Automatically check Netbackup for activity and jobs starting ?

[Michael_G_Ander]

Hello

We have had an issue where Netbackup was standing still, but all the services was up

We wonder what others do to automatically check Netbackup for activity and expected jobs starting up

Regards

Michael

[Marianne]

Maybe run 'nbpemreq' on a regular basis? Even put it in OS scheduler and email output to 'someone-who-cares'?

nbpemreq -predict_all -date <mm/dd/yyyy HH:MM:SS>

[Michael_G_Ander]

Thanks Marianne, but I don't understand how that caches netbackup stand still

[Marianne]

This is the problem with gathering too much info over the years - indexing system is not working!

I'm trying to find a TN - something about scheduled backups not starting automatically after abnormal shutdown of master... Even my Google skills are letting me down ...
Will be back if I find anything.

You did not mention your OS or NBU patch version?
Was there 'abnormal shutdown' of master recently?
Can manual backups be kicked off when this happens?

**** EDIT ***** 

Found the TN. It's quite old and was supposed to be fixed in 6.5.2, but I know for a fact that I've seen this in an early 7.x installation: http://www.symantec.com/docs/TECH61569

[Michael_G_Ander]

Did not mention versions as this was general inquiry about what people do to catch this siuation, where everything looks fine but the system is attucally not running any backups.

Our version is currently 7.5.0.4  on Windows 2008 R2 MSCS.

But the master didn't shutdown, because then it would have failed over to the other node.

[Marianne]

We run 'nbpemreq -predict_all ...' to ensure that our interpretation of schedules is the same as NBU's.

Else, check system logs and nbpem logs for errors.

[Lars_Hesdorf · Accepted solution]

This command gives all activity for last 5 minutes # bperror -all -U -d `gdate --date="-5 minutes" +"%m/%d/%Y %H:%M:%S"` Need to have GNU date command installed! If 5 minutes if too small then just inrease minut argument. At least this gives an idear of system is active or not. If nothing meaningfull comes from this command then I think situation with 99.99% confidence is that system is hung.

[Lars_Hesdorf]

And if need for creating an automated e-mail then embed the bperror command into

gtimeout 10s bperror -all -U -d `gdate --date="-5 minutes" +"%m/%d/%Y %H:%M:%S"`  >/dev/null 2>&1; if [[ $? == 0 ]] ; then .....

gtimeout yet another GNU command worth knowing.

That will also catch situation NetBackup not started at all and bperror not responding.