Segmentation fault - DB Restore

[ttrinh7975]

Has anyone ran into this problem?

I am getting Segmentation fault when attempts to perform DB restore on AIX 5.3ML3
TSM server code 5.4.0.2. DB restore was successful on some day but not others.
Tivoli is stump on this issue as well. They have escalated to level 3 support.

TSM DB restore goes for a little over an hour and then the "Segemenation Fault" came
up.
ANR4643I Processed 262144 log records.
ANR4643I Processed 266240 log records.
ANR4643I Processed 270336 log records.
ANR4643I Processed 274432 log records.
ANR4642I Sequential media log undo pass in progress.
Segmentation fault

 

[moon-buddy]

Every time I see segmentation faults happening, it brings me to two possible conclusions: 1) Software corruption - the OS or application, and 2) Hardware, in this case, a memory problem - ECC bit-error, or sync errors.

Might want to ask for IBM service (or yourself if you have the diagnostic CDs) to run a full (hardware) diagnostics of your TSM server.

 

[ttrinh7975]

moon-buddy-

Thanks for the respond. The DB restore work some of the day and not the other.
I have 7days worth of DB backup tapes. DB restore sucessful okay every other 2 days.
On the same hardware and OS and TSM server version.

The mksysb was created so the environment is exact the same.

 

[heada]

When it fails, is there anything in common? (same location in the restore, same tape, same tape drive, same time of day?) You're on ML03, have you tried updating to TL06? From what I remember, ML03 was kinda buggy and TL04CSP was ideal...but TL06 seems to be good as well.

-Aaron

 

[ttrinh7975]

heada-

Yes, the restore DB failed seem to failed at the same place. Of course on different tape on different day. I just confirmation from System Admin that she did the AIX OS upgrade from ML 03 to 06. We are currently running v5.3.0.6

 

[heada]

I would assume theres nothing in errpt worth looking at.

Are the restores failing on ML03? Normally when you restore a mksysb in a NIM setup, NIM will upgrade the ML to the latest in the lppsource. If the mksysb is ML03 and the lppsource is ML06 you'll restore the ML03 and then before the final reboot it will do an update_all from the lppsource. If thats what happened, you might want to try restoring the mksysb using a lppsource that is still ML03 to see if the upgrade is whats doing it. There are LOTS of changes from ML03 to ML06.

-Aaron