement: CREATE OR REPLACE PROCEDURE emp_dml_handler(in_any IN ANYDATA) IS lcr SYS.LCR$_ROW_RECORD; rc PLS_INTEGER; command VARCHAR2(30); old_values SYS.LCR$_ROW_LIST; BEGIN -- Access the LCR rc := in_any.GETOBJECT(lcr); -- Get the object command type command := lcr.GET_COMMAND_TYPE(); -- I am inserting the XML equivalent of the LCR into the monitoring table. insert into streams_monitor (txt_msg) values (command || DBMS_STREAMS.CONVERT_LCR_TO_XML(in_any) ); -- Set the command_type in the row LCR to INSERT lcr.SET_COMMAND_TYPE('INSERT'); -- Set the object_name in the row LCR to EMP_DEL lcr.SET_OBJECT_NAME('EMPLOYEE_AUDIT'); -- Set the new values to the old values for update and delete IF command IN ('DELETE', 'UPDATE') THEN -- Get the old values in the row LCR old_values := lcr.GET_VALUES('old'); -- Set the old values in the row LCR to the new values in the row LCR lcr.SET_VALUES('new', old_values); -- Set the old values in the row LCR to NULL lcr.SET_VALUES('old', NULL); END IF; -- Add a SYSDATE value for the timestamp column lcr.ADD_COLUMN('new', 'UPD_DATE', ANYDATA.ConvertDate(SYSDATE)); -- Add a user value for the timestamp column lcr.ADD_COLUMN('new', 'user_name', lcr.GET_EXTRA_ATTRIBUTE('USERNAME') ); -- Add an action column lcr.ADD_COLUMN('new', 'ACTION', ANYDATA.ConvertVarChar2(command)); DECLARE enqueue_options DBMS_AQ.enqueue_options_t; message_properties DBMS_AQ.message_properties_t; message_handle RAW(16); recipients DBMS_AQ.aq$_recipient_list_t; BEGIN recipients(1) := sys.aq$_agent( 'anydata_subscriber', 'strmadmin.lrc_emp_q@second.world', NULL); message_properties.recipient_list := recipients; DBMS_AQ.ENQUEUE( queue_name => 'strmadmin.lrc_emp_q', enqueue_options => enqueue_options, message_properties => message_properties, payload => anydata.convertObject(lcr), msgid => message_handle); EXCEPTION WHEN OTHERS THEN insert into streams_monitor (txt_msg) values ('Anydata: ' || DBMS_UTILITY.FORMAT_ERROR_STACK ); END; -- Make the changes lcr.EXECUTE(true); commit; END; / The declaration section above created some variable required for an enqueue. We created a subscriber (that's the name of the consumer). We will use that name to dequeue the record in the SECOND instance. We then enqueued our LCR as an ANYDATA datatype. I put the exception handler there in case there are any problems with our enqueue. That's all it takes. Insert some records into the HR.employees table and commit them. Then log into strmadmin@second and select * from the lrc_emp_t table. You should have as many records there as you inserted. There are not a lot of moving parts so there aren't many things that will go wrong. Propagation is where I have the most troubles. You can query DBA_PROPAGATION to see if you have any propagation errors. That's it for moving the data from 10g to 9i. In my next article, I will show you how to dequeue the data and put it into the employee_audit table on the SECOND side. If you have any problems or any questions please post them. Take care, LewisC Note 10: CDC 9.2 ================ A change table is required for each source table. The publisher uses the procedure DBMS_LOGMNR_CDC_PUBLISH .CREATE_CHANGE_TABLE to create change tables, as shown in Listing 1. In this example, the change tables corresponding to PRICE_LIST and SALES_TRAN are named CDC_PRICE_LIST and CDC_SALES_TRAN respectively. This procedure creates a change table in a specified schema. execute DBMS_CDC_PUBLISH.CREATE_CHANGE_TABLE(OWNER => 'cdc1', \ CHANGE_TABLE_NAME => 'emp_ct', \ CHANGE_SET_NAME => 'SYNC_SET', \ SOURCE_SCHEMA => 'scott', \ SOURCE_TABLE => 'emp', \ COLUMN_TYPE_LIST => 'empno number, ename varchar2(10), job varchar2(9), mgr number, hiredate date, deptno number', \ CAPTURE_VALUES => 'both', \ RS_ID => 'y', \ ROW_ID => 'n', \ USER_ID => 'n', \ TIMESTAMP => 'n', \ OBJECT_ID => 'n',\ SOURCE_COLMAP => 'n', \ TARGET_COLMAP => 'y', \ OPTIONS_STRING => NULL); This procedure adds columns to, or drops columns from, an existing change table. EXECUTE DBMS_LOGMNR_CDC_PUBLISH.ALTER_CHANGE_TABLE (OWNER => 'cdc1') \ CHANGE_TABLE_NAME => 'emp_ct' \ OPERATION => ADD \ ADD_COLUMN_LIST => '' \ RS_ID => 'Y' \ ROW_ID => 'N' \ USER_ID => 'N' \ TIMESTAMP => 'N' \ OBJECT_ID => 'N' \ SOURCE_COLMAP => 'N' \ TARGET_COLMAP => 'N'); This procedure allows a publisher to drop a subscriber view in the subscriber's schema. EXECUTE sys.DBMS_CDC_SUBSCRIBE.DROP_SUBSCRIBER_VIEW( \ SUBSCRIPTION_HANDLE =>:subhandle, \ SOURCE_SCHEMA =>'scott', \ SOURCE_TABLE => 'emp'); Note 11: Asktom thread ====================== You Asked I am looking for specific example of setting up streams for bi-directional schema level replication. What are your thoughts on using Oracle Streams to implement active-active configuration of databases for high availability? Thanks, Pratap and we said... replication is for replication. replication is definitely nothing I would consider for HA. For HA there is: o RAC -- active active servers in a room. o Data Guard -- active/warm for failover in the even the room disappears. Replication is a study in complexity. Update anywhere will make your application o infinitely hard to design o fairly impossible to test o more fragile (more moving pieces, more things that can go wrong. which conflicts with your stated goal of HA) I would not consider replication for HA in any circumstance. Data Guard is the feature you are looking for. Tom, Correct me if I am wrong. My understanding is that for a Dataguard failover, manual intervention of the DBA is required. But if I have a replicated database (2 masters - synchronous from primary to replicated database and asynchronous the other way around - and only the primary being updated in normal circumstances), the failover would be automatic and does not require the DBA to be on site immediately. Thanks Anandhi Followup March 8, 2004 - 8am US/Eastern: the problem is you have to design your entire system from day 1 to be replicated since when you "failover" (lose the ability to connect to db1) there will be QUEUED transactions that have not yet taken place on db2 (eg: your users will say "hey, I know i did that already and do it all over again") when db1 "recovers" it'll push its transactions and db2 will push its transactions. bamm -- update conflicts. So, replication is a tool developers can use to build a replicated database. dataguard is a tool DBA's can use to set up a highly available environment. they are not to be confused - you cannot replicate 3rd party applications like Oracle Apps, people soft, SAP, etc. You cannot replication most custom developed applications without major design/coding efforts. you can data guard ANYTHING. and yes, when failover is to take place, you want a human deciding that. failover is something that happens in minutes, it is in response to a disaster (hence the name DR). It is a fire, an unrecoverable situation. You do not want to failover because a system blue screened (wait for it to reboot). You do not want to failover some people but not others (as would happen with db1, db2 and siteA, siteB if siteA cannot route to db1 but can route to db2 but siteB can still route to db1 - bummer, now you have transactions taking place on BOTH and unless you designed the systems to be "replicatable" you are in a hole world of hurt) DR is something you want a human to be involved in. They need to pull the trigger. Hi Tom, can you please provide a classification of streams and change data capture. I guess the main difference is that streams covers event capture, transport (transformation) and consumption. CDC only the capture. But if you consider only event capture, are there technical differences between streams and change data capture? What was the main reason to made CDC as a separate product? thx Jaromir http://www.db-nemec.com Followup March 26, 2004 - 9am US/Eastern: think of streams like a brick. think of CDC like a building made of brick. streams can be used to build CDC. CDC is built on top of streams (async CDC is anyway, sync CDC is trigger based). they are complimentary, not really competing. Hi Tom, Is Oracle advanced queuing(AQ) is renamed as Oracle Stream in 10g? Thanks Followup February 13, 2005 - 4pm US/Eastern: no, AQ is a foundation technology used in the streams implemenation (and advanced replication), but streams is not AQ. Note 12: ANYDATA: ================= This datatype could be useful in an application that stores generic attributes -- attributes you don't KNOW what the datatypes are until you actually run the code. In the past, we would have stuffed everything into a VARCHAR2 -- dates, numbers, everything. Now, you can put a date in and have it stay as a date (and the system will enforce it is in fact a valid date and let you perform date operations on it -- if it were in a varchar2 -- someone could put "hello world" into your "date" field) SQL> connect adm/vga88nt Connected. SQL> create table t ( x sys.anyData ); Table created. SQL> insert into t values ( sys.anyData.convertNumber(5) ); 1 row created. SQL> SQL> insert into t values ( sys.anyData.convertDate(sysdate) ); 1 row created. SQL> SQL> insert into t values ( sys.anyData.convertVarchar2('hello world') ); 1 row created. SQL> commit; Commit complete. SQL> select t.x.gettypeName() typeName from t t; TYPENAME -------------------------------------------------------------------------------- SYS.NUMBER SYS.DATE SYS.VARCHAR2 SQL> select * from t; X() -------------------------------------------------------------------------------- ANYDATA() ANYDATA() ANYDATA() Unfortunately, they don't have a method to display the contents of ANYDATA in a query (most useful in programs that will fetch the data, figure out what it is and do something with it -- eg: the application has some intelligence as to how to handle the data) Fortunately we can write one tho: create or replace function getData( p_x in sys.anyData ) return varchar2 2 as 3 l_num number; 4 l_date date; 5 l_varchar2 varchar2(4000); 6 begin 7 case p_x.gettypeName 8 when 'SYS.NUMBER' then 9 if ( p_x.getNumber( l_num ) = dbms_types.success ) 10 then 11 l_varchar2 := l_num; 12 end if; 13 when 'SYS.DATE' then 14 if ( p_x.getDate( l_date ) = dbms_types.success ) 15 then 16 l_varchar2 := l_date; 17 end if; 18 when 'SYS.VARCHAR2' then 19 if ( p_x.getVarchar2( l_varchar2 ) = dbms_types.success ) 20 then 21 null; 22 end if; 23 else 24 l_varchar2 := '** unknown **'; 25 end case; 26 27 return l_varchar2; 28 end; 29 / Function created. select getData( x ) getdata from t; GETDATA -------------------- 5 19-MAR-02 hello world Note 12: Materialized Views =========================== thread 1: --------- create materialized view emp_rollback enable query rewrite as select deptno, sum(sal) sal from emp group by deptno; Now, given that all the necessary settings have been done (see the data warehousing guide for a comprehensive example) your end users can query: select deptno, sum(sal) from emp where deptno in ( 10, 20) group by deptno; and the database engine will rewrite the query to go against the precomputed rollup, not the details -- giving you the answer in a fraction of the time it would normally take. CREATE MATERIALIZED VIEW LOG ON sales WITH SEQUENCE, ROWID (prod_id, cust_id, time_id, channel_id, promo_id, quantity_sold, amount_sold) INCLUDING NEW VALUES; CREATE MATERIALIZED VIEW sum_sales PARALLEL BUILD IMMEDIATE REFRESH FAST ON COMMIT AS SELECT s.prod_id, s.time_id, COUNT(*) AS count_grp, SUM(s.amount_sold) AS sum_dollar_sales, COUNT(s.amount_sold) AS count_dollar_sales, SUM(s.quantity_sold) AS sum_quantity_sales, COUNT(s.quantity_sold) AS count_quantity_sales FROM sales s GROUP BY s.prod_id, s.time_id; This example creates a materialized view that contains aggregates on a single table. Because the materialized view log has been created with all referenced columns in the materialized view's defining query, the materialized view is fast refreshable. If DML is applied against the sales table, then the changes will be reflected in the materialized view when the commit is issued. CREATE MATERIALIZED VIEW cust_sales_mv PCTFREE 0 TABLESPACE demo STORAGE (INITIAL 16k NEXT 16k PCTINCREASE 0) PARALLEL BUILD IMMEDIATE REFRESH COMPLETE ENABLE QUERY REWRITE AS SELECT c.cust_last_name, SUM(amount_sold) AS sum_amount_sold FROM customers c, sales s WHERE s.cust_id = c.cust_id GROUP BY c.cust_last_name; thread 2: --------- Use the CREATE MATERIALIZED VIEW statement to create a materialized view. A materialized view is a database object that contains the results of a query. The FROM clause of the query can name tables, views, and other materialized views. Collectively these objects are called master tables (a replication term) or detail tables (a data warehousing term). This reference uses "master tables" for consistency. The databases containing the master tables are called the master databases. Note: The keyword SNAPSHOT is supported in place of MATERIALIZED VIEW for backward compatibility. thread 3: --------- The following statement creates the primary-key materialized view on the table emp located on a remote database. SQL> CREATE MATERIALIZED VIEW mv_emp_pk REFRESH FAST START WITH SYSDATE NEXT SYSDATE + 1/48 WITH PRIMARY KEY AS SELECT * FROM emp@remote_db; Materialized view created. Note: When you create a materialized view using the FAST option you will need to create a view log on the master tables(s) as shown below: SQL> CREATE MATERIALIZED VIEW LOG ON emp; Materialized view log created. thread 4: --------- Refreshing Materialized Views When creating a materialized view, you have the option of specifying whether the refresh occurs ON DEMAND or ON COMMIT. In the case of ON COMMIT, the materialized view is changed every time a transaction commits, thus ensuring that the materialized