I was sure that I had disabled the media mount timeout and restarted
bprd and bpdbm before I posted to the list, but it occured to me later
that the bpbrm still governing the supspended/incomplete jobs probably
still retained the old timeout in its job config. After completely
shutting down netbackup and resuming the incomplete jobs, they and all
subsequent jobs turned out fine. Lesson learned: don't forget the backup
processes not just the daemons. Don't know why the media mount timeout
was set (to 300), I was sure that I had disabled the timeout.. either
way, I'll give it another few nights to see if this was the real
problem. If you don't hear from me on this again, consider this a SUMMARY :)
Thanks, all who replied!
Robert Cantu
Consultant, Perfect Order
Veritas Certified Data Protection Specialist
robert.cantu at perfectorder.com
http://www.perfectorder.com
Tim Hoke wrote:
>Hey Robert,
>
>I'm assuming that the customer must have a media mount timeout actually
>set?
>
>If so, set it to zero and see what ultimately results. The timeout is
>probably just masking the real problem.
>
>Is it always the same tape/drive?
>
>Those are the places to start looking.
>
>HTH
>-Tim
>
>On Wed, 25 Aug 2004, Robert Cantu wrote:
>
>
>
>>Hi all, first time poster
>>
>>We have a customer with an L700 with 4 IBM LTO2 native fibre drives,
>>v880 (fully configured) running NBU 5.0 MP3 as a single master/media
>>server. We moved to MP3 to fix the spurious error 69's at the end of
>>some windows backups. The problem we're having is that for a single tape
>>at a time, the drive is loaded and I show that it's trying to position
>>the tape to file ###. At this point, something times out and I get error
>>52, "timed out waiting for media manager to mount volume". bptm just
>>says that bpbrm gave it the kill signal and internally reports error 82
>>"media manager killed by signal". This puts all backups writing to the
>>tape (usually with multiplexing 6 or 4 depending on full/incr) into
>>incomplete. Usually I can retry it after a while and it's fine, but
>>recently, it's consistently failing, and I can count on at least one
>>group of jobs failing every policy.
>>
>>I'm not sure what's going on here, as the drives are loading and
>>mounting. I can't find a "positioning" timeout, as that's what I think
>>is happening. Any thoughts please? Please CC my email address, I only
>>get the digests from the list.
>>
>>Robert Cantu
>>Consultant, Perfect Order
>>Veritas Certified Data Protection Specialist
>>robert.cantu at perfectorder.com
>>http://www.perfectorder.com
>>
>>_______________________________________________
>>Veritas-bu maillist - Veritas-bu AT mailman.eng.auburn DOT edu
>>http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu
>>
>>
>>
>
>
>
|
