October 11, 2012

A few troubles when doing PSU patching on AIX RAC

when help a friend patching an AIX Rac server, i got a few troubles, record them here.

1. opatch failed and reporting:

Required amount of space(23428.225MB) is not available.
UtilSession failed: 
Prerequisite check "CheckSystemSpace" failed.
Log file location: /oracle/app/11.2.0/grid/cfgtoollogs/opatch/opatch2012-10-07_18-26-07PM_1.log
OPatch failed with error code 73
The mount point have 20gb free space but still we get the error.
For this bug we can use below parameter to skip space check when using opatch:

2. opatch failed and report:
Patching component oracle.rdbms,

Copy failed from '/oracle/ rver/13919095/files b bclsra11.so' to '/oracle/app/11.2.0/grid b bclsra11.so'...  
Please verify all applications associated with the Oracle Home '/oracle/app/11.2.0/grid' are shut down. If this is AIX, please perform solution documented in Note 739963.1 on https://myoraclesupport.oracle.com.

Do you want to retry copying the file? [y|n]

opatch reports a file is in use while currently from fuser command we can see there is no process holding the file.
It is a known issue in AIX, and the fix is to run below command as root:

4. opatch failed and report:
.storage is locked by another opatch session.
This is because another opatch activity just terminated due to the third issue and left its lock file t there.
The solution is remove *lock* file under $ORACLE_HOME/.storage

5. Finally patch succeed, but when starting grid service on second node, failed with below error:
2012-10-07 13:23:18.069: [    AGFW][5912] {0:0:31} Created alert :  (:CRSAGF00123:) :  Failed to start the agent process:  /oracle/grid/ Category: -2 Operation: setpcred Loc:  spawnproc26 OS error: 2 Other : child couldn't setpcred

On metalink there is only one bug report for same error messsage and same case as us----CRS failed to startup after patching.

No workaround.


Martin Klier said...

More details and a poor mans workaround:

Post a Comment