view always contains the latest data. Alternatively, you can control the time when refresh of the materialized views occurs by specifying ON DEMAND. In this case, the materialized view can only be refreshed by calling one of the procedures in the DBMS_MVIEW package. DBMS_MVIEW provides three different types of refresh operations. DBMS_MVIEW.REFRESH Refresh one or more materialized views. DBMS_MVIEW.REFRESH_ALL_MVIEWS Refresh all materialized views. DBMS_MVIEW.REFRESH_DEPENDENT Refresh all materialized views that depend on a specified master table or materialized view or list of master tables or materialized views. Note 10: ======== /********************************************************************************* * @author : Chandar * @version : 1.0 * * Name of the Application : SetupStreams.sql * Creation/Modification History : * * Chandar 02-Feb-2003 Created * * Overview of Script: * This SQL scripts sets up the streams for bi-directional replication between two * databases. Replication is set up for the table named tabone in strmuser schema * created by the script in both the databases. * Ensure that you have created a streams administrator before executing this script. * The script StreamsAdminConfig.sql can be used to create a streams administrator * and configure it. * After running this script you can use AddTable.sql script to add another active * table to streams environment. ***********************************************************************************/ SET VERIFY OFF SET ECHO OFF SPOOL streams_setup.log --define variables to store global names of two databases variable site1 varchar2(128); variable site2 varchar2(128); variable scn number; ----------------------------------------------------------------------------------- -- get TNSNAME , SYS password and streams admin user details for both the databases ----------------------------------------------------------------------------------- PROMPT -- TNSNAME for database 1 ACCEPT db1 PROMPT 'Enter TNS Name of first database :' PROMPT -- SYS password for database 1 ACCEPT syspwddb1 PROMPT 'Enter password for sys user of first database :' PROMPT -- Streams administrator username for database 1 ACCEPT strm_adm_db1 PROMPT 'Enter username for streams admin of first database :' PROMPT -- Streams administrator password for database 1 ACCEPT strm_adm_pwd_db1 PROMPT 'Enter password for streams admin on first database :' PROMPT -- TNSNAME for database 2 ACCEPT db2 PROMPT 'Enter TNS Name of second database :' PROMPT -- SYS password for database 2 ACCEPT syspwddb2 PROMPT 'Enter password for sys user of second database :' PROMPT -- Streams administrator username for database 2 ACCEPT strm_adm_db2 PROMPT 'Enter username for streams admin of second database :' PROMPT -- Streams administrator password for database 2 ACCEPT strm_adm_pwd_db2 PROMPT 'Enter password for streams admin on second database :' PROMPT PROMPT Connecting as SYS user to database 1 CONN sys/&syspwddb1@&db1 AS SYSDBA; -- Store global name in site1 variable EXECUTE SELECT global_name INTO :site1 FROM global_name; PROMPT Granting execute privileges on dbms_lock and dbms_pipe to streams admin GRANT EXECUTE ON DBMS_LOCK TO &strm_adm_db1; GRANT EXECUTE ON DBMS_PIPE to &strm_adm_db1; -- create a user name strmuser and grant necessary privileges PROMPT Creating user named strmuser GRANT CONNECT, RESOURCE TO strmuser IDENTIFIED BY strmuser; PROMPT Connecting as strmuser to database1 CONN strmuser/strmuser@&db1 -- create a sample table named tabone for which the replication will be set up PROMPT PROMPT Creating table tabone CREATE TABLE tabone (id NUMBER(5) PRIMARY KEY, name VARCHAR2(50)); -- grant all permissions on tabone to stream administration PROMPT Adding supplemetal logging for table tabone ALTER TABLE tabone ADD SUPPLEMENTAL LOG GROUP tabone_log_group ( id,name) ALWAYS; PROMPT Granting permissions on table tabone to streams administration GRANT ALL ON strmuser.tabone TO &strm_adm_db1; ------------------------------------ -- Repeat above steps for database 2 ------------------------------------ PROMPT Connecting as SYS user to database2 CONN sys/&syspwddb2@&db2 AS SYSDBA; -- Store global name in site2 variable EXECUTE SELECT global_name INTO :site2 FROM global_name; PROMPT Granting execute privileges on dbms_lock and dbms_pipe to streams admin GRANT EXECUTE ON DBMS_LOCK TO &strm_adm_db2; GRANT EXECUTE ON DBMS_PIPE to &strm_adm_db2; -- create a user name strmuser and grant necessary privileges PROMPT Creating user named strmuser GRANT CONNECT, RESOURCE TO strmuser IDENTIFIED BY strmuser; PROMPT Connecting as strmuser CONN strmuser/strmuser@&db2 -- create a sample table named tabone for which the replication will be set up PROMPT PROMPT Creating table tabone CREATE TABLE tabone (id NUMBER(5) PRIMARY KEY, name VARCHAR2(50)); PROMPT Adding supplemetal logging for table tabone ALTER TABLE tabone ADD SUPPLEMENTAL LOG GROUP tabone_log_group ( id,name) ALWAYS; -- grant all permissions on tabone to stream administration PROMPT Granting all permissions on tabone to streams administrator GRANT ALL ON strmuser.tabone TO &strm_adm_db2; ---------------------------------------------------------------------------------- -- Set up replication for table tabone from database 1 to database 2 using streams ---------------------------------------------------------------------------------- -- connect as streams admin to database 1 PROMPT Connecting as streams adimistrator to database 1 conn &strm_adm_db1/&strm_adm_pwd_db1@&db1 -- create and set up streams queue at database 1 PROMPT PROMPT Creating streams queue BEGIN DBMS_STREAMS_ADM.SET_UP_QUEUE( queue_table => 'strmuser_queue_table', queue_name => 'strmuser_queue', queue_user => 'strmuser'); END; / -- Add table propagation rules for table tabone to propagate captured changes -- from database 1 to database 2 PROMPT Adding propagation rules for table tabone BEGIN DBMS_STREAMS_ADM.ADD_TABLE_PROPAGATION_RULES( table_name => 'strmuser.tabone', streams_name => 'db1_to_db2_prop', source_queue_name => '&strm_adm_db1..strmuser_queue', destination_queue_name => '&strm_adm_db2..strmuser_queue@'||:site2, include_dml => true, include_ddl => true, source_database => :site1); END; / -- create a capture process and add table rules for table tabone to capture the -- changes made to tabone in database 1 PROMPT Creating capture process at database 1 and adding table rules for table tabone. BEGIN DBMS_STREAMS_ADM.ADD_TABLE_RULES( table_name => 'strmuser.tabone', streams_type => 'capture', streams_name => 'capture_db1', queue_name => '&strm_adm_db1..strmuser_queue', include_dml => true, include_ddl => true); END; / -- create a database link to database 2 connecting as streams administrator PROMPT Creating database link to database 2 DECLARE sql_command VARCHAR2(200); BEGIN sql_command :='CREATE DATABASE LINK ' ||:site2|| ' CONNECT TO'|| '&strm_adm_db2 IDENTIFIED BY &strm_adm_pwd_db2 USING ''&db2'''; EXECUTE IMMEDIATE sql_command; END; / -- get the current SCN of database 1 PROMPT Getting current SCN of database 1 EXECUTE :scn := DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER(); -- connect to database 2 as streams administrator PROMPT Connecting as streams administrator to database 2 conn &strm_adm_db2/&strm_adm_pwd_db2@&db2 -- Set table instantiation SCN for table tabone at database 2 to current -- SCN of database 1 -- We need not use import/export for instantiation because table tabone -- does not contain any data PROMPT PROMPT Setting instantiation SCN for table tabone at database 2 BEGIN DBMS_APPLY_ADM.SET_TABLE_INSTANTIATION_SCN(source_object_name => 'strmuser.tabone', source_database_name => :site1, instantiation_scn => :scn); END; / -- create and set up streams queue at database 2 PROMPT Setting up streams queue at database 2 BEGIN DBMS_STREAMS_ADM.SET_UP_QUEUE( queue_table => 'strmuser_queue_table', queue_name => 'strmuser_queue', queue_user => 'strmuser'); END; / -- create an apply process and add table rules for table tabone to apply -- any changes propagated from database 1 PROMPT Creating Apply process at database 2 and adding table rules for table tabone BEGIN DBMS_STREAMS_ADM.ADD_TABLE_RULES( table_name => 'strmuser.tabone', streams_type => 'apply', streams_name => 'apply_db2', queue_name => '&strm_adm_db2..strmuser_queue', include_dml => true, include_ddl => true, source_database => :site1); END; / -- start the apply process at database 2 PROMPT Starting the apply process BEGIN DBMS_APPLY_ADM.START_APPLY( apply_name => 'apply_db2'); END; / -- connect to database 1 as streams administrator PROMPT Connecting as streams administrator to database 1 conn &strm_adm_db1/&strm_adm_pwd_db1@&db1 -- start the capture process PROMPT PROMPT Starting the capture process at database 1 BEGIN DBMS_CAPTURE_ADM.START_CAPTURE( capture_name => 'capture_db1'); END; / -- make dml changes to tabone to check if streams is working PROMPT Inserting row in tabone at database 1 INSERT INTO strmuser.tabone VALUES(11,'chan'); COMMIT; -- wait for some time so that changes are applied to database 2. EXECUTE DBMS_LOCK.SLEEP(35); ----------------------------------------------------------------------------------- -- Set up replication for table tabone from database 2 to database 1 using streams ----------------------------------------------------------------------------------- -- connect to database 2 as streams administrator PROMPT Connecting as streams administrator to database 2 conn &strm_adm_db2/&strm_adm_pwd_db2@&db2 -- select table tabone to see if changes from database 1 are applied PROMPT PROMPT Selecting rows from tabone at database 2 to see if changes are propagated select * from strmuser.tabone; PROMPT PROMPT Setting up bi-directional replication of table tabone -- create a database link to database 1 connecting as streams administrator PROMPT PROMPT Creating database link from database 2 to database 1 DECLARE sql_command varchar2(200); BEGIN sql_command :='CREATE DATABASE LINK ' ||:site1|| ' CONNECT TO'|| '&strm_adm_db1 IDENTIFIED BY &strm_adm_pwd_db1 USING ''&db1'''; EXECUTE IMMEDIATE sql_command; END; / -- Add table propagation rules for table tabone to propagate capture changes -- from database 2 to database 1 PROMPT Adding table propagation rules for tabone at database 2 BEGIN DBMS_STREAMS_ADM.ADD_TABLE_PROPAGATION_RULES( table_name =>'strmuser.tabone', streams_name => 'db2_to_db1_prop', source_queue_name => '&strm_adm_db2..strmuser_queue', destination_queue_name => '&strm_adm_db1..strmuser_queue@'||:site1, include_dml => true, include_ddl => true, source_database => :site2); END; / -- create a capture process and add table rules for table tabone to -- capture the changes made to tabone in database 2 PROMPT Creating capture process at database 2 and adding table rules for table tabone BEGIN DBMS_STREAMS_ADM.ADD_TABLE_RULES( table_name => 'strmuser.tabone', streams_type => 'capture', streams_name => 'capture_db2', queue_name => '&strm_adm_db2..strmuser_queue', include_dml => true, include_ddl => true); END; / -- get the current SCN of database 2 PROMPT Getting the current SCN of database 2 EXECUTE :scn := DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER(); -- connect to database 1 as streams administrator PROMPT Connecting as streams administrator to database 1 CONN &strm_adm_db1/&strm_adm_pwd_db1@&db1 -- Set table instantiation SCN for table tabone at database 1 to current -- SCN of database 2 PROMPT PROMPT Setting instantiation SCN for tabone at database 2 BEGIN DBMS_APPLY_ADM.SET_TABLE_INSTANTIATION_SCN(source_object_name => 'strmuser.tabone', source_database_name => :site2, instantiation_scn => :scn); END; / -- create an apply process and add table rules for table tabone to apply -- any changes propagated from database 2 PROMPT Creating apply process at database 1 and adding table rules for tabone BEGIN DBMS_STREAMS_ADM.ADD_TABLE_RULES( table_name => 'strmuser.tabone', streams_type => 'apply', streams_name => 'apply_db1', queue_name => '&strm_adm_db1..strmuser_queue', include_dml => true, include_ddl => true, source_database => :site2); END; / -- start the apply process PROMPT Starting the apply process BEGIN DBMS_APPLY_ADM.START_APPLY( apply_name => 'apply_db1'); END; / -- connect to database 2 as streams administrator PROMPT Connecting to database 2 as streams administrator CONN &strm_adm_db2/&strm_adm_pwd_db2@&db2; -- start the capture process PROMPT PROMPT Starting the capture process at database 2 BEGIN DBMS_CAPTURE_ADM.START_CAPTURE( capture_name => 'capture_db2'); END; / -- perform dml on tabone at database 2 to check if changes are propagated PROMPT Inserting a row into tabone at database 2 INSERT INTO strmuser.tabone VALUES(12,'kelvin'); COMMIT; -- wait for some time so that changes are applied to database 1. EXECUTE DBMS_LOCK.SLEEP(35); -- connect to database 1 as streams administrator PROMPT Connecting as streams administrator to database 1 CONN &strm_adm_db1/&strm_adm_pwd_db1@&db1 PROMPT Checking if the changes made at database 2 are applied at database 1 SELECT * FROM strmuser.tabone; SET ECHO OFF SPOOL OFF PROMPT End of Script Note 11: ======== The Data Propagator replication engine uses a change-capture mechanism and a log to replicate data between a source system and a target system. A capture process running on the source system captures changes as they occur in the source tables and stores them temporarily in the change data tables. The database administrator of the source database must ensure that the Data Propagator capture process is active on the source system. The apply process reads the change data tables and applies the changes to the target tables. You can create a Data Propagator subscription with the DB2 Everyplace XML Scripting tool. In