Backups Hung & NBU Services Need To Be Restarted?

[T3chBallad]

Hi Guru's,

Im not too familiar with veritas nbu.

Currently im facing an issue with the backups being hung.

I identifiy it as hung when a backup which usually takes e.g 2minutes, actually runs for more than 10 hours.

Despite that, more backups queues than the one which are active.

Usually when a backup hungs, i will restart the nbu services on the master server.

But it effects the other backups which are running ok.

My concern is, now i want to try restart the services of the media servers.

Is there any comands on unix platforms  to identify that the nbu services needs a restart or in any other way.

Thank you!

[Marianne]

Ensure that bptm and bpbrm log directories exist on all media servers. Each backup stream have a unique pair of bpbrm and bptm processes that are represented by the PID's in the log files (in square brackets [ ] ) and the process table (bpps output) on the media server.

You should be able to identify the relevant PID's for the 'hanging' backup by searching for the start of backup for particular job and follow this PID through the log file to see what's happening.
bpbrm will contain comms info with client and bptm will record each full data buffer received from client.
You might also need to troubleshoot the client being backed up and verify network transfer rate and processes running on the client.

Which NBU version and which OS?

[Yogesh9881]

you can use nbrbutil -dump from below location on unix box (media server)

/usr/openv/netbackup/bin/admincmd/./nbrbutil -dump

for more about nbrbutil navigate below URL

http://www.symantec.com/business/support/index?page=content&id=HOWTO42528

https://www-secure.symantec.com/connect/forums/understanding-nbrbutil-dump-output

[watsons · Accepted solution]

A hung job can be a bug in Netbackup (nbrb or nbjm) or when the server is too busy with thousands of jobs, this is especially common in huge SLP duplications.

For your case, trying to identify whether the hung is just particular to a certain client, or just a media server.

You don't have to restart NBU on master server if not everything is hanging. Let say you have identified jobs hang only for a media server, you can do the following (in master server) to clear the nbrb queue for that media server:

\Netbackup\bin\admincmd\nbrbutil -resetMediaServer <mediaServerHostname>

If that does not help, you can then restart only NBU service on that media server only, without having to restart NBU on master server.