piątek, 20 maja 2011

When a metalink lies about bug fixes

OS: Oracle Enterprise Linux 5.4
Oracle: 10.2.0.5


I tried to run RMAN backup using the following command

osb3@ORCL2(09:47)[/home/oracle]$ rman target /

Recovery Manager: Release 10.2.0.5.0 - Production on Fri May 20 09:47:38 2011

Copyright (c) 1982, 2007, Oracle.  All rights reserved.

connected to target database: ORCL2 (DBID=3949336569)

RMAN> run {
backup database include current controlfile TAG FULLY
SKIP READONLY
SKIP OFFLINE;
}2> 3> 4> 5>

Starting backup at 2011-05-20 09:48:55
using channel ORA_DISK_1
skipping read-only file 5
skipping offline file 6
RMAN-06060: WARNING: skipping datafile compromises tablespace SOE_INDEX recoverability
RMAN-06060: WARNING: skipping datafile compromises tablespace SOE_INDEX recoverability
channel ORA_DISK_1: starting full datafile backupset
channel ORA_DISK_1: specifying datafile(s) in backupset
input datafile fno=00010 name=/u01/app/oracle/oradata/ORCL2/undotbs.dbf
input datafile fno=00001 name=/u01/app/oracle/oradata/ORCL2/system01.dbf
input datafile fno=00004 name=/u01/app/oracle/oradata/ORCL2/users01.dbf
input datafile fno=00003 name=/u01/app/oracle/oradata/ORCL2/sysaux01.dbf
channel ORA_DISK_1: starting piece 1 at 2011-05-20 09:48:55
RMAN-03009: failure of backup command on ORA_DISK_1 channel at 05/20/2011 09:49:30
ORA-19694: some changed blocks were not found in the change tracking file
continuing other job steps, job failed will not be re-run
channel ORA_DISK_1: starting full datafile backupset
channel ORA_DISK_1: specifying datafile(s) in backupset
including current control file in backupset
channel ORA_DISK_1: starting piece 1 at 2011-05-20 09:49:30
channel ORA_DISK_1: finished piece 1 at 2011-05-20 09:49:31
piece handle=/u01/app/oracle/ORCL2/backupset/2011_05_20/o1_mf_ncnnf_FULLY_6xd78bsv_.bkp tag=FULLY comment=NONE
channel ORA_DISK_1: backup set complete, elapsed time: 00:00:01
RMAN-00571: ===========================================================
RMAN-00569: =============== ERROR MESSAGE STACK FOLLOWS ===============
RMAN-00571: ===========================================================

RMAN-03009: failure of backup command on ORA_DISK_1 channel at 05/20/2011 09:49:30
ORA-19694: some changed blocks were not found in the change tracking file


The block change tracking is enabled.

SQL> select*from v$block_change_tracking;

STATUS    FILENAME                                            BYTES
---------------------------------------------------------------------
ENABLED   /u01/app/oracle/ORCL2/block_change_tracking.trc     11599872


On the metalink I found the note:
BLOCK CHANGE TRACKING GIVES ORA-19694 WHEN USING DIRECT LOAD [ID 300989.1]

It is written:


Likelihood of Occurrence 
~~~~~~~~~~~~~~~~~~~~~~~~ 
  This will affect those using SQL Direct Load and RMAN Fast Incremental
  Backups with Change Tracking enabled.
 
Possible Symptoms 
~~~~~~~~~~~~~~~~~ 
  RMAN-571: ===========================================================
  RMAN-569: =============== ERROR MESSAGE STACK FOLLOWS ===============
  RMAN-571: ===========================================================
  RMAN-3009: failure of backup command on ORA_DISK_1 channel at 02/14/2005
  ORA-19694: some changed blocks were not found in the change tracking file

When this error occurs the old incremental backups are suspect and should
not be kept/used for recovery. The incremental level 0 backups are not
affected as that is like a full backup and reads all blocks.
 
Workaround 
~~~~~~~~~~
  Disable Change Tracking

To disable change tracking, use this SQL statement:

  SQL> ALTER DATABASE DISABLE BLOCK CHANGE TRACKING;

Patches
~~~~~~~
  
  Fixed in Patchset10.1.0.5, and one-off patches are available 
  for 10.1.0.4 and 10.1.0.3.


I have 10.2.0.5 version with PSU 2.

SQL> select * from registry$history;

ACTION_TIME       ACTION NAMESPACE VERSION  ID COMMENTS       BUNDLE_SERIES
----------------- ------ --------- -------- -- -------------- -------------
11/02/03 08:37:29 APPLY  SERVER    10.2.0.5 2  PSU 10.2.0.5.2 PSU


osb3@ORCL2(12:07)[/u01/app/oracle/product/10.2.0.5/OPatch]$ ./opatch lsinventory
Invoking OPatch 10.2.0.5.1

Oracle Interim Patch Installer version 10.2.0.5.1
Copyright (c) 2010, Oracle Corporation.  All rights reserved.


Oracle Home       : /u01/app/oracle/product/10.2.0.5
Central Inventory : /u01/app/oracle/oraInventory
   from           : /etc/oraInst.loc
OPatch version    : 10.2.0.5.1
OUI version       : 10.2.0.5.0
OUI location      : /u01/app/oracle/product/10.2.0.5/oui
Log file location : /u01/app/oracle/product/10.2.0.5/cfgtoollogs/opatch/opatch2011-05-20_12-07-59PM.log

Patch history file: /u01/app/oracle/product/10.2.0.5/cfgtoollogs/opatch/opatch_history.txt

Lsinventory Output file location : /u01/app/oracle/product/10.2.0.5/cfgtoollogs/opatch/lsinv/lsinventory2011-05-20_12-07-59PM.txt

--------------------------------------------------------------------------
Installed Top-level Products (2):

Oracle Database 10g                                                  10.2.0.1.0
Oracle Database 10g Release 2 Patch Set 4                           10.2.0.5.0a
There are 2 products installed in this Oracle Home.


Interim patches (1) :

Patch  10248542     : applied on Thu Feb 03 08:35:01 CET 2011
Unique Patch ID:  13098962
   Created on 28 Dec 2010, 00:46:13 hrs PST8PDT
   Bugs fixed:
     6402302, 9713537, 8350262, 9949948, 8394351, 10327179, 8546356, 9711859
     9714832, 9952230, 10248542, 8544696, 9963497, 9772888, 8664189, 10249537
     7519406, 9952270, 8277300, 9726739


Of course the workaround works. The backup runs fine when I disabled the block change tracking.
But the bug should be fixed in 10.2.0.5
Once again, don`t trust what Oracle guys write in the Metalink notes.

Brak komentarzy:

Prześlij komentarz