DB2 Everyplace Version 8.2, you cannot create or configure Data Propagator subscriptions by using the DB2® Everyplace® Sync Server Administration Console. You can use the DB2 Everyplace Sync Server Administration Console only to view and assign Data Propagator subscriptions to subscription sets. All the Data Propagator subscriptions use the Data Propagator replication engine. Each Data Propagator replication environment consists of a source system and a mirror system. The source system contains the source database, the tables that you want to replicate, and the capture process that is used to capture the data changes. The mirror system contains the mirror database and tables. DB2 Everyplace starts the apply process on the mirror system. When capturing changes to the change data tables, the capture process that is running on both the source system and the mirror system will consume processor resources and input/output resources. As a result of this additional load on the source system, replication competes with the source applications for system resources. Additionally, with the Data Propagator engine, the number of moves that the changed data has to make between the tables in the mirror system is higher than with the JDBC engine. As a result, the mirror database requires a substantially larger logging space than the JDBC replication engine. Capacity planners should balance the needs of the replication tasks and source application to determine the size of the source system accordingly. -How the Data Propagator replication engine handles data changes to the source system When the source application changes a table in the source system, the changes are synchronized Data Propagator replication engine first captures the changes, synchronizes them to the mirror system, and then applies them to the target system (the mobile device). -How the Data Propagator replication engine handles data changes to the client sys9tem When the client application on the mobile device changes a table in the client system, the Data Propagator replication engine first synchronizes the changes to the mirror system, captures them, and then applies them to the source system. Note 12: commitscn ================== The COMMIT SCN - an undocumented feature May 1999 -------------------------------------------------------------------------------- Try the following experiment: create table t (n1 number); insert into t values (userenv('commitscn')); select n1 from t; N1 -------- 43526438 rem Wait a few seconds if there are other people working rem on your system, or start a second session execute a rem couple of small (but real) transactions and commits then commit; select n1 from t; N1 -------- 43526441 Obviously your values for N1 will not match the values above, but you should see that somehow the data you inserted into your table was not the value that was finally committed, so what's going on ? The userenv('commitscn') function has to be one of the most quirky little undocumented features of Oracle. You can only use it in a very restricted fashion, but if you follow the rules the value that hits the database is the current value of the SCN (System Commit Number), but when you commit your transaction the number changes to the latest value of the SCN which is always just one less than the commit SCN used by your transaction. Why on earth, you say, would Oracle produce such a wierd function - and how on earth do they stop it from costing a fortune in processing time. To answer the first question think replication. Back to the days of 7.0.9, when a client asked me to build a system which used asynchronous replication between London and New York; eventually I persuaded him this was not a good idea, especially on early release software when the cost to the business of an error would be around $250,000 per shot; nevertheless I did have to demonstrate that in principal it was possible. The biggest problem, though, was guaranteeing that transactions were applied at the remote site in exactly the same order that they had been committed at the local site; and this is precisely where Oracle uses userenv('commitscn'). Each time a commit hits the database, the SCN is incremented, so each transaction is 'owned' by an SCN and no two transactions can belong to a single SCN - ultimately the SCN generator is the single-thread through which all the database must pass and be serialised. Although there is a small arithmetical quirk that the value of the userenv('commitscn') is changed to one less than the actual SCN used to commit the transaction, nevertheless each transaction gets a unique, correctly ordered value for the function. If you have two transactions, the one with the lower value of userenv('commitscn') is guaranteeably the one that committed first. So how does Oracle ensure that the cost of using this function is not prohbitive. Well you need to examine Oracle errors 1735 and 1721 in the $ORACLE_HOME/rdbms/admin/mesg/oraus.msg file. ORA-01721: USERENV(COMMITSCN) invoked more than once in a transaction ORA-01735: "USERENV('COMMITSCN') not allowed here You may only use userenv('commitscn') to update exactly one column of one row in a transaction, or insert exactly one value for one row in a transaction, and (just to add that final touch of peculiarity) the column type has to be an unconstrained number type otherwise the subsequent change does not take place. -------------------------------------------------------------------------------- Build Your Own Replication: Given this strange function, here's the basis of what you have to do to write your own replication code: create table control_table(sequence_id number, commit_id number); begin transaction insert into control_table (sequence_id,commit_id) select meaningless_sequence.nextval, null from dual; -- save the value of meaningless_sequence -- left as a language-specific exercise update control_table set commit_id = userenv('commitscn') where sequence_id = {saved value of meaningless_sequence}; -- now do all the rest of the work, and include the saved -- meaningless_sequence.currval in every row of every table commit; end transaction If you now transport the changed data to the remote site, using the commit_id to send the transactions in the correct order, and the sequence_id to find the correct items of data, most of your problems are over. (Although you still have some messy details which are again left as an exercise.) Note 13: ======== Oracle stream not working as Logminer is down Posted: Dec 17, 2007 11:58 PM Reply Hi, Oracle streams capture process is not capturing any updates made on table for which capture & apply process are configured. Capture process & apply process are running fine showing enabled as status & no error. But, No new records are captured in ‘streams_queue_table’ when I update record in table, which is configured for capturing changes. This setup was working till I got ‘ORA-01341: LogMiner out-of-memory’ error in alert.log file. I guess logminer is not capturing the updates from redo log. Current Alert log is showing following lines for logminer init process LOGMINER: Parameters summary for session# = 1 LOGMINER: Number of processes = 3, Transaction Chunk Size = 1 LOGMINER: Memory Size = 10M, Checkpoint interval = 10M But same log was like this before LOGMINER: Parameters summary for session# = 1 LOGMINER: Number of processes = 3, Transaction Chunk Size = 1 LOGMINER: Memory Size = 10M, Checkpoint interval = 10M >>LOGMINER: session# = 1, reader process P002 started with pid=18 OS id=5812 >>LOGMINER: session# = 1, builder process P003 started with pid=36 OS id=3304 >>LOGMINER: session# = 1, preparer process P004 started with pid=37 OS id=1496 We can clearly see reader, builder & preparer process are not starting after I got Out of memory exception in log miner. To allocate more space to logminer, I tried to setup tablespace to logminer I got 2 exception which was contradicting each other error. SQL> exec DBMS_LOGMNR.END_LOGMNR(); BEGIN DBMS_LOGMNR.END_LOGMNR(); END; * ERROR at line 1: >>ORA-01307: no LogMiner session is currently active ORA-06512: at "SYS.DBMS_LOGMNR", line 76 ORA-06512: at line 1 SQL> EXECUTE DBMS_LOGMNR_D.SET_TABLESPACE('logmnrts'); BEGIN DBMS_LOGMNR_D.SET_TABLESPACE('logmnrts'); END; * ERROR at line 1: >>ORA-01356: active logminer sessions found ORA-06512: at "SYS.DBMS_LOGMNR_D", line 232 ORA-06512: at line 1 When I tried stopping logminer exception was ‘no logminer session is active’, But when I tried to setup tablespace exception was ‘active logminer sessions found’. I am not sure how to resolve this issue. Please let me know how to resolve this issue. Thanks Posts: 25 From: Brazil Registered: 6/19/06 Re: Oracle stream not working as Logminer is down Posted: Dec 19, 2007 3:34 AM in response to: sgurusam Reply The Logminer session associated with a capture process is a special kind of session which is called a "persistent session". You will not be able to stop it using DBMS_LOGMNR. This package controls only non-persistent sessions. To stop the persistent LogMiner session you must stop the capture process. However, I think your problem is more related to a lack of RAM space instead of tablespace (i. e, disk) space. Try to increase the size of the SGA allocated to LogMiner, by setting capture parameter _SGA_SIZE. I can see you are using the default of 10M, which may be not enough for your case. Of course, you will have to increase the values of init parameters streams_pool_size, sga_target/sga_max_size accordingly, to avoid other memory problems. To set the _SGA_SIZE parameter, use the PL/SQL procedure DBMS_CAPTURE_ADM.SET_PARAMETER. The example below would set it to 100Megs: begin DBMS_CAPTURE_ADM.set_parameter(' I have set up asyncronous hotlog change data capture from a 9iR2 mainframe > oracle database to an AIX 10gR2 database. The mainframe process didn't work > and put the capture into Abended status. > > *** SESSION ID:(21.175) 2006-08-01 17:28:51.777 > error 10010 in STREAMS process > ORA-10010: Begin Transaction > ORA-00308: cannot open archived log '//'EDB.RL11.ARCHLOG.T1.S5965.DBF'' > ORA-04101: OS/390 implementation layer error > ORA-04101, FUNC=LOCATE , RC=8, RS=C5C7002A, ERRORID=1158 > ::OPIRIP: Uncaught error 447. Error stack::ORA-00447: fatal error in > background > A-00308: cannot open archived log '//'EDB.RL11.ARCHLOG.T1.S5965.DBF'' > ORA-04101: OS/390 implementation layer error > ORA-04101, FUNC=LOCATE , RC=8, RS=C5C7002A, ERRORID=1158 > > This is because I had lower case characters in the log file format in the > init.ora on the mainframe. The actual log file that was created was a > completely different name. > > I shut down the database and fixed the init.ora. Switched the log file. I > dropped all the objects that I created for CDC. I recreated the capture and > altered the start scn of the capture to the current log which I found by > running: select to_char(max(first_change#)) from v$log; > > I created the other objects, but when I run > dbms_cdc_publish.alter_hotlog_change_source to enable, it immediately changes > the capture from disabled to abended status, and gives me the same error > message as above. > > How do I get the capture out of abended status, and how do I get it to NOT > try to find the old archive log file (which isn't there anyways)? > Any help would be greatly appreciated! ============================================================================================== Note 15: Async CDC extended TEST: ============================================================================================== Purpose: Test Async CDC Hotlog and solve errors 1. long running txn detected 2. stop of capture Date : 26/02/2008 DB : 10.2.0.3 ------------------------------------------------------------------------ SOURCE TABLE OWNER: ALBERT SOURCE TABLE : PERSOON PUBLISHER : publ_cdc CDC_SET : CDC_DEMO_SET SUBSCRIBER : subs_cdc CHANGE TABLE : CDC_PERSOON CHANGE_SOURCE : SYNC_SOURCE ------------------------------------------------------------------------ set ORACLE_HOME=C:\ora10g\product\10.2.0\db_1 Init: -- specific: TEST10G: startup mount pfile=c:\oracle\admin\test10g\pfile\init.ora TEST10G2: startup mount pfile=c:\oracle\admin\test10g2\pfile\init.ora -- common: alter database archivelog; archive log start; alter database force logging; alter database add supplemental log data; alter database open; archive log list -- Archive Mode show parameter aq_tm_processes -- min 3 show parameter compatible -- must be 10.1.0 or above show parameter global_names -- must be TRUE show parameter job_queue_processes -- min 2 recommended 4-6 show parameter open_links -- not less than the default 4 show parameter shared_pool_size -- must be 0 or at least 200MB show parameter streams_pool_size -- min. 480MB (10MB/capture 1MB/apply) show parameter undo_retention -- min. 3600 (1 hr.) (900) -- Examples of altering initialization parameters alter system set aq_tm_processes=3 scope=BOTH; alter system set compatible='10.2.0.1.0' scope=SPFILE; alter system set global_names=TRUE scope=BOTH; alter system set job_queue_processes=6 scope=BOTH; alter system set open_links=4 scope=SPFILE; alter system set streams_pool_size=200M scope=BOTH; -- very slow if making smaller alter system set undo_retention=3600 scope=BOTH; /* JOB_QUEUE_PROCESSES (current value) + 2 PARALLEL_MAX_SERVERS (current value) + (5 * (the number of change sets planned)) PROCESSES (current value) + (7 * (the number of change sets planned)) SESSIONS (current value) + (2 * (the number of change sets planned)) */ ------------------------------------------------------------------------ Admin Queries: connect / as sysdba select * FROM DBA_SOURCE_TABLES; SELECT SET_NAME,CHANGE_SOURCE_NAME,BEGIN_SCN,END_SCN,CAPTURE_ENABLED,PURGING,QUEUE_NAME