Tape changes volume pool type

[pjtee]

Hi

I notice that my scratch tape has changed volume pool overnight after my backups have run. In the morning I have 3 Daily tapes when the night before I had two and the scratch. This is causing some problems now because some of my SQL backups will not run properly without a scratch tapes present. So I am getting "status 96" errors in the morning for these SQL jobs. Just wondering is it common for tapes to change volume pools automatically?


Thanks

[Andy_Welburn]

it will take it out of scratch & put it in whatever volume pool it requires, this is presumably what it would do anyway for your SQL jobs (take the tape out of scratch & put it in the pool that your SQL backups utilise)?

There could be ways 'around' this (e.g.max number of partially full media etc)- but would depend on what your requirements are & your environment/set-up/versions etc etc.

[pjtee]

Andy, thanks for the reply. Im still as a loss as to why its using my scratch tape. All of my backups combined should fit very easily on my two daily tapes, yet its spanning the 3 and I lose my scratch tape.

[Andy_Welburn · Accepted solution]

Or is it only partially filling them? If the latter it may be that it is using all of them concurrently (i.e. 3 drives at once). Maybe different retention periods for your backups (by default NB doesn't mix retention periods)

If the backups are filling the tapes you'll have to look at purchasing more!

***EDIT***

You can use bpmedialist (e.g. bpmedialist -U -m media_id) to determine status of your media (e.g. FULL, FROZEN, SUSPENDED or blank for ACTIVE).

A few things you may be able to change (depending on you requirements, environment & even NB version) if you wish to restrict or control the number of ACTIVE tapes used in a volume pool:

Storage Unit: max concurrent write drives, max streams/drive
Policy: limit jobs/policy, all multiple data streams
Schedule: media multiplexing
Volume Pool: maximum number of partially full media.

Some of these will not directly affect the number of active media utilised, but by cutting down the number of jobs or streams that are run concurrently may do so indirectly.