Eric Sturdivant <[EMAIL PROTECTED]> wrote:
> We have a modified butc which works on solaris with modern TSM
> clients.  Each of our fileservers runs butc processes to dump
> volumes to the TSM server. This works just fine. The problem we
> have is in doing entire server or partition restores, butc will
> restore each volume in sequence, causing the TSM server to
> thrash around in the tapes, getting each little diff from all
> the volumes. Co-location has helped the problem a bit, cutting
> down on the number of tape mounts, but not the seeks through
> the tape.
> 
> The only thing we can think to do at this point is to throw
> disk at the problem, and keep as much of the dumps on disk as
> possible. (This, of course, is expensive...)

As far as I know the people around here that uses TSM with AFS do
a 'vos dump' to put the volume on a stagingdisk and then take a
backup of the disk.

We do the same but use NetBackup.

'butc' destroyed our first tapedrive as it was not fast enough to
get the tape streaming.

_______________________________________________
OpenAFS-info mailing list
[email protected]
https://lists.openafs.org/mailman/listinfo/openafs-info

Reply via email to