Catalog Backup and nbdb.log

[Dirk_Mueller]

Hi@ll,

I have NBU 7.1.0.4 on Win 2003-Servers.

Our generally problem is that a full-catalog-backups runs over 33 hours. The catalog is nearly 800 GB.

Our actually problem:

the last full-catalog-backup has started 27.05. 10:24 and has ended 28.05 19:43

Today in the morning NetBackup has hung, because the netbackup-partition was full.

The nbdb.log in the folder NetBackupDB/data was about 17,5 GB

After increasing the partition I have restarted netbackup and I have stated a incremential catalog backup.

After that the nbdb.log wandered over to NetBAckupDB/staging.

After a few minutes it was gone.

The inc-catalog-backup is still running.

Who can explain me the steps and reasons of a catalog-backup and the nbdb.log-staging???

Kind regards

Dirk

[Nicolai]

http://www.symantec.com/docs/TECH173231

The staging folder contain as copy of the relation databases during catalog backup. 

A catalog backup protect both the relation databases like EMM and the flat files located in /usr/openv/netbackujp/db. Thats's why a Catalog backup job always consist of multiple jobs.

Hope it explain the qustion.

Update:

Another T/N - It not related to the question (bug report ) but explain what the catalog is used for in-directly.

http://www.symantec.com/docs/TECH179524

[Dirk_Mueller]

Thanks Nicolai,

with your descriptions and some descriptions of the support I think I understand the steps of a catalog backup. It was not so easy, because our environment does not behave normally.

The support told me that 17,5 GB nbdb.log is normal for a large NB-environment.

And sometimes it happens that after a catalog backup the nbdb.log is not reduced.

Can you confirm that???

Regards

Dirk

[mph999]

After a successful Catalog backup the NBDB.log should always be reduced.  It gets truncated to '0' size, but, if the system is busy, it will increase in size quite fast, so by the time you look it is posisble that in fact it has got back to quite a large size.

THe catalog backup as Nicolai explains, is really quite simle : x4 jobs

1.  Parent job

2.  Stage NBDB database(s) to staging area

3.  Backup up database in staging area

4.  Backup image DB (/usr/openv/netbackup/db/images/* )

(A few other files get backed up as part of the above (like /usr/openv/netbackup/db/ ), but that is a 'simple' explanation.

Martin

[Nicolai · Accepted solution]

incremental catalog backup does not truncate the nbdb.log - only fulls.

[mph999]

Thanks for pointing that out Nicolai , you are correct ...

M