FROM CHANGE_SETS; SELECT OWNER, QUEUE_TABLE, TYPE, OBJECT_TYPE, RECIPIENTS FROM DBA_QUEUE_TABLES; SELECT SET_NAME,STATUS,EARLIEST_SCN,LATEST_SCN,to_char(LAST_PURGED, 'DD-MM-YYYY;HH24:MI'), to_char(LAST_EXTENDED, 'DD-MM-YYYY;HH24:MI'),SUBSCRIPTION_NAME FROM DBA_SUBSCRIPTIONS; SELECT PROPAGATION_SOURCE_NAME, PROPAGATION_NAME, STAGING_DATABASE, DESTINATION_QUEUE FROM CHANGE_PROPAGATIONS; SELECT tablespace_name, force_logging FROM dba_tablespaces; SELECT supplemental_log_data_min, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all, force_logging FROM gv$database; SELECT owner, name, QUEUE_TABLE, ENQUEUE_ENABLED, DEQUEUE_ENABLED FROM dba_queues; SELECT capture_name, total_messages_captured, total_messages_enqueued, elapsed_enqueue_time FROM dba_hist_streams_capture; SELECT apply_name, reader_total_messages_dequeued, reader_lag, server_total_messages_applied FROM dba_hist_streams_apply_sum; SELECT table_name, scn, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all FROM dba_capture_prepared_tables; SELECT table_name, scn, supplemental_log_data_pk, supplemental_log_data_ui, supplemental_log_data_fk, supplemental_log_data_all FROM dba_capture_prepared_tables; SELECT DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() from dual; SELECT EQ_NAME,EQ_TYPE,TOTAL_WAIT#,FAILED_REQ#,CUM_WAIT_TIME,REQ_DESCRIPTION FROM V_$ENQUEUE_STATISTICS WHERE CUM_WAIT_TIME>0 ; SELECT set_name,capture_name,queue_name,queue_table_name,capture_enabled FROM cdc_change_sets$; SELECT set_name,capture_name,capture_enabled FROM cdc_change_sets$; SELECT set_name, CAPTURE_ENABLED, BEGIN_SCN, END_SCN,LOWEST_SCN,CAPTURE_ERROR FROM cdc_change_sets$; SELECT set_name, change_source_name, capture_enabled, stop_on_ddl, publisher FROM change_sets; SELECT subscription_name, handle, set_name, username, earliest_scn, description FROM cdc_subscribers$; SELECT username FROM dba_users u, streams$_privileged_user s WHERE u.user_id = s.user#; SELECT cap.CAPTURE_NAME, cap.FIRST_SCN, cap.APPLIED_SCN, cap.REQUIRED_CHECKPOINT_SCN FROM DBA_CAPTURE cap, CHANGE_SETS cset WHERE cset.SET_NAME = 'CDC_DEMO_SET' AND cap.CAPTURE_NAME = cset.CAPTURE_NAME; SELECT r.SOURCE_DATABASE,r.SEQUENCE#,r.NAME,r.DICTIONARY_BEGIN,r.DICTIONARY_END FROM DBA_REGISTERED_ARCHIVED_LOG r, DBA_CAPTURE c WHERE c.CAPTURE_NAME = 'CDC$C_CHANGE_SET_ALBERT' AND r.CONSUMER_NAME = c.CAPTURE_NAME; SELECT CONSUMER_NAME,PURGEABLE,THREAD#, FIRST_SCN,NEXT_SCN, SEQUENCE# FROM DBA_REGISTERED_ARCHIVED_LOG ------------------------------------------------------------------------ >>>>>>>>>>> connect / as sysdba Initial: -- TS CREATE TABLESPACE TS_CDC DATAFILE 'C:\ORACLE\ORADATA\TEST10G\TS_CDC.DBF' SIZE 50M EXTENT MANAGEMENT LOCAL AUTOALLOCATE SEGMENT SPACE MANAGEMENT AUTO LOGGING FORCE LOGGING; - USERS: create user albert identified by albert default tablespace ts_cdc temporary tablespace temp QUOTA 10M ON sysaux QUOTA 20M ON users QUOTA 50M ON ts_cdc ; create user publ_cdc identified by publ_cdc default tablespace ts_cdc temporary tablespace temp QUOTA 10M ON sysaux QUOTA 20M ON users QUOTA 50M ON TS_CDC ; create user subs_cdc identified by subs_cdc default tablespace ts_cdc temporary tablespace temp QUOTA 10M ON sysaux QUOTA 20M ON users QUOTA 50M ON TS_CDC ; -- GRANTS: GRANT create session TO albert; GRANT create table TO albert; GRANT create sequence TO albert; GRANT create procedure TO albert; GRANT connect TO albert; GRANT resource TO albert; GRANT create session TO publ_cdc; GRANT create table TO publ_cdc; GRANT create sequence TO publ_cdc; GRANT create procedure TO publ_cdc; GRANT connect TO publ_cdc; GRANT resource TO publ_cdc; GRANT dba TO publ_cdc; GRANT create session TO subs_cdc; GRANT create table TO subs_cdc; GRANT create sequence TO subs_cdc; GRANT create procedure TO subs_cdc; GRANT connect TO subs_cdc; GRANT resource TO subs_cdc; GRANT dba TO subs_cdc; GRANT execute_catalog_role TO publ_cdc; GRANT select_catalog_role TO publ_cdc; GRANT execute_catalog_role TO subs_cdc; GRANT select_catalog_role TO subs_cdc; -- object privileges GRANT execute ON dbms_cdc_publish TO publ_cdc; GRANT execute ON dbms_cdc_subscribe TO publ_cdc; GRANT execute ON dbms_lock TO publ_cdc; GRANT execute ON dbms_cdc_publish TO subs_cdc; GRANT execute ON dbms_cdc_subscribe TO subs_cdc; GRANT execute ON dbms_lock TO subs_cdc; execute dbms_streams_auth.grant_admin_privilege('publ_cdc'); SQL> SELECT * 2 FROM dba_streams_administrator; USERNAME LOC ACC ------------------------------ --- --- publ_cdc YES YES SQL> desc dba_streams_administrator; SQL> SELECT username 2 FROM dba_users u, streams$_privileged_user s 3 WHERE u.user_id = s.user#; USERNAME ------------------------------ publ_cdc ------------------------------------------------------------------------ CDC: ==== -- CREATE CHANGE_SET >>>>>>>>>>> connect albert/albert create table persoon ( userid number, name varchar(30), lastname varchar(30), constraint pk_userid primary key (userid) ); GRANT SELECT ON PERSOON TO publ_cdc; GRANT SELECT ON PERSOON TO subs_cdc; ALTER TABLE persoon ADD SUPPLEMENTAL LOG DATA (ALL) COLUMNS; >>>>>>>>>>> connect / as sysdba SQL> SELECT DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() from dual; DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() ----------------------------------------- 608789 exec dbms_capture_adm.prepare_table_instantiation(table_name => ALBERT.PERSOON); SQL> SELECT table_name, scn, supplemental_log_data_pk, supplemental_log_data_ui, 2 supplemental_log_data_fk, supplemental_log_data_all 3 FROM dba_capture_prepared_tables; TABLE_NAME SCN SUPPLEME SUPPLEME SUPPLEME SUPPLEME ------------------------------ ---------- -------- -------- -------- -------- PERSOON 608809 IMPLICIT IMPLICIT IMPLICIT EXPLICIT SQL> SELECT DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() from dual; DBMS_FLASHBACK.GET_SYSTEM_CHANGE_NUMBER() ----------------------------------------- 608879 >>>>>>>>>>> connect publ_cdc/publ_cdc (done as publisher!!) exec dbms_cdc_publish.create_change_set('CDC_DEMO_SET', 'CDC Demo 2 Change Set', 'HOTLOG_SOURCE', 'Y', NULL, NULL); Note the 'HOTLOG_SOURCE' !! SQL> exec dbms_cdc_publish.create_change_set('CDC_DEMO_SET', 'CDC Demo 2 Change Set', 'HOTLOG_SOURCE ', 'Y', NULL, NULL); PL/SQL procedure successfully completed. Note: if you need to drop a change set, use: DBMS_CDC_PUBLISH.DROP_CHANGE_SET(change_set_name IN VARCHAR2); >>>>>>>>>>> conn / as sysdba SQL> SELECT set_name, capture_name, queue_name, queue_table_name 2 FROM cdc_change_sets$; SET_NAME CAPTURE_NAME QUEUE_NAME QUEUE_TABLE_NAME ------------------------------ ------------------------------ ------------------------------ ------- SYNC_SET CDC_DEMO_SET CDC$C_CDC_DEMO_SET CDC$Q_CDC_DEMO_SET CDC$T_CDC_DEMO_SET SQL> SQL> SELECT set_name, CAPTURE_ENABLED, BEGIN_SCN, END_SCN,LOWEST_SCN,CAPTURE_ERROR 2 FROM cdc_change_sets$; SET_NAME C BEGIN_SCN END_SCN LOWEST_SCN C ------------------------------ - ---------- ---------- ---------- - SYNC_SET Y 0 N CDC_DEMO_SET N 0 N SQL> SELECT set_name, CAPTURE_ENABLED, BEGIN_SCN, END_SCN,LOWEST_SCN,CAPTURE_ERROR 2 FROM cdc_change_sets$; SET_NAME C BEGIN_SCN END_SCN LOWEST_SCN C ------------------------------ - ---------- ---------- ---------- - SYNC_SET Y 0 N CDC_DEMO_SET N 0 N SQL> SELECT set_name, change_source_name, capture_enabled, stop_on_ddl, publisher 2 FROM change_sets; SET_NAME CHANGE_SOURCE_NAME C S PUBLISHER ------------------------------ ------------------------------ - - ------------------------------ SYNC_SET SYNC_SOURCE Y N CDC_DEMO_SET HOTLOG_SOURCE N Y SYS SQL> SQL> SELECT subscription_name, handle, set_name, username, earliest_scn, description 2 FROM cdc_subscribers$; no rows selected -- CREATE CHANGE TABLE: >>>>>>>>>>> conn publ_cdc/publ_cdc BEGIN dbms_cdc_publish.create_change_table('publ_cdc', 'CDC_PERSOON', 'CDC_DEMO_SET', 'ALBERT', 'PERSOON', 'userid number, name varchar(30), lastname varchar(30)', 'BOTH', 'Y', 'Y', 'Y', 'Y', 'N', 'N', 'Y', 'TABLESPACE TS_CDC'); END; / The publisher can use this procedure for asynchronous and synchronous Change Data Capture. However, the default values for the following parameters are the only supported values for synchronous change sets: begin_date, end_date, and stop_on_ddl. SQL> BEGIN 2 dbms_cdc_publish.create_change_table('publ_cdc', 'CDC_PERSOON', 'CDC_DEMO_SET', 3 'ALBERT', 'PERSOON', 'userid number, name varchar(30), lastname varchar(30)', 4 'BOTH', 'Y', 'Y', 'Y', 'Y', 'N', 'N', 'Y', 'TABLESPACE TS_CDC'); 5 END; 6 / PL/SQL procedure successfully completed. GRANT select ON CDC_PERSOON TO subs_cdc; Note: To drop a change table use: -- drop the change table DBMS_CDC_PUBLISH.DROP_CHANGE_TABLE( owner IN VARCHAR2, change_table_name IN VARCHAR2, force_flag IN CHAR); exec dbms_cdc_publish.drop_change_table('publ_cdc','CDC_PERSOON','Y'); >>>>>>>>>>> connect / as sysdba SQL> SELECT change_set_name, source_schema_name, source_table_name 2 FROM cdc_change_tables$; CHANGE_SET_NAME SOURCE_SCHEMA_NAME SOURCE_TABLE_NAME ------------------------------ ------------------------------ ------------------------------ CDC_DEMO_SET ALBERT PERSOON SQL> SELECT set_name,capture_name,capture_enabled 2 FROM cdc_change_sets$; SET_NAME CAPTURE_NAME C ------------------------------ ------------------------------ - SYNC_SET Y CDC_DEMO_SET CDC$C_CDC_DEMO_SET N >>>>>>>>>>> connect publ_cdc/publ_cdc exec dbms_cdc_publish.alter_change_set(change_set_name=>'CDC_DEMO_SET', enable_capture=> 'Y'); SQL> exec dbms_cdc_publish.alter_change_set(change_set_name=>'CDC_DEMO_SET', enable_capture=> 'Y'); PL/SQL procedure successfully completed. >>>>>>>>>>> connect / as sysdba SQL> SELECT set_name,capture_name,capture_enabled 2 FROM cdc_change_sets$; SET_NAME CAPTURE_NAME C ------------------------------ ------------------------------ - SYNC_SET Y CDC_DEMO_SET CDC$C_CDC_DEMO_SET Y SQL> SELECT owner, name, QUEUE_TABLE, ENQUEUE_ENABLED, DEQUEUE_ENABLED 2 FROM dba_queues; OWNER NAME QUEUE_TABLE ENQUEUE DEQUEUE ------------------------------ ------------------------------ ------------------------------ ------- SYS CDC$Q_CDC_DEMO_SET CDC$T_CDC_DEMO_SET YES YES SYS AQ$_CDC$T_CDC_DEMO_SET_E CDC$T_CDC_DEMO_SET NO NO ........ ........ SQL> SELECT OWNER, QUEUE_TABLE, TYPE, OBJECT_TYPE, RECIPIENTS 2 FROM DBA_QUEUE_TABLES; OWNER QUEUE_TABLE TYPE OBJECT_TYPE ------------------------------ ------------------------------ ------- ------------------------- .......... .......... SYS CDC$T_CDC_DEMO_SET OBJECT SYS.ANYDATA .......... SQL> SELECT set_name, change_source_name, capture_enabled, stop_on_ddl, publisher 2 FROM change_sets; SET_NAME CHANGE_SOURCE_NAME C S PUBLISHER ------------------------------ ------------------------------ - - ------------------------------ SYNC_SET SYNC_SOURCE Y N CDC_DEMO_SET HOTLOG_SOURCE Y Y SYS SQL> SQL> SELECT subscription_name, handle, set_name, username, earliest_scn, description 2 FROM cdc_subscribers$; no rows selected >>>>>>>>>>> connect subs_cdc/subs_cdc exec dbms_cdc_subscribe.create_subscription('CDC_DEMO_SET', 'cdc_demo subx', 'CDC_DEMO_SUB'); SQL> exec dbms_cdc_subscribe.create_subscription('CDC_DEMO_SET', 'cdc_demo subx', 'CDC_DEMO_SUB'); PL/SQL procedure successfully completed. >>>>>>>>>>> connect / as sysdba SQL> SELECT subscription_name, handle, set_name, username, earliest_scn, description 2 FROM cdc_subscribers$; SUBSCRIPTION_NAME HANDLE SET_NAME USERNAME EARLIEST_SCN DESCRIP ------------------------------ ---------- ------------------------------ --------------------------- CDC_DEMO_SUB 1 CDC_DEMO_SET subs_cdc 1 cdc_dem Note: If you want to drop a subscription, use: DBMS_CDC_SUBSCRIBE.DROP_SUBSCRIPTION(subscription_name IN VARCHAR2); DBMS_CDC_SUBSCRIBE.DROP_SUBSCRIPTION('SUBSCRIPTION_ALBERT'); >>>>>>>>>> connect subs_cdc/subs_cdc BEGIN dbms_cdc_subscribe.subscribe('CDC_DEMO_SUB', 'ALBERT', 'PERSOON', 'userid, name, lastname', 'CDC_DEMO_SUB_VIEW'); END; / SQL> BEGIN 2 dbms_cdc_subscribe.subscribe('CDC_DEMO_SUB', 'ALBERT', 'PERSOON', 3 'userid, name, lastname', 'CDC_DEMO_SUB_VIEW'); 4 END; 5 / PL/SQL procedure successfully completed. SQL> SELECT set_name, subscription_name, status 2 FROM user_subscriptions; SET_NAME SUBSCRIPTION_NAME S ------------------------------ ------------------------------ - CDC_DEMO_SET CDC_DEMO_SUB N exec dbms_cdc_subscribe.activate_subscription('CDC_DEMO_SUB'); SQL> exec dbms_cdc_subscribe.activate_subscription('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> SELECT set_name, subscription_name, status 2 FROM user_subscriptions; SET_NAME SUBSCRIPTION_NAME S ------------------------------ ------------------------------ - CDC_DEMO_SET CDC_DEMO_SUB A >>>>>>>>>>> connect albert/albert SQL> insert into persoon 2 values 3 (1,'piet','pietersen'); 1 row created. SQL> commit; Commit complete. SQL> insert into persoon 2 values 3 (2,'jan','janssen'); 1 row created. SQL> commit; Commit complete. >>>>>>>>>>>>>>> connect subs_cdc/subs_cdc exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); SQL> select * from publ_cdc.CDC_PERSOON; OP CSCN$ COMMIT_TI XIDUSN$ XIDSLT$ XIDSEQ$ RSID$ ROW_ID$ USERNAME$ -- ---------- --------- ---------- ---------- ---------- ---------- ------------------ ------------- I 627180 27-FEB-08 2 44 323 1 AAAM1CAAGAAAAAQAAA ALBERT I 627232 27-FEB-08 10 7 326 2 AAAM1CAAGAAAAAQAAB ALBERT SQL> select * from CDC_DEMO_SUB_VIEW; OP CSCN$ COMMIT_TI XIDUSN$ XIDSLT$ XIDSEQ$ ROW_ID$ RSID$ TARGET_COLMAP$ -- ---------- --------- ---------- ---------- ---------- ------------------ ---------- ------------- I 627180 27-FEB-08 2 44 323 AAAM1CAAGAAAAAQAAA 1 FE7F000000000000000000000000000 I 627232 27-FEB-08 10 7 326 AAAM1CAAGAAAAAQAAB 2 FE7F000000000000000000000000000 SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAQAAA 1 piet I 27-FEB-08 AAAM1CAAGAAAAAQAAB 2 jan >>>>>>>>>>>>>> connect albert/albert insert into persoon values (3,'kees','pot'); >>>>>>>>>>>>>> connect subs_cdc/subs_cdc SQL> select * from publ_cdc.CDC_PERSOON; OP CSCN$ COMMIT_TI XIDUSN$ XIDSLT$ XIDSEQ$ RSID$ ROW_ID$ USERNAME$ -- ---------- --------- ---------- ---------- ---------- ---------- ------------------ ------------- I 627180 27-FEB-08 2 44 323 1 AAAM1CAAGAAAAAQAAA ALBERT I 627232 27-FEB-08 10 7 326 2 AAAM1CAAGAAAAAQAAB ALBERT I 628175 27-FEB-08 9 16 351 10001 AAAM1CAAGAAAAAQAAC ALBERT SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAQAAA 1 piet I 27-FEB-08 AAAM1CAAGAAAAAQAAB 2 jan exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); SQL> exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAQAAA 1 piet I 27-FEB-08 AAAM1CAAGAAAAAQAAB 2 jan I 27-FEB-08 AAAM1CAAGAAAAAQAAC 3 kees exec dbms_cdc_subscribe.purge_window('CDC_DEMO_SUB'); SQL> exec dbms_cdc_subscribe.purge_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select * from publ_cdc.CDC_PERSOON; OP CSCN$ COMMIT_TI XIDUSN$ XIDSLT$ XIDSEQ$ RSID$ ROW_ID$ USERNAME$ -- ---------- --------- ---------- ---------- ---------- ---------- ------------------ ------------- I 627180 27-FEB-08 2 44 323 1 AAAM1CAAGAAAAAQAAA ALBERT I 627232 27-FEB-08 10 7 326 2 AAAM1CAAGAAAAAQAAB ALBERT I 628175 27-FEB-08 9 16 351 10001 AAAM1CAAGAAAAAQAAC ALBERT SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; no rows selected >>>>>>>>>>>>>> connect albert/albert Connected. SQL> insert into persoon 2 values 3 (4,'joop','joopsen'); >>>>>>>>>>>>>>> connect subs_cdc/subs_cdc Connected. