On Mon, 26 Nov 2001 15:02:45 +0100, it was written:
>We are backing up some large Novell servers (more than 1 million files per
>Server). The users want to be able to restore every version of a file 30
>days back and every version of a file at the end of a month 12 months back.
>In order to meet these requirements at least partly we set the backup
>management class as follows:
>
>Versions Data Exists: 10
>Versions Data Deleted: 5
>Retain Extra Versions: 366
>Retain Only Version: 366
>
>Because of these settings, the TSM database grew to 15 GB in the last 6
>months and is growing fast even further. Restore performance is bad for
>these servers as well.
>
>Is there a possibility / idea to get over this apart from using archives? Is
>it possible to create archives (not backupsets) from existing data on the
>TSM server?

No, archives require an additional pull of data across your network.

If your stated business requirements are set in concrete, then you
need to rethink the resources you have available to you. You will
probably need *very* fast disk access and (probably) a more potent
server, not to mention more library resource.

If business requirements are *not* set in concrete, then rethink your
prohibition on backupsets. They are quite a joy to create and use (as
long as you track them carefully).

I would bet a really good beer that none of your users will *ever*
need to retrieve a file that's more than a week old.

--
Mark Stapleton (stapleton AT berbee DOT com)
Mark Stapleton (stapleton AT berbee DOT com)