Netbackup6.5.2 and Agent (SAP and ORA) and inline copies: catastrophic failure.

[jim_dalton]

Just to let folks know I've encountered a feature, verified by Symantec (though it could be a config issue), whereby running inline copies of SAP or ORA agent based backups results in acute failure of Netbackup. The jobs start and sit queued, never moving. Nothing else can run, some jobs dont even get as far as the activity monitor.

Restarting netbackup has no effect; clearing down ipcs related resources has no effect. Only a reboot fixes the problem. Standard (flat file) inline copies are fine, single SAP or ORA agent backups also fine.

Of course, if you know a fix to this, I'm very interested. 

Solaris10/Single master plus one media server/Solaris 10 client/6.5.2,6.5.2A (at least) 

Thanks,Tim 

[Omar_Villa]

This is known issue with nbpem fixed on 6.5.2a.

[jim_dalton · Accepted solution]

 I do have the solution: the policies were taken from a fully functioning 5.1 server but contain an option that whilst benign in 5.1 is catastrophic in 6.5:we had retentions defined for both copies in the 'Automatic Full Backup' whereas the doc tells you to just have the one retention (retention in this case refers to how long netbackup will keep a record of what backup ran, not the retention of backup images). Having a single retention in the Automatic fixed the problem with both SAP and Oracle inline copy backups.