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; no rows selected SQL> select * from publ_cdc.CDC_PERSOON; OP CSCN$ COMMIT_TI XIDUSN$ XIDSLT$ XIDSEQ$ RSID$ ROW_ID$ USERNAME$ -- ---------- --------- ---------- ---------- ---------- ---------- ------------------ ------------- I 627180 27-FEB-08 2 44 323 1 AAAM1CAAGAAAAAQAAA ALBERT I 627232 27-FEB-08 10 7 326 2 AAAM1CAAGAAAAAQAAB ALBERT I 628175 27-FEB-08 9 16 351 10001 AAAM1CAAGAAAAAQAAC ALBERT I 628841 27-FEB-08 5 3 350 20001 AAAM1CAAGAAAAAPAAA ALBERT SQL> exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAPAAA 4 joop >>>>>>>>>>>>>> connect albert/albert SQL> insert into persoon 2 values 3 (5,'gerrit','gerritsen'); 1 row created. SQL> commit; >>>>>>>>>>>>>> connect subs_cdc/subs_cdc SQL> select * from publ_cdc.CDC_PERSOON; OP CSCN$ COMMIT_TI XIDUSN$ XIDSLT$ XIDSEQ$ RSID$ ROW_ID$ USERNAME$ -- ---------- --------- ---------- ---------- ---------- ---------- ------------------ ------------- I 636854 27-FEB-08 2 7 333 30001 AAAM1CAAGAAAAAOAAA ALBERT I 627180 27-FEB-08 2 44 323 1 AAAM1CAAGAAAAAQAAA ALBERT I 627232 27-FEB-08 10 7 326 2 AAAM1CAAGAAAAAQAAB ALBERT I 628175 27-FEB-08 9 16 351 10001 AAAM1CAAGAAAAAQAAC ALBERT I 628841 27-FEB-08 5 3 350 20001 AAAM1CAAGAAAAAPAAA ALBERT SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAPAAA 4 joop SQL> exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAOAAA 5 gerrit I 27-FEB-08 AAAM1CAAGAAAAAPAAA 4 joop >>>>>>>>>>>>>>>> connect albert/albert@test10g Connected. SQL> insert into persoon 2 values 3 (6,'marie','bruinsma'); 1 row created. SQL> commit; Commit complete. >>>>>>>>>>>>>>>> connect subs_cdc/subs_cdc Connected. SQL> select * from publ_cdc.CDC_PERSOON; OP CSCN$ COMMIT_TI XIDUSN$ XIDSLT$ XIDSEQ$ RSID$ ROW_ID$ USERNAME$ -- ---------- --------- ---------- ---------- ---------- ---------- ------------------ ------------- I 636854 27-FEB-08 2 7 333 30001 AAAM1CAAGAAAAAOAAA ALBERT I 643057 27-FEB-08 9 13 364 40001 AAAM1CAAGAAAAAPAAB ALBERT I 627180 27-FEB-08 2 44 323 1 AAAM1CAAGAAAAAQAAA ALBERT I 627232 27-FEB-08 10 7 326 2 AAAM1CAAGAAAAAQAAB ALBERT I 628175 27-FEB-08 9 16 351 10001 AAAM1CAAGAAAAAQAAC ALBERT I 628841 27-FEB-08 5 3 350 20001 AAAM1CAAGAAAAAPAAA ALBERT 6 rows selected. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAOAAA 5 gerrit I 27-FEB-08 AAAM1CAAGAAAAAPAAA 4 joop SQL> exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 27-FEB-08 AAAM1CAAGAAAAAOAAA 5 gerrit I 27-FEB-08 AAAM1CAAGAAAAAPAAB 6 marie I 27-FEB-08 AAAM1CAAGAAAAAPAAA 4 joop >>>>>>>>>>> Now about RMAN: A redo log file used by Change Data Capture must remain available on the staging database until Change Data Capture has captured it. However, it is not necessary that the redo log file remain available until the Change Data Capture subscriber is done with the change data. To determine which redo log files are no longer needed by Change Data Capture for a given change set, the publisher alters the change set's Streams capture process, which causes Streams to perform some internal cleanup and populates the DBA_LOGMNR_PURGED_LOG view. The publisher follows these steps: Uses the following query on the staging database to get the three SCN values needed to determine an appropriate new first_scn value for the change set, CHICAGO_DAILY: SELECT cap.CAPTURE_NAME, cap.FIRST_SCN, cap.APPLIED_SCN, cap.REQUIRED_CHECKPOINT_SCN FROM DBA_CAPTURE cap, CHANGE_SETS cset WHERE cset.SET_NAME = 'CDC_DEMO_SET' AND cap.CAPTURE_NAME = cset.CAPTURE_NAME; SQL> SELECT cap.CAPTURE_NAME, cap.FIRST_SCN, cap.APPLIED_SCN, 2 cap.REQUIRED_CHECKPOINT_SCN 3 FROM DBA_CAPTURE cap, CHANGE_SETS cset 4 WHERE cset.SET_NAME = 'CDC_DEMO_SET' AND 5 cap.CAPTURE_NAME = cset.CAPTURE_NAME; CAPTURE_NAME FIRST_SCN APPLIED_SCN REQUIRED_CHECKPOINT_SCN ------------------------------ ---------- ----------- ----------------------- CDC$C_CDC_DEMO_SET 610086 672502 665072 SQL> SELECT recid, first_change#, sequence#, next_change# 2 FROM V$LOG_HISTORY; RECID FIRST_CHANGE# SEQUENCE# NEXT_CHANGE# ---------- ------------- ---------- ------------ 1 534907 1 555371 2 555371 2 557968 ......... 68 648702 68 651777 69 651777 69 653085 70 653085 70 655053 71 655053 71 655234 72 655234 72 656658 73 656658 73 657846 74 657846 74 659879 75 659879 75 662288 76 662288 76 662292 77 662292 77 662297 78 662297 78 662312 79 662312 79 662322 80 662322 80 662329 81 662329 81 662337 --> 82 662337 82 664708 83 664708 83 665724 84 665724 84 670061 85 670061 85 674246 85 rows selected. SQL> SELECT first_change#, next_change#, sequence#, archived, substr(name, 1, 40) 2 FROM V$ARCHIVED_LOG; FIRST_CHANGE# NEXT_CHANGE# SEQUENCE# ARC SUBSTR(NAME,1,40) ------------- ------------ ---------- --- ---------------------------------------------------------- 570647 572710 19 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR ......... 657846 659879 74 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 659879 662288 75 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 662288 662292 76 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 662292 662297 77 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 662297 662312 78 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 662312 662322 79 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 662322 662329 80 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 662329 662337 81 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 662337 664708 82 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR --> 664708 665724 83 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 665724 670061 84 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 670061 674246 85 YES C:\ORACLE\FLASH_RECOVERY_AREA\TEST10G\AR 104 rows selected. >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> TEST: <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< Dinsdag 13:15:47 2008 in alertlog C001: long running txn detected, xid: 0x0003.02a.00000160 C001: long txn committed, xid: 0x0003.02a.00000160 Dinsdag 16.00: >>>>>>>>>>> connect albert/albert SQL> insert into persoon 2 values 3 (8,'appie','sel'); 1 row created. SQL> commit; >>>>>>>>>>> connect subs_cdc/subs_cdc SQL> select OPERATION$,RSID$,USERID,NAME,LASTNAME from publ_cdc.CDC_PERSOON; OP RSID$ USERID NAME LASTNAME -- ---------- ---------- ------------------------------ ------------------------------ I 30001 5 gerrit gerritsen I 40001 6 marie bruinsma I 50001 7 lubbie lubbie I 50002 8 appie sel <-------- inderdaad toegevoegd I 1 1 piet pietersen I 2 2 jan janssen I 10001 3 kees pot I 20001 4 joop joopsen 8 rows selected. SQL> connect sys/vga88nt@test10g as sysdba Connected. SQL> SELECT cap.CAPTURE_NAME, cap.FIRST_SCN, cap.APPLIED_SCN, 2 cap.REQUIRED_CHECKPOINT_SCN 3 FROM DBA_CAPTURE cap, CHANGE_SETS cset 4 WHERE cset.SET_NAME = 'CDC_DEMO_SET' AND 5 cap.CAPTURE_NAME = cset.CAPTURE_NAME; CAPTURE_NAME FIRST_SCN APPLIED_SCN REQUIRED_CHECKPOINT_SCN ------------------------------ ---------- ----------- ----------------------- CDC$C_CDC_DEMO_SET 610086 683349 683184 Dinsdag 20.00h: SQL> connect sys/vga88nt@test10g as sysdba Connected. SQL> SELECT cap.CAPTURE_NAME, cap.FIRST_SCN, cap.APPLIED_SCN, 2 cap.REQUIRED_CHECKPOINT_SCN 3 FROM DBA_CAPTURE cap, CHANGE_SETS cset 4 WHERE cset.SET_NAME = 'CDC_DEMO_SET' AND 5 cap.CAPTURE_NAME = cset.CAPTURE_NAME; CAPTURE_NAME FIRST_SCN APPLIED_SCN REQUIRED_CHECKPOINT_SCN ------------------------------ ---------- ----------- ----------------------- CDC$C_CDC_DEMO_SET 610086 690551 683349 NO LONG RUNNING TRANSACTIONS Dinsdag 21.20h >>>>>>>>>>>>>>> conn albert/albert SQL> connect albert/albert@test10g Connected. SQL> insert into persoon 2 values 3 (10,'marietje','popje'); 1 row created. Woe 08.00h >>>>>>>>>>>>>> conn subs_cdc/subs_cdc SQL> select * from albert.persoon; USERID NAME LASTNAME ---------- ------------------------------ -------------- 5 gerrit gerritsen 4 joop joopsen 6 marie bruinsma 7 lubbie lubbie 8 appie sel 1 piet pietersen 2 jan janssen 3 kees pot 8 rows selected. SQL> select OPERATION$,RSID$,USERID,NAME,LASTNAME from publ_cdc.CDC_PERSOON; OP RSID$ USERID NAME LASTNAME -- ---------- ---------- ------------------------------ ------------------------------ I 50001 7 lubbie lubbie I 50002 8 appie sel SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; no rows selected >>>>>>>>>>>>>>>>>>>>>>connect sys/vga88nt@test10g as sysdba Connected. SQL> select * from DBA_SOURCE_TABLES; SOURCE_SCHEMA_NAME SOURCE_TABLE_NAME ------------------------------ ------------------------- ALBERT PERSOON Woe 08.30: >>>>>>>>>>>>>>>>>>>>>> connect albert/albert@test10g Connected. SQL> insert into persoon 2 values 3 (9,'nadia','nadia'); 1 row created. SQL> commit; Commit complete. >>>>>>>>>>>>>>>>>>>>> connect subs_cdc/subs_cdc@test10g Connected. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; no rows selected SQL> exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 28-FEB-08 AAAM1CAAGAAAAAPAAC 7 lubbie I 28-FEB-08 AAAM1CAAGAAAAAPAAD 8 appie I 28-FEB-08 AAAM1CAAGAAAAAQAAD 9 nadia SQL> select OPERATION$,RSID$,USERID,NAME,LASTNAME from publ_cdc.CDC_PERSOON; OP RSID$ USERID NAME LASTNAME -- ---------- ---------- ------------------------------ ------------------------------ I 50001 7 lubbie lubbie I 50002 8 appie sel I 60001 9 nadia nadia Woe 8.45: >>>>>>>>>>>>>>> connect albert/albert@test10g SQL> insert into persoon 2 values 3 (10,'lejah','lejah'); 1 row created. No commit done Woe 9:22:12 2008 C001: long running txn detected, xid: 0x0006.025.0000018f etc.. 12:02:31 2008 C001: long running txn detected, xid: 0x0006.025.0000018f 12.15 COMMIT RMAN FULL BACKUP MADE 12.30 SQL> insert into persoon 2 values 3 (11,'mira','mira'); 1 row created. No COMMIT 13:02:38 2008 C001: long running txn detected, xid: 0x000a.019.000001a2 13:12:38 2008 C001: long running txn detected, xid: 0x000a.019.000001a2 13:22:40 2008 C001: long running txn detected, xid: 0x000a.019.000001a2 COMMIT 13:26:25 2008 C001: long txn committed, xid: 0x000a.019.000001a2 ? knllgobjinfo: MISSING Streams multi-version data dictionary 13.29: SQL> create table persoon2 2 ( 3 userid number, 4 name varchar(30), 5 lastname varchar(30), 6 constraint pk_userid2 PRIMARY KEY (userid)); Table created. SQL> insert into persoon2 2 values 3 (1,'piet','piet'); 1 row created. SQL> commit; Commit complete. 13.46: SQL> insert into persoon2 -- COMPLETELY INDEPENDENT OF CDC 2 values 3 (2,'karel','karel'); 1 row created. SQL> NO COMMIT 16.00h NEVER LONG RUNNING TRANSACTION DETECTED. 27 Feb 16.00h >>>>>>>>>>>> conn albert/albert SQL> insert into persoon 2 values 3 (12,'xyz','xyz'); 1 row created. SQL> commit; >>>>>>>>>>>>> conn subs_cdc/subs_cdc SQL> select * from albert.persoon; USERID NAME LASTNAME ---------- ------------------------------ ------------------------------ 5 gerrit gerritsen 4 joop joopsen 6 marie bruinsma 7 lubbie lubbie 8 appie sel 1 piet pietersen 2 jan janssen 3 kees pot 9 nadia nadia 10 lejah lejah 11 mira mira 12 xyz xyz 12 rows selected. SQL> select OPERATION$,COMMIT_TIMESTAMP$,USERNAME$,USERID,NAME from publ_cdc.cdc_persoon; OP COMMIT_TI USERNAME$ USERID NAME -- --------- ------------------------------ ---------- ------------------------------ I 28-FEB-08 ALBERT 7 lubbie I 28-FEB-08 ALBERT 8 appie I 28-FEB-08 ALBERT 9 nadia I 29-FEB-08 ALBERT 10 lejah I 29-FEB-08 ALBERT 11 mira I 29-FEB-08 ALBERT 12 xyz 6 rows selected. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 28-FEB-08 AAAM1CAAGAAAAAPAAC 7 lubbie I 28-FEB-08 AAAM1CAAGAAAAAPAAD 8 appie I 28-FEB-08 AAAM1CAAGAAAAAQAAD 9 nadia SQL> exec dbms_cdc_subscribe.extend_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select OPERATION$,COMMIT_TIMESTAMP$,USERNAME$,USERID,NAME from publ_cdc.cdc_persoon; OP COMMIT_TI USERNAME$ USERID NAME -- --------- ------------------------------ ---------- ------------------------------ I 28-FEB-08 ALBERT 7 lubbie I 28-FEB-08 ALBERT 8 appie I 28-FEB-08 ALBERT 9 nadia I 29-FEB-08 ALBERT 10 lejah I 29-FEB-08 ALBERT 11 mira I 29-FEB-08 ALBERT 12 xyz 6 rows selected. