Video Screencast Help

Media allocation/use NBU 6.0MP5 on Solaris 10

Created: 14 May 2010 | 8 comments

I have a situation where backups and or image copies when intially started, allocate an Active/Avail media from the proper volume pool for that host/policy but when initial media becomes full all subsequent media allocations for job grab a new scatch tape from scratch pool and none of the many other avail/active media within the pool? Any thoughts why? I was under impression that NB would grab all required tapes for backup job from AVAIL media within it's volume pool and not just the initlal tape.

PS all meida same type, retention all the same, unlimited mounts, etc etc.

Comments 8 CommentsJump to latest comment

J.H Is gone's picture

Are the tapes frozen or suspended?

I don't have to know how to spell....I work on Unix.
NetBackup 7.0.1 - AIX & Windows

Steve Cassandra's picture

nope all show as Active and like i said no problem grabbing them if its first tape allocated for job just all subsequent tapes once intial tape gets Full then taken from scratch pool  instead of from one of the many other Avail/Active tapes in pool.

Marianne's picture

Do you have more than one media server? Media servers don't share media (not until 6.5 with media sharing enabled).

Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links

Andy Welburn's picture

From my experience NetBackup will not continue a backup on a tape that already has images on it. It always appears to prefer an empty tape as opposed to one thats already been used but is still active.

Can you confirm those active tapes you've identified in the non-scratch pool already have images on?

You may also want to look at retention periods - NetBackup does not mix retentions on tape by default (but it can, altho' I wouldn't recommend it). This would mean that a new job would possibly not start on an existing tape if images if the retentions differed.

You may want to look at using the "Maximum Number of Partially Full Media" setting for the volume pool - you may be able to use this to cut down on part-full tapes & 'force' NetBackup to use these active tapes.

Marianne's picture

My experience has actually been that NetBackup will try to append to existing (assigned) media containing backups that's in the same pool, the same retentions period, not FULL/FROZEN/SUSPENDED, belongs to the same media sever and is in the robot, before using unassingned media (from same pool or scratch).

This is what Admin Guide II says about media selection when using a robot:

NetBackup selects a volume in a robot in the following order:
■ NetBackup searches the media catalog for a volume that is already mounted in a drive and meets the following criteria:
■ Configured to contain backups at the retention level that the backup schedule requires. However, if the NetBackup Media host property Allow multiple retentions per media is specified for the server, NetBackup does not search by retention level.
■ In the volume pool that is required by the backup.
■ Of the same density that is required by the requested backup and in the robot that is requested by the backup.
■ Not currently in use by another backup or a restore.
■ Not written in a protected format. NetBackup detects tape format after the volume is mounted. If the volume is in a protected format, NetBackup unmounts the volume and resumes the search.

If a suitable volume is found, NetBackup uses it.
■ If NetBackup cannot find a mounted volume that satisfies all of the previous conditions, it checks the media catalog for any volume that is suitable.
■ If a suitable volume is in a robot, NetBackup issues the commands that move the volume to a drive, position the heads to the beginning of the volume, and assign it to the request. No manual intervention is required.
■ If a suitable volume is not in a robot but is in a stand-alone drive, NetBackup automatically mounts and assigns it. No manual intervention is required.

If a suitable volume is not in a robot or a stand-alone drive, NetBackup does one of the following:
■ Pends a mount request if the request is media-specific (restore, import, or tpreq).
■ Attempts to to use another volume (for backup jobs where any other media can be used).
If the media catalog does not have a suitable volume or if a suitable volume is at end of media (EOM), a new volume is assigned. NetBackup may assign a new volume even if a volume is not full (because NetBackup received an EOM message from the drive). The new volume must meet all of the following criteria:
■ Is the correct media type.
■ Is for the correct robot type (if applicable).
■ Is located in the requested robotic peripheral (if applicable).
■ Resides on the requested host.
■ Is in the correct volume pool.
■ Is not currently assigned (not already allocated to NetBackup).
■ Is not expired (if an expiration date is defined in NetBackup).
■ Has not exceeded the maximum number of mounts allowed.
■ If more than one volume qualifies, NetBackup chooses the volume that was least recently used. NetBackup then adds it to the media catalog and assigns it the specified retention level.

Supporting Storage Foundation and VCS on Unix and Windows as well as NetBackup on Unix and Windows
Handy NBU Links

Andy Welburn's picture

I only see this behaviour when NB starts a new batch of scheduled jobs.

If it fills a tape part way thru' a 'batch' it has always gone for a new, empty tape as opposed to a partially full one. I know this really used to bug me as every now & then we'd end up with two or three partially full (ACTIVE) media for each volume pool & the number would never drop (NB always rotated between them & then start a new tape leaving the other part-fulls to be used next time).

We haven't got max no' of part full media set so maybe that's one possible explanation for our observations?

Steve Cassandra's picture

I give up. Wasted too much time on this. Obviously some sort of bug and I think that's why I keep seeing people referring to 6.5 "Maximum Number of Partially Full Media" parameter which 6.0 apparently does not have. So maybe Symantec created that parameter to fix this issue.
Over all summary to everything above and just to make sure I clarified, is that I have a Volume Pool based on retention in this case 1yr. I have numerous policies/servers that use this volume pool as they all require 1 year retentions. There is only one media server. The volume pool has up to 40+ partially active media/tapes some with as little as 20GB. None are frozen/suspended etc. There obviously not all already assigned to other jobs.  Most if not all the policies allow multiplexing. When any job initiates, it appears to grab a tape out of the 1 yr volume pool as expected. Once it fills that tape and if it requires any more from that point, it grabs empty one from scratch pool instead of any of the 40+ active tapes in its own pool. So these 40+ active tapes for the most part sit around wasting media resource.IT's a never ending cycle and the amount of active tapes never seems to decrease to a what I would expect to be a normal level of just a couple.

Andy Welburn's picture

Volume Pool? i.e. actually physically loaded and being written to?

When we had a similar issue I just took out all the ACTIVE tapes from the library but left enough in for NB to start its next lot of jobs. Those that were removed were left to expire gracefully & then loaded back in to the library as scratch. That way I managed to get the number of ACTIVE media for that pool down to a more 'realistic' level.