Scratch pool assigning more tapes than necessary

[Chris_Tobolski]

Hello!

I would like to preface this by warning that I am a novice concerning Netbackup, and am doing my best to learn. I appreciate any help/information on this issue.

I have recently configured a scratch pool because I was occasionally getting 96 errors on a specific policy, and was told by a support representative that a best practice is to let the Scratch pool assign tapes rather than manually assigning them.

I have a seperate policy that has written a slightly less than 6TB in the current week. This amount of data is typically stored on 3 tapes per week. However, since creating the Scratch pool, 7 tapes are assigned to this policy this week, meaning that 4 have been allocated from the Scratch pool.

I read this article on how tapes are assigned: http://www.symantec.com/business/support/index?page=content&id=HOWTO32854

I'm thinking maybe some of the volumes are at the end of media state, perhaps. Or, after reading this tech article http://goo.gl/x8qHRh, that maybe NetBackup is trying to store entire jobs on one tape rather than conserve space.

Do either of these seem plausible? Why is the scratch pool allocating so many tapes for this policy?

Thanks,

Chris

[Dan_42]

Can you tell us a little more about your environment, especially hardware? Do you have a tape library/robot with multiple drives, for instance? 

[SymTerry]

NetBackup will try to use a whole tape, till the tape library reports full. Also are you using tape compression? The amount of data will depend on the type of data and compression rate. 

[Nicolai]

Chris - The best advice I can give to you is accepting Netbackup pulls tape from the scratch pool. I you start controlling when tape is being retrieved you are looking into loosing battle because Netbackup isn't built to control what single tape are being written. Netbackup assumes all tapes can be used for the job and try to arrange it as effective as possible.

But I can assure you Netbackup will fill tapes up, when they are allocated from the scratch pool (sooner or later). Do don't worry - using a scratch pool is the right ting to do.

[Chris_Tobolski · Accepted solution]

Thank you all for your posts. I am a student worker and am only here a few days a week, so please forgive me for not responding quickly.

As it turns out, before the previous backup admin went on vacation, he sent the retention levels from 60 days to 90 days, which overlaps with our rotation. So the tapes were full of unexpired information, which I did not discover until last Friday.

So it makes sense that the scratch pool assigned so many tapes.

Thanks again for your responses.

Chris T.