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; OP COMMIT_TI ROW_ID$ USERID NAME -- --------- ------------------ ---------- ------------------------------ I 28-FEB-08 AAAM1CAAGAAAAAPAAC 7 lubbie I 28-FEB-08 AAAM1CAAGAAAAAPAAD 8 appie I 28-FEB-08 AAAM1CAAGAAAAAQAAD 9 nadia I 29-FEB-08 AAAM1CAAGAAAAAQAAE 10 lejah I 29-FEB-08 AAAM1CAAGAAAAAQAAF 11 mira I 29-FEB-08 AAAM1CAAGAAAAAQAAG 12 xyz 6 rows selected. SQL> exec dbms_cdc_subscribe.purge_window('CDC_DEMO_SUB'); PL/SQL procedure successfully completed. SQL> select OPERATION$,COMMIT_TIMESTAMP$,USERNAME$,USERID,NAME from publ_cdc.cdc_persoon; OP COMMIT_TI USERNAME$ USERID NAME -- --------- ------------------------------ ---------- ------------------------------ I 28-FEB-08 ALBERT 7 lubbie I 28-FEB-08 ALBERT 8 appie I 28-FEB-08 ALBERT 9 nadia I 29-FEB-08 ALBERT 10 lejah I 29-FEB-08 ALBERT 11 mira I 29-FEB-08 ALBERT 12 xyz 6 rows selected. SQL> select OPERATION$, COMMIT_TIMESTAMP$, ROW_ID$, USERID, NAME from CDC_DEMO_SUB_VIEW; no rows selected SQL> Note about "long running txn": ------------------------------ I have found the following definition of a long running transaction: - A long-running transaction is a transaction that has not received any LCRs for over 10 minutes. Open transactions (ie, transactions where the commit or rollback has not been received) without new LCRs in 10 minutes will spill to the apply spill table. In dba_apply_parameters you can find parameters of the apply process. APPLY_NAME PARAMETER VALUE ------------------------------ ------------------------------ ---------- CDC$A_CHANGE_SET_ALBERT ALLOW_DUPLICATE_ROWS N CDC$A_CHANGE_SET_ALBERT COMMIT_SERIALIZATION NONE CDC$A_CHANGE_SET_ALBERT DISABLE_ON_ERROR Y CDC$A_CHANGE_SET_ALBERT DISABLE_ON_LIMIT Y CDC$A_CHANGE_SET_ALBERT MAXIMUM_SCN INFINITE CDC$A_CHANGE_SET_ALBERT PARALLELISM 1 CDC$A_CHANGE_SET_ALBERT STARTUP_SECONDS 0 CDC$A_CHANGE_SET_ALBERT TIME_LIMIT INFINITE CDC$A_CHANGE_SET_ALBERT TRACE_LEVEL 0 CDC$A_CHANGE_SET_ALBERT TRANSACTION_LIMIT INFINITE CDC$A_CHANGE_SET_ALBERT TXN_LCR_SPILL_THRESHOLD 10000 CDC$A_CHANGE_SET_ALBERT WRITE_ALERT_LOG Y =============================================================================== TEST CASE: Export CDC objects from DATABASE TEST10G to DATABASE TEST10G2 - Make database properties in TEST10G2 same as in TEST10G (example, archive logging, pools etc..) - Create same CDC related tablespaces - Create users in TEST10G2 DB - GRANT ALL APROPRIATE PERMISSIONS - Export from TEST10G - IMPORT INTO TEST10G2 =============================================================================== =============================================================================== PROBLEMS: ========= 1. long running txn detected: ----------------------------- Not serious. 2. RMAN-08137: WARNING: archive log not deleted as it is still needed: ---------------------------------------------------------------------- WARNING: archive log not deleted as it is still needed Cause An archivelog that should have been deleted was not as it was required by Streams or Data Guard. The next message identifies the archivelog. Action This is an informational message. The archivelog can be deleted after it is no longer needed. See the documentation for Data Guard to alter the set of active Data Guard destinations. See the documentation for Streams to alter the set of active streams. Starting backup at 27-FEB-08 channel t1: starting archive log backupset channel t1: specifying archive log(s) in backup set input archive log thread=1 sequence=600 recid=570 stamp=647820534 channel t1: starting piece 1 at 27-FEB-08 channel t1: finished piece 1 at 27-FEB-08 piece handle=ipj9q29g_1_1 tag=TAG20080227T233511 comment=API Version 2.0,MMS Version 5.3.3.0 channel t1: backup set complete, elapsed time: 00:00:04 RMAN-08137: WARNING: archive log not deleted as it is still needed archive log filename=/dbms/tdbaaccp/accptrid/recovery/archive/arch_1_600_630505403.arch thread=1 sequence=600 Finished backup at 27-FEB-08 Thu Feb 28 00:00:01 2008 Starting backup at 28-FEB-08 channel t1: starting archive log backupset channel t1: specifying archive log(s) in backup set input archive log thread=1 sequence=600 recid=570 stamp=647820534 channel t1: starting piece 1 at 28-FEB-08 channel t1: finished piece 1 at 28-FEB-08 piece handle=isj9q3o5_1_1 tag=TAG20080228T000004 comment=API Version 2.0,MMS Version 5.3.3.0 channel t1: backup set complete, elapsed time: 00:00:04 RMAN-08137: WARNING: archive log not deleted as it is still needed archive log filename=/dbms/tdbaaccp/accptrid/recovery/archive/arch_1_600_630505403.arch thread=1 sequence=600 Finished backup at 28-FEB-08 Thu Feb 28 01:00:01 2008 Starting backup at 28-FEB-08 channel t1: starting archive log backupset channel t1: specifying archive log(s) in backup set input archive log thread=1 sequence=600 recid=570 stamp=647820534 channel t1: starting piece 1 at 28-FEB-08 channel t1: finished piece 1 at 28-FEB-08 piece handle=ivj9q78l_1_1 tag=TAG20080228T010004 comment=API Version 2.0,MMS Version 5.3.3.0 channel t1: backup set complete, elapsed time: 00:00:04 channel t1: deleting archive log(s) archive log filename=/dbms/tdbaaccp/accptrid/recovery/archive/arch_1_600_630505403.arch recid=570 stamp=647820534 Finished backup at 28-FEB-08 Also handled. 3. ORA-00600: internal error code, arguments: [knlcLoop-200], [], [], [], [], [], [], [] ---------------------------------------------------------------------------------------- LOGMINER: End mining logfile: /dbms/tdbaaccp/accptrid/recovery/redo_logs/redo03.log Thu Feb 28 09:10:30 2008 LOGMINER: Begin mining logfile: /dbms/tdbaaccp/accptrid/recovery/redo_logs/redo01.log Thu Feb 28 09:21:01 2008 Thread 1 advanced to log sequence 608 Current log# 2 seq# 608 mem# 0: /dbms/tdbaaccp/accptrid/recovery/redo_logs/redo02.log Thu Feb 28 09:21:01 2008 LOGMINER: End mining logfile: /dbms/tdbaaccp/accptrid/recovery/redo_logs/redo01.log Thu Feb 28 09:21:01 2008 LOGMINER: Begin mining logfile: /dbms/tdbaaccp/accptrid/recovery/redo_logs/redo02.log Thu Feb 28 09:22:46 2008 Errors in file /dbms/tdbaaccp/accptrid/admin/dump/bdump/accptrid_c001_1491066.trc: ORA-00600: internal error code, arguments: [knlcLoop-200], [], [], [], [], [], [], [] Thu Feb 28 09:22:59 2008 Streams CAPTURE C001 with pid=25, OS id=1491066 stopped Thu Feb 28 09:22:59 2008 Errors in file /dbms/tdbaaccp/accptrid/admin/dump/bdump/accptrid_c001_1491066.trc: ORA-00600: internal error code, arguments: [knlcLoop-200], [], [], [], [], [], [], [] On AIX: IDENTIFIER TIMESTAMP T C RESOURCE_NAME DESCRIPTION A63BEB70 0221101308 P S SYSPROC SOFTWARE PROGRAM ABNORMALLY TERMINATED SQL> select SEQUENCE#, FIRST_CHANGE#,STATUS,ARCHIVED from v$log; SEQUENCE# FIRST_CHANGE# STATUS ARC ---------- ------------- ---------------- --- 607 9579856 INACTIVE YES 608 9594819 CURRENT NO 606 9579542 INACTIVE YES SQL> / SEQUENCE# FIRST_CHANGE# STATUS ARC ---------- ------------- ---------------- --- 607 9579856 INACTIVE YES 608 9594819 CURRENT NO 606 9579542 INACTIVE YES Warning: Errors detected in file /dbms/tdbaaccp/accptrid/admin/dump/bdump/accptrid_c001_1499336.trc > /dbms/tdbaaccp/accptrid/admin/dump/bdump/accptrid_c001_1499336.trc > Oracle Database 10g Enterprise Edition Release 10.2.0.3.0 - 64bit Production > With the Partitioning, OLAP and Data Mining options > ORACLE_HOME = /dbms/tdbaaccp/ora10g/home > System name: AIX > Node name: pl003 > Release: 3 > Version: 5 > Machine: 00CB560D4C00 > Instance name: accptrid > Redo thread mounted by this instance: 1 > Oracle process number: 23 > Unix process pid: 1499336, image: oracle@pl003 (C001) > > *** 2008-02-28 10:49:04.501 > *** SERVICE NAME:(SYS$USERS) 2008-02-28 10:49:04.488 > *** SESSION ID:(195.1286) 2008-02-28 10:49:04.488 > KnlcLoop: priorCkptScn currentCkptScn > 0x0000.00926d84 0x0000.00927073 > knlcLoop: buf_txns_knlcctx:1:: lowest bufLcrScn:0x0000.00926cca > knlcPrintCharCachedTxn:xid: 0x000b.005.000001ac > *** 2008-02-28 10:49:04.501 > ksedmp: internal or fatal error > ORA-00600: internal error code, arguments: [knlcLoop-200], [], [], [], [], [], [], [] > OPIRIP: Uncaught error 447. Error stack: > ORA-00447: fatal error in background process > ORA-00600: internal error code, arguments: [knlcLoop-200], [], [], [], [], [], [], [] ============================================================================================== END OF Async CDC extended TEST: ============================================================================================== exec dbms_cdc_subscribe.extend_window('CHANGE_SET_ALBERT'); exec dbms_cdc_subscribe.purge_window('CHANGE_SET_ALBERT'); exec DBMS_CDC_PUBLISH.DROP_CHANGE_SET('CHANGE_SET_ALBERT'); exec dbms_capture_adm.abort_table_instantiation('HR.CDC_DEMO'); -- drop the change set exec dbms_cdc_publish.drop_change_set('CDC_DEMO_SET'); ============= 26 X$ TABLES: ============= Listed below are some of the important subsystems in the Oracle kernel. This table might help you to read those dreaded trace files and internal messages. For example, if you see messages like this, you will at least know where they come from: OPIRIP: Uncaught error 447. Error stack: KCF: write/open error block=0x3e800 online=1 Kernel Subsystems: OPI Oracle Program Interface KK Compilation Layer - Parse SQL, compile PL/SQL KX Execution Layer - Bind and execute SQL and PL/SQL K2 Distributed Execution Layer - 2PC handling NPI Network Program Interface KZ Security Layer - Validate privs KQ Query Layer RPI Recursive Program Interface KA Access Layer KD Data Layer KT Transaction Layer KC Cache Layer KS Services Layer KJ Lock Manager Layer KG Generic Layer KV Kernel Variables (eg. x$KVIS and X$KVII) S or ODS Operating System Dependencies Where can one get a list of all hidden Oracle parameters? Oracle initialization or INIT.ORA parameters with an underscore in front are hidden or unsupported parameters. One can get a list of all hidden parameters by executing this query: select * from SYS.X$KSPPI where substr(KSPPINM,1,1) = '_'; The following query displays parameter names with their current value: select a.ksppinm "Parameter", b.ksppstvl "Session Value", c.ksppstvl "Instance Value" from x$ksppi a, x$ksppcv b, x$ksppsv c where a.indx = b.indx and a.indx = c.indx and substr(ksppinm,1,1)='_' order by a.ksppinm; Remember: Thou shall not play with undocumented parameters! Oracle's x$ Tables See also: Speculation of X$ Table Names x$ tables are the sql interface to viewing oracle's memory in the SGA. The names for the x$ tables can be queried with select kqftanam from x$kqfta; x$activeckpt x$bh Information on buffer headers. Contains a record (the buffer header) for each block in the buffer cache. This select statement lists how many blocks are Available, Free and Being Used. select count(*), State from ( select decode (state, 0, 'Free', 1, decode (lrba_seq, 0, 'Available', 'Being Used'), 3, 'Being Used', state) State from x$bh ) group by state The meaning of state: 0 FREE no valid block image 1 XCUR a current mode block, exclusive to this instance 2 SCUR a current mode block, shared with other instances 3 CR a consistent read (stale) block image 4 READ buffer is reserved for a block being read from disk 5 MREC a block in media recovery mode 6 IREC a block in instance (crash) recovery mode The meaning of tch: tch is the touch count. A high touch count indicates that the buffer is used often. Therefore, it will probably be at the head of the MRU list. See also touch count. The meaning of tim: touch time. class represents a value designated for the use of the block. lru_flag set_ds maps to addr on x$kcbwds. le_addr can be outer joined on x$le.le_addr. flag is a bit array. Bit if set 0 Block is dirty 4 temporary block 9 or 10 ping 14 stale 16 direct 524288 (=0x80000) Block was read in a full table scan See this link x$bufqm x$class_stat x$context x$globalcontext x$hofp x$hs_session The x$kc... tables x$kcbbhs x$kcbmmav x$kcbsc x$kcbwait x$kcbwbpd Buffer pool descriptor, the base table for v$buffer_pool. How is the buffer cache split between the default, the recycle and the keep buffer pool. x$kcbwds Set descriptor, see also x$kcbwbpd The column id can be joined with v$buffer_pool.id. The column bbwait corresponds to the buffer busy waits wait event. Information on working set buffers addr can be joined with x$bh.set_ds. set_id will be between lo_setid and hi_setid in v$buffer_pool for the relevant buffer pool. x$kccal x$kccbf x$kccbi x$kccbl x$kccbp x$kccbs x$kcccc x$kcccf x$kccdc x$kccdi x$kccdl x$kccfc x$kccfe x$kccfn x$kccic x$kccle Controlfile logfile entry. Use select max(lebsz) from x$kccle to find out the size of a log block. The log block size is the unit for the following init params: log_checkpoint_interval, _log_io_size, and max_dump_file_size. x$kcclh x$kccor x$kcccp Checkpoint Progress: The column cpodr_bno displays the current redo block number. Multiplied with the OS Block Size (usually 512), it returns the amount of bytes of redo currently written to the redo logs. Hence, this number is reset at each log switch. k$kcccp can (together with x$kccle) be used to monitor the progress of the writing of online redo logs. The following query does this. select le.leseq "Current log sequence No", 100*cp.cpodr_bno/le.lesiz "Percent Full", cp.cpodr_bno "Current Block No", le.lesiz "Size of Log in Blocks" from x$kcccp cp, x$kccle le where LE.leseq =CP.cpodr_seq and bitand(le.leflg,24)=8; bitand(le.leflg,24)=8 makes sure we get the current log group How much redo is written by Oracle uses a variation of this SQL statement to track how much redo is written by different DML Statements. x$kccrs x$kccrt x$kccsl x$kcctf x$kccts x$kcfio x$kcftio x$kckce x$kckty x$kclcrst x$kcrfx x$kcrmf x$kcrmx x$kcrralg x$kcrrarch x$kcrrdest x$kcrrdstat x$kcrrms x$kcvfh x$kcvfhmrr x$kcvfhonl x$kcvfhtmp x$kdnssf The x$kg... tables KG stands for kernel generic x$kghlu This view shows one row per shared pool area. If there's a java pool, an additional row is displayed. x$kgicc x$kgics x$kglcursor x$kgldp x$kgllk This table lists all held and requested library