TSM Virtual Environments Collocation

[jmmartin]

Hello,

Does anyone have experience or suggestions about Collocating offsite copies of TSMVE data? The TSM documentation says to collocate by filespace but that seems like it would use a lot of tapes.

We have an offsite copypool that is collocated by group. Would it be appropriate to use this copy pool for TSMVE data? The node would be the datacenter node, so I would imagine that all filespaces under that Node would be collocated. It's 4.1 TB of data, so I assume that would be all collocated together.

Thanks!

joe

 

[moon-buddy]

The problem with TSM for VE is that the data is actually images.

If you don't collocate, you will end up with a restore nightmare like what I experienced last year. Aside from the fact that the restore is slowwww, the restores can only be done serially in a non-collocated mode.

You can try grouping VM Ware images by server groups but this also needs to be planned out well. You should consider group-by-tier restore priorities.

On the 'not so enticing side', doing it by filespace is a nightmare from a tape count perspective. Imagine an environment of 1000 VM Ware quest partitions which would translate to at least 1000 tapes!

The answer is a balaning act - a combinatgion of group and node collocation plus server tiering (how crtical is the server for the Business). You may also want to look at non-tape backups to speed restores.

 

[jmmartin]

Thanks for your experiences! I'll keep all that in consideration.