cancel
Showing results for 
Search instead for 
Did you mean: 

incr rman agent oracle backup fails but full completes successfully

Gene_B_
Not applicable

I have an incremental oracle backup running on red hat linux 2.6.18 (client sftw vers 7.5.0.7 ) and the master ( no media server in between) is running solaris 5.1.0 ( sftw vers  7.5.0.7).  The stream fails with status 13 and the parent job fails with a 6.  The full runs successfully every time, and they are pretty much the same scripts.  Any suggestions would be appreciated.  Will add logs to post.

 

5 REPLIES 5

Nicolai
Moderator
Moderator
Partner    VIP   

Try adding CLIENT_READ_TIMEOUT = 3600 to bp.conf on client and master server. Netbackup does not need to be restarted.

http://www.symantec.com/docs/HOWTO33199

Will_Restore
Level 6

Check the RMAN .out file for clues.  It's probably taking a long time to gather what needs to be backed up, in which case you can increase the Client Read Timeout value on the Media Server.

 

Michael_G_Ander
Level 6
Certified

An option is also to implement Block Change Tracking on the Oracle database, so rman don't have to scan all the blocks to find the changed ones. Does require a restart of the Oracle and a new full (level 0) backup

CLIENT_CONNECT_TIMEOUT is usually also worth looking in these case

Hope this helps

The standard questions: Have you checked: 1) What has changed. 2) The manual 3) If there are any tech notes or VOX posts regarding the issue

jim_dalton
Level 6

How big is the DB? And how much change do you and your DBA anticipate? If its a big DB then Michaels suggeston comes into play as you can wait a very long time for the data to arrive which is why the timeout becomes significant. CBT/BCT reduces this wait, but applies to specific Oracle version eg 10g and beyond.

If its a small DB then if you have capacity it might be worth forgetting about incremental and just do full as the full can take less time. Less time indeed. And can be simpler to deal with on recovery/restore. Just run a few tests.

Jim

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Valuable advice so far but 6 days of no response from the OP....

Please help us understand what 'pretty much the same scripts' means, 
provide the rman output file,
provide client's dbclient log,
check Client Read and Connect Timeout on the media server
.

If dbclient log folder does not exist, create it under /usr/openv/netbackup/logs and give 777 permissions to the folder.