object locks for all sessions. It is more complete than v$lock. The column kglnaobj displays the first 80 characters of the name of the object. select kglnaobj, kgllkreq from x$kgllk x join v$session s on s.saddr = x.kgllkses; kgllkreq = 0 means, the lock is held, while kgllkreq > 0 means that the lock is requested. x$kglmem x$kglna x$kglna1 x$kglob Library Cache Object x$kglsim x$kglst x$kgskasp x$kgskcft x$kgskcp x$kgskdopp x$kgskpft x$kgskpp x$kgskquep x$kjbl x$kjbr x$kjdrhv x$kjdrpcmhv x$kjdrpcmpf x$kjicvt x$kjilkft x$kjirft x$kjisft x$kjitrft x$kksbv x$kkscs x$kkssrd x$klcie x$klpt x$kmcqs x$kmcvc x$kmmdi x$kmmrd x$kmmsg x$kmmsi x$knstacr x$knstasl x$knstcap x$knstmvr x$knstrpp x$knstrqu x$kocst The x$kq... tables x$kqfco This table has an entry for each column of the x$tables and can be joined with x$kqfta. The column kqfcosiz indicates the size (in bytes?) of the columns. select t.kqftanam "Table Name", c.kqfconam "Column Name", c.kqfcosiz "Column Size" from x$kqfta t, x$kqfco c where t.indx = c.kqfcotab x$kqfdt x$kqfsz x$kqfta It seems that all x$table names can be retrieved with the following query. select kqftanam from x$kqfta; This table can be joined with x$kqfco which contains the columns for the tables: select t.kqftanam "Table Name", c.kqfconam "Column Name" from x$kqfta t, x$kqfco c where t.indx = c.kqfcotab x$kqfvi x$kqfvt x$kqlfxpl x$kqlset x$kqrfp x$kqrfs x$kqrst x$krvslv x$krvslvs x$krvxsv The x$ks... tables KS stands for kernel services. x$ksbdd x$ksbdp x$ksfhdvnt x$ksfmcompl x$ksfmelem x$ksfmextelem x$ksfmfile x$ksfmfileext x$ksfmiost x$ksfmlib x$ksfmsubelem x$ksfqp x$ksimsi x$ksled x$kslei x$ksles x$kslld x$ksllt x$ksllw x$kslwsc x$ksmfs x$ksmfsv This SGA map. x$ksmge x$ksmgop x$ksmgsc x$ksmgst x$ksmgv x$ksmhp x$ksmjch x$ksmjs x$ksmlru Memory least recently used Whenever a select is performed on x$ksmlru, its content is reset! This table show which memory allocations in the shared pool caused the throw out of the biggest memory chunks since it was last queried. x$ksmls x$ksmmem This 'table' seems to allow to address (that is read (write????)) every byte in the SGA. Since the size of the SGA equals the size of select sum(value) from v$sga, the following query must return 0 (at least on a four byte architecture. Don't know about 8 bytes.) select (select sum(value) from v$sga ) - (select 4*count(*) from x$ksmmem) "Must be Zero!" from dual; x$ksmsd x$ksmsp x$ksmsp_nwex x$ksmspr x$ksmss x$ksolsfts x$ksolsstat x$ksppcv x$ksppcv2 Contains the value kspftctxvl for each parameter found in x$ksppi. Determine if this value is the default value with the column kspftctxdf. x$ksppi This table contains a record for all documented and undocumented (starting with an underscore) parameters. select ksppinm from x$ksppi to show the names of all parameters. Join indx+1 with x$ksppcv2.kspftctxpn. x$ksppo x$ksppsv x$ksppsv2 x$kspspfile x$ksqeq x$ksqrs x$ksqst Enqueue management statistics by type. ksqstwat: The number of wait for the enqueue statistics class. ksqstwtim: Cumulated waiting time. This column is selected when v$enqueue_stat.cum_wait_time is selected. The types of classes are: BL Buffer Cache Management CF Controlfile Transaction CI Cross-instance call invocation CU Bind Enqueue DF Datafile DL Direct Loader index creation DM Database mount DP ??? DR Distributed Recovery DX Distributed TX FB acquired when formatting a range of bitmap blocks far ASSM segments. id1=ts#, id2=relative dba FS File Set IN Instance number IR Instance Recovery IS Instance State IV Library cache invalidation JD Something to do with dbms_job JQ Job queue KK Redo log kick LA..LP Library cache lock MD enqueue for Change data capture materialized view log (gotten internally for DDL on a snapshot log) id1=object# of the snapshot log. MR Media recovery NA..NZ Library cache pin PF Password file PI Parallel slaves PR Process startup PS Parallel slave synchronization SC System commit number SM SMON SQ Sequence number enqueue SR Synchronized replication SS Sort segment ST Space management transaction SV Sequence number value SW Suspend writes enqueue gotten when someone issues alter system suspend|resume TA Transaction recovery UL User defined lock UN User name US Undo segment, serialization WL Redo log being written XA Instance attribute lock XI Instance registration lock XR Acquired for alter system quiesce restricted x$kstex x$ksull x$ksulop x$ksulv x$ksumysta x$ksupr x$ksuprlat x$ksurlmt x$ksusd Contains a record for all statistics. x$ksuse x$ksusecon x$ksusecst x$ksusesta x$ksusgif x$ksusgsta x$ksusio x$ksutm x$ksuxsinst x$ktadm x$targetrba x$ktcxb The SGA transaction table. x$ktfbfe x$ktfthc x$ktftme x$ktprxrs x$ktprxrt x$ktrso x$ktsso x$ktstfc x$ktstssd x$kttvs Lists save undo for each tablespace: The column kttvstnm is the name of the tablespace that has saved undo. The column is null otherwise. x$kturd x$ktuxe Kernel transaction, undo transaction entry x$kvis Has (among others) a row containing the db block size: select kvisval from x$kvis where kvistag = 'kcbbkl' x$kvit x$kwddef x$kwqpd x$kwqps x$kxfpdp x$kxfpns x$kxfpsst x$kxfpys x$kxfqsrow x$kxsbd x$kxscc x$kzrtpd x$kzspr x$kzsrt x$le Lock element: contains an entry for each PCM lock held for the buffer cache. x$le can be left outer joined to x$bh on le_addr. x$le_stat x$logmnr_callback x$logmnr_contents x$logmnr_dictionary x$logmnr_logfile x$logmnr_logs x$logmnr_parameters x$logmnr_process x$logmnr_region x$logmnr_session x$logmnr_transaction x$nls_parameters x$option x$prmsltyx x$qesmmiwt x$qesmmsga x$quiesce x$uganco x$version x$xsaggr x$xsawso x$xssinfo A perlscript to find x$ tables #!/usr/bin/perl -w use strict; open O, ("/appl/oracle/product/9.2.0.2/bin/oracle"); open F, (">x"); my $l; my $p = ' ' x 40; my %x; while (read (O,$l,10000)) { $l = $p.$l; foreach ($l =~ /(x\$\w{3,})/g) { $x{$_}++; } $p = substr ($l,-40); } foreach (sort keys %x) { print F "$_\n"; } Obviously, it is also possible to extract those names through x$kqfta =============== 27 OTHER STUFF: =============== 27.1 How to retrieve DDL from sqlplus: ======================================= Use DBMS_METADATA.GET_DDL() Examples: SELECT dbms_metadata.get_ddl('TABLE','EMPLOYEE','RM_LIVE') from dual; SQL> set pagesize 0 SQL> set long 90000 SELECT dbms_metadata.get_ddl('TABLE', table_name, 'RM_LIVE') FROM DBA_TABLES WHERE OWNER = 'RM_LIVE' and table_name like 'CDC_%'; SELECT dbms_metadata.get_ddl('INDEX, table_name, 'RM_LIVE') FROM DBA_TABLES WHERE OWNER = 'RM_LIVE' More on this procedure: If there is a task in Oracle for which the wheel has been reinvented many times, it is that of generating database object DDL. There are numerous scripts floating in different forums doing the same thing. Some of them work great, while others work only until a specific version. Sometimes the DBAs prefer to create the scripts themselves. Apart from the testing overhead, these scripts require substantial insight into the data dictionary. As new versions of the database are released, the scripts need to be modified to fit the new requirements. Starting from Oracle 9i Release 1, the DBMS_METADATA package has put an official end to all such scripting effort. This article provides a tour of the reverse engineering features of the above package, with a focus on generating the creation DDL of existing database objects. The article also has a section covering the issue of finding object dependencies. Why do we need to reverse engineer object creation DDL We need them for several reasons: Database upgrade from earlier versions when for various reason export-import is the only way out. But huge databases would require a precreated structure - importing data with several parallel processes into individual tables. Moving development objects into production. The cleanest method is to reverse engineer the DDL of the existing objects and run them in the production. For learning the various parameters that an object has been created with. When we create an object, we do not specify all the options, letting Oracle pick the defaults. We might want to view the defaults that have been picked up, or we might want to crosscheck the parameters of the object. For that we need Enterprise Manager, Toad, or some other tool, or self-developed queries in the data dictionary. Now DBMS_METADATA get the clean complete DDL with all options. Modes of usage of the Metadata Package A set of functions that can be used with SQL. This is known as the browsing interface. The functions in the browsing interface are GET_DDL, GET_DEPENDENT_DDL, GET_GRANTED_DDL A set of functions that can be used in PLSQL, which is in fact a superset of (1). They support filtering, and optional turning on and turning off of some clause in the DDL. The flexibilities provided by the programmer interface are rarely required. For general use the browsing interface is sufficient - more so if the programmer knows SQL well. Retrieving DDL information by SQL As mentioned in the section above, GET_DDL, GET_DEPENDENT_DDL and GET_GRANTED_DDL are the three functions in this mode. The next few sections discuss them in detail. The objects on which the examples are tested are given in Table 9. GET_DDL The general syntax of GET_DDL is GET_DDL(object_type, name, schema, version, model, transform). Version, model and transform take the default values "COMPATIBLE", "ORACLE", and "DDL" - further discussion of these is not in the scope of this article. object_type can be any of the object types given in Table 8 below. Table 1 shows a simple usage of the GET_DDL function to get all the tables of a schema. This function can only be used to fetch named objects, that is, objects with type N or S in Table 8. We will see in a later section how the "/" at the end of the DDL can be turned on by default. Table 1 (DBMS_METADATA.GET_DDL Usage) SQL> set head off SQL> set long 1000 SQL> set pages 0 SQL> show user USER is "REVRUN" SQL> SQL> select DBMS_METADATA.GET_DDL('TABLE','EMPLOYEE')||'/' from dual; CREATE TABLE "REVRUN"."EMPLOYEE" ( "LASTNAME" VARCHAR2(60) NOT NULL ENABLE, "FIRSTNAME" VARCHAR2(20) NOT NULL ENABLE, "MI" VARCHAR2(2), "SUFFIX" VARCHAR2(10), "DOB" DATE NOT NULL ENABLE, "BADGE_NO" NUMBER(6,0), "EXEMPT" VARCHAR2(1) NOT NULL ENABLE, "SALARY" NUMBER(9,2), "HOURLY_RATE" NUMBER(7,2), PRIMARY KEY ("BADGE_NO") USING INDEX PCTFREE 10 INITRANS 2 MAXTRANS 255 STORAGE(INITIAL 65536 NEXT 1048576 MINEXTENTS 1 MAXEXTENTS 2147483645 PCTINCREASE 0 FREELISTS 1 FREELIST GROUPS 1 BUFFER_POOL DEFAULT) TABLESPACE "SYSTEM" ENABLE ) PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 NOCOMPRESS LOGGING STORAGE(INITIAL 65536 NEXT 1048576 MINEXTENTS 1 MAXEXTENTS 2147483645 PCTINCREASE 0 FREELISTS 1 FREELIST GROUPS 1 BUFFER_POOL DEFAULT) TABLESPACE "SYSTEM" / GET_DEPENDENT_DDL The general syntax of GET_DEPENDENT_DDL is GET_DEPENDENT_DDL(object_type, base_object_name, base_object_schema, version, model, transform, object_count) Version, model and transform take the default values "COMPATIBLE", "ORACLE" and "DDL", and are not discussed futher. object_count takes the default of 10000 and can be left like that for most cases. object_type can be any object of type D in Table 8. base_object_name is the base object on which the object_type objects are dependent. The GET_DEPENDENT_DDL function allows the fetching of metadata for dependent objects with a single call. For some object types, other functions can be used for the same effect. For example, GET_DDL can be used to fetch an index by its name or GET_DEPENDENT_DDL can be used to fetch the same index by specifying the table on which it is defined. An added reason for using GET_DEPENDENT_DDL in this case might be that it gives the DDL of all dependent objects of that base object and the specific object type. Table 2 shows a simple usage of GET_DEPENDENT_DDL. Table 2 (GET_DEPENDENT_DDL example) SQL> column aa format a132 SQL> SQL> select DBMS_METADATA.GET_DEPENDENT_DDL('TRIGGER','EMPLOYEE') aa from dual; CREATE OR REPLACE TRIGGER "REVRUN"."HOURLY_TRIGGER" before update of hourly_rate on employee for each row begin :new.hourly_rate:=:old.hourly_rate;end; ALTER TRIGGER "REVRUN"."HOURLY_TRIGGER" ENABLE CREATE OR REPLACE TRIGGER "REVRUN"."SALARY_TRIGGER" before insert or update of salary on employee for each row WHEN (new.salary > 150000) CALL check_sal(:new.salary) ALTER TRIGGER "REVRUN"."SALARY_TRIGGER" ENABLE GET_GRANTED_DDL The general syntax of GET_GRANTED_DDL is GET_GRANTED_DDL(object_type, grantee, version, model, transform, object_count) Version, model and transform take the default values "COMPATIBLE", "ORACLE" and "DDL", and need no further discussion. object_count takes the default of 10000, and can be left like that for most cases. grantee is the user who is granting the object_types. The object types that can work in GET_GRANTED_DDL are the ones with type G in Table 8. Table 3 shows a simple usage of the GET_GRANTED_DDL function. Table 3 (GET_GRANTED_DDL Usage) SQL> set long 99999 SQL> column aa format a132 SQL> select DBMS_METADATA.GET_GRANTED_DDL('OBJECT_GRANT','REVRUN_USER') aa from dual; GRANT UPDATE ("SALARY") ON "REVRUN"."EMPLOYEE" TO "REVRUN_USER" GRANT UPDATE ("HOURLY_RATE") ON "REVRUN"."EMPLOYEE" TO "REVRUN_USER" GRANT INSERT ON "REVRUN"."TIMESHEET" TO "REVRUN_USER" GRANT UPDATE ON "REVRUN"."TIMESHEET" TO "REVRUN_USER" Table 4 below classifies some common objects as Dependent Object (D), Named Object (N) or Granted Object (G). Some objects exhibit more than one such property. For a complete list, refer to the Oracle Documentation. However, the list below will meet most requirements. Metadata information retrieval by programmatic interface The programmatic interface is for fine-grained detailed control on DDL generation. The list of procedures available