Reporting and Auditing Tapes

[TravoTon]

While I am lookin for information I thought I might post a question to the forums.

I am currently doing research on the best way to audit our tapes to see how much each policy is taking up.

I have found information using the command bpimmedia -policy <policy> -L

I have had a number of jobs fail lately and they had to be restarted.

Is there any way to discover what tape the failed jobs are on, if they have any other data on them, and is there a way to delete the unuseable data but maintain the good data?

[TravoTon]

From what I see in the forums you cannot delete an image without deleting everything else on the tape.

Is there a better way to determine what is on a tape?

[RamNagalla]

hello,

if the job got failled after writing some data on the tape.. netbackup will not keep track of it and it will simple overwrite with next job..

so there is nothing that we need to do manually to clear the failed job data from tapes..

[jim_dalton · Accepted solution]

What angle are you coming from regards "how much each policy is taking up"? If you want to report on data volumes, then opscenter is probably the place to go, that is its job. It is included in Netbackup. Then if you need to recharge that will give you some precise numbers.

Data on tape isnt really of much meaning since if you are compressing data (and you ought to be) then you dont know how each of your data streams compresses...some might get 1:1 , others 10:1.

If it really is about recharging, you could send different cost centres data to different medai pools and configure backup policies accordingly. Then it will be clear, but may affect backup windows and so forth.

Jim  

Sounds like folks are getting hung up on the number of tapes being used.  

[TravoTon]

We are backing up to a VTL (where compression happens) then from there copying to physical tapes for long term restores.

Our physical library is running out of tapes.

I have only recently aquired the position of backup administrator and am new to Netbackup.

I don't know why we have not used opscenter in the past but I will have to look into using that.

Thank you for your suggestion.