for use in the programmatic interface is as follows: OPEN SET_FILTER SET_COUNT GET_QUERY SET_PARSE_ITEM ADD_TRANSFORM SET_TRANSFORM_PARAM FETCH_xxx CLOSE To make use of this interface one must write a PLSQL block. Considering the fact that several CLOB columns are involved, this is not simple. However, the next section shows how to use the SET_TRANSFORM_PARM function in SQLPLUS in order to perform most of the jobs done by this interface. If one adds simple SQL skills to it, the programmatic interface can be bypassed in almost all cases. To get details of the programmatic interface, the reader should refer to the documentation. Using the SET_TRANSFORM_PARAM function in SQL Session This function determines how the output of the DBMS_METADATA is displayed. The general syntax is SET_TRANSFORM_PARAM(transform_handle, name, value). transform_handle for SQL Sessions is DBMS_METADATA.SESSION_TRANSFORM name is the name of the transform, and value is essentially TRUE or FALSE. Table 4 shows how to get the DDL of tables not containing the word LOG in a good indented form and with SQL Terminator without a storage clause. Table 4 (SET_TRANSFORM_PARAM Usage) SQL> execute DBMS_METADATA.SET_TRANSFORM_PARAM(DBMS_METADATA.SESSION_TRANSFORM,'STORAGE',false); PL/SQL procedure successfully completed. SQL> execute DBMS_METADATA.SET_TRANSFORM_PARAM(DBMS_METADATA.SESSION_TRANSFORM,'PRETTY',true); PL/SQL procedure successfully completed. SQL> execute DBMS_METADATA.SET_TRANSFORM_PARAM(DBMS_METADATA.SESSION_TRANSFORM,'SQLTERMINATOR',true); PL/SQL procedure successfully completed. SQL> select dbms_metadata.get_ddl('TABLE',table_name) from user_tables 2 where table_name not like '%LOG'; CREATE TABLE "REVRUN"."EMPLOYEE" ( "LASTNAME" VARCHAR2(60) NOT NULL ENABLE, "FIRSTNAME" VARCHAR2(20) NOT NULL ENABLE, "MI" VARCHAR2(2), "SUFFIX" VARCHAR2(10), "DOB" DATE NOT NULL ENABLE, "BADGE_NO" NUMBER(6,0), "EXEMPT" VARCHAR2(1) NOT NULL ENABLE, "SALARY" NUMBER(9,2), "HOURLY_RATE" NUMBER(7,2), PRIMARY KEY ("BADGE_NO") USING INDEX PCTFREE 10 INITRANS 2 MAXTRANS 255 TABLESPACE "SYSTEM" ENABLE ) PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 NOCOMPRESS LOGGING TABLESPACE "SYSTEM" ; CREATE TABLE "REVRUN"."TIMESHEET" ( "BADGE_NO" NUMBER(6,0), "WEEK" NUMBER(2,0), "JOB_ID" NUMBER(5,0), "HOURS_WORKED" NUMBER(4,2), FOREIGN KEY ("BADGE_NO") REFERENCES "REVRUN"."EMPLOYEE" ("BADGE_NO") ENABLE ) PCTFREE 10 PCTUSED 40 INITRANS 1 MAXTRANS 255 NOCOMPRESS LOGGING TABLESPACE "SYSTEM" ; SQL> Thus we see how a DDL requirement even with some filtering condition and a formatting requirement was met by the SQL browsing interface along with SET_SESSION_TRANSFORM. Table 5 shows the name and meaning of the SET_SESSION_TRANSFORM parameters. Table 5 (SET_SESSION_TRANSFORM "name" Parameters) PRETTY (all objects) - If TRUE, format the output with indentation and line feeds. Defaults to TRUE. SQLTERMINATOR (all objects) - If TRUE, append a SQL terminator (; or /) to each DDL statement. Defaults to FALSE. DEFAULT (all objects) - Calling SET_TRANSFORM_PARAM with this parameter set to TRUE has the effect of resetting all parameters for the transform to their default values. Setting this FALSE has no effect. There is no default. INHERIT (all objects) - If TRUE, inherits session-level parameters. Defaults to FALSE. If an application calls ADD_TRANSFORM to add the DDL transform, then by default the only transform parameters that apply are those explicitly set for that transform handle. This has no effect if the transform handle is the session transform handle. SEGMENT_ATTRIBUTES (TABLE and INDEX) - If TRUE, emit segment attributes (physical attributes, storage attributes, tablespace, logging). Defaults to TRUE. STORAGE (TABLE and INDEX) - If TRUE, emit storage clause. (Ignored if SEGMENT_ATTRIBUTES is FALSE.) Defaults to TRUE. TABLESPACE (TABLE and INDEX) - If TRUE, emit tablespace. (Ignored if SEGMENT_ATTRIBUTES is FALSE.) Defaults to TRUE. CONSTRAINTS (TABLE) - If TRUE, emit all non-referential table constraints. Defaults to TRUE. REF_CONSTRAINTS (TABLE) - If TRUE, emit all referential constraints (foreign key and scoped refs). Defaults to TRUE. CONSTRAINTS_AS_ALTER (TABLE) - If TRUE, emit table constraints as separate ALTER TABLE (and, if necessary, CREATE INDEX) statements. If FALSE, specify table constraints as part of the CREATE TABLE statement. Defaults to FALSE. Requires that CONSTRAINTS be TRUE. FORCE (VIEW) - If TRUE, use the FORCE keyword in the CREATE VIEW statement. Defaults to TRUE. DBMS_METADATA Security Model The object views of the Oracle metadata model implement security as follows: Non-privileged users can see the metadata only of their own objects. SYS and users with SELECT_CATALOG_ROLE can see all objects. Non-privileged users can also retrieve object and system privileges granted to them or by them to others. This also includes privileges granted to PUBLIC. If callers request objects they are not privileged to retrieve, no exception is raised; the object is simply not retrieved. If non-privileged users are granted some form of access to an object in someone else's schema, they will be able to retrieve the grant specification through the Metadata API, but not the object's actual metadata. Finding objects that are dependent on a given object This is another type of requirement. While dropping a seemingly unimportant table or procedure from a schema one might like to know the objects that are dependent on this object. The data dictionary view DBA_DEPENDENCIES or USER_DEPENDENCIES or ALL_DEPENDENCIES is the answer to these requirements. The columns of the ALL_DEPENDENCIES view are discussed in Table 6. ALL_DEPENDENCIES describes dependencies between procedures, packages, functions, package bodies, and triggers accessible to the current user, including dependencies on views created without any database links. Only tables are left out of this view. However for finding table dependencies we can use ALL_CONSTRAINTS. The ALL_DEPENDENCIES view comes to the rescue in the very important area of finding dependencies between stored code objects. Table 6 (Columns of ALL_DEPENDENCIES table) Column Description ------ ----------- OWNER Owner of the object NAME Name of the object TYPE Type of object REFERENCED_OWNER Owner of the parent object REFERENCED_NAME Type of parent object REFERENCED_TYPE Type of referenced object REFERENCED_LINK_NAME Name of the link to the parent object (if remote) SCHEMAID ID of the current schema DEPENDENCY_TYPE Whether the dependency is a REF dependency (REF) or not (HARD) Table 7 below shows how to use the above view to get the dependencies. The example shows a case where we might want to drop the procedure CHECK_SAL, but we would like to find any objects dependent on it. The query below shows that a TRIGGER named SALARY_TRIGGER is dependent on it. Table 7 (Use of the ALL_DEPENDENCIES view) SQL> select name, type, owner 2 from all_dependencies 3 where referenced_owner = 'REVRUN' 4 and referenced_name = 'CHECK_SAL'; NAME TYPE OWNER ------------------------------ ----------------- ---------------------- SALARY_TRIGGER TRIGGER REVRUN CONCLUSION This article is intended to give the minimum effort answer to elementary and intermediate level object dependency related issues. For advanced object dependency issues, this article points to the solution. As Oracle keeps on upgrading its versions, it is clear that they will be upgrading the DBMS_METADATA interface and ALL_DEPENDENCIES view along with it. The solutions developed along those lines will persist. Table 8 (Classifying common database objects as Named, Dependent, Granted and Schema objects) CONSTRAINT (Constraints) SND DB_LINK (Database links) SN DEFAULT_ROLE (Default roles) G FUNCTION (Stored functions) SN INDEX (Indexes) SND MATERIALIZED_VIEW (Materialized views) SN MATERIALIZED_VIEW_LOG (Materialized view logs) D OBJECT_GRANT (Object grants) DG PACKAGE (Stored packages) SN PACKAGE_SPEC (Package specifications) SN PACKAGE_BODY (Package bodies) SN PROCEDURE (Stored procedures) SN ROLE (Roles) N ROLE_GRANT (Role grants) G SEQUENCE (Sequences) SN SYNONYM (Synonyms) S SYSTEM_GRANT (System privilege grants) G TABLE (Tables) SN TABLESPACE (Tablespaces) N TRIGGER (Triggers) SND TYPE (User-defined types) SN TYPE_SPEC (Type specifications) SN TYPE_BODY (Type bodies) SN USER (Users) N VIEW (Views) SN Table 9 (Creation script of the REVRUN Schema) connect system/manager drop user revrun cascade; drop user revrun_user cascade; drop user revrun_admin cascade; create user revrun identified by revrun; GRANT resource, connect, create session , create table , create procedure , create sequence , create trigger , create view , create synonym , alter session TO revrun; create user revrun_user identified by user1; create user revrun_admin identified by admin1; grant connect to revrun_user; grant connect to revrun_admin; connect revrun/revrun Rem Creating employee tables... create table employee ( lastname varchar2(60) not null, firstname varchar2(20) not null, mi varchar2(2), suffix varchar2(10), DOB date not null, badge_no number(6) primary key, exempt varchar(1) not null, salary number (9,2), hourly_rate number (7,2) ) / create table timesheet (badge_no number(6) references employee (badge_no), week number(2), job_id number(5), hours_worked number(4,2) ) / create table system_log (action_time DATE, lastname VARCHAR2(60), action LONG ) / Rem grants... grant update (salary,hourly_rate) on employee to revrun_user; grant ALL on employee to revrun_admin with grant option; grant insert,update on timesheet to revrun_user; grant ALL on timesheet to revrun_admin with grant option; Rem indexes... create index i_employee_name on employee(lastname); create index i_employee_dob on employee(DOB); create index i_timesheet_badge on timesheet(badge_no); Rem triggers create or replace procedure check_sal( salary in number) as begin return; -- Demo code end; / create or replace trigger salary_trigger before insert or update of salary on employee for each row when (new.salary > 150000) call check_sal(:new.salary) / create or replace trigger hourly_trigger before update of hourly_rate on employee for each row begin :new.hourly_rate:=:old.hourly_rate;end; SELECT substr(username, 1, 20), account_status, default_tablespace, temporary_tablespace, created FROM dba_users WHERE created > SYSDATE -10; ============= 11g Features: ============= Note 1: ------- Vraag: Wat zijn de belangrijkste nieuwe performance features in Oracle Database 11g? Antwoord: De drie Result Caches: SQL Result Cache, PL/SQL Function Result Cache en de OCI Client Result Cache. De SQL result Cache bewaart de uitkomst van een vaak uitgevoerd SQL query statement in de SGA. De Query Optimizer houdt zelf bij welke queries in aanmerking komen, rekening houdende met de DML en query frequentie. Met name queries op lookup tabellen hebben hier enorm veel profijt van. De PL/SQL Result Cache doet hetzelfde, maar dan voor PL/SQL Functions. De OCI Client Cache bewaart het query resultaat op de client zodat er geen netwerk trip naar de database nodig is voor geselecteerde SQL queries. Verder nog SQL Plan Management, een feature die SQL performance regressie voorkomt door executie plannen op te slaan in de database als basis voor toekomstige executie plannen. Als er in de toekomst een ander executie plan beschikbaar komt, doordat er bijvoorbeeld een index is aangemaakt, kan zo’n nieuw plan alleen geaccepteerd worden als het daadwerkelijk tot een betere performance leidt. Dus de SQL optimizer is hiermee zelflerend geworden. Vraag: Wat zijn de belangrijkste nieuwe Backup and Recovery features? Antwoord: De Data Recovery Advisor: in plaats van zelf te bedenken hoe je een Recovery vraagstuk het beste kunt aanpakken vraag je nu gewoon aan RMAN een advies en natuurlijk kan RMAN dat advies ook voor je uitvoeren. Dus 'advise failure' en ‘repair failure' is alles wat een Oracle 11g DBA-er hoeft te weten. En qua backup performance is de RMAN multisection backup een belangrijke verbetering die het mogelijk maakt om één file met meerdere channels te backuppen om intra-file parallelisme te bewerkstellen. Een grote time saver is de RMAN “duplicate from active database” feature die het mogelijk maakt om een database te dupliceren zonder dat daarvoor een opgeslagen backup nodig is. Vraag: Wat zijn de belangrijkste nieuwe security features? Antwoord: Tablespace encryptie is belangrijk als het om data protectie gaat. Hiermee is het mogelijk om de gehele inhoud van een Tablespace te encrypten ongeacht de gebruikte datatypen. Hiermee is de data niet alleen beveiligd binnen de database maar ook tegen aanvallen buiten de database om. Een andere belangrijke verbetering is het nieuwe password algoritme en de mogelijkheid om de DBA passwords onder te brengen in een LDAP server zodat ze centraal beheerd kunnen worden. Vraag: Wat zijn de belangrijkste nieuwe data opslag features? Antwoord: De nieuwe LOB implementatie is ronduit geweldig met een veel betere performance en ingebouwde encryptie mogelijkheid die voor iedereen met LOB’s in de database veel voordelen biedt. Maar ook de integratie van NFS in de database, Direct NFS, is een feature met veel performance voordelen en meer keuze vrijheid wat het onder liggende disk systeem betreft. De Oracle 11g database kan nu rechtstreeks met een NFS server praten zonder tussenkomst van de NFS layer uit het operating systeem. Als laatste zijn de vele nieuwe partitionerings methoden ronduit overweldigend en is alles zo ongeveer mogelijk wat men maar wensen kan. Misschien wel de meest belangrijke vorm van partitioneren, range partitioning, is geautomatiseerd met de introductie van interval partitioning, maar ook de virtuele partition key biedt vele nieuwe mogelijkheden.