Hi,
My solution to this problem was to write a series of scripts which managed
all the daily operations (backup to copy pools, reclamation, etc.). The
scripts kick off the various operations and then only let them run for a
specific interval. At the end of the interval, the operation is cancelled
with a "cancel process".
My scripts aren't documented and I don't know how useful they would be to
anyone. They are written largely in Korn shell and awk and are very, very
ugly. In addition, there's alot of cruft which is site-specific - a
registration package, email notification of client failure to back up file
systems, and a lot of stuff to manage tape volumes written before DRM (I now
use DRM a little bit in the scripts).
I found the adsm administrative scheduler to be of limited value - it could
kick off operations, but not monitor their progress or cancel them.
To manage limited resources (such as a server with only two tape drives) you
need a full-featured language with variables, branches and loops.
Just as an aside - I think that an adsm server should have a minimum of four
tape drives.
Jim
> ----------
> From: Michael L. Gimm[SMTP:mlg AT FOURLEAF DOT DK]
> Reply To: ADSM: Dist Stor Manager
> Sent: Thursday, March 18, 1999 5:25 AM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Reclamation blocks backup (migration)
>
> Hi all
>
> One of my customers have a ploblem. Sometimes the reclamation process uses
> his 2 mountpoints all night, so his migration just waits all night for
> mountpoints.
> This results in failed backups because the clients timeout when
> encountering too many "Server out of storage space".
>
> What are the possibilities here - Prioritize?
>
>
> Best regards
> Four Leaf Technologies A/S, Viby J.
> Michael L Gimm
> Counsultant
>
> E-Mail : mlg AT fourleaf DOT dk
> Direct : +45 87 38 60 54
> Phone : +45 70 25 45 00
> Fax: : +45 70 25 45 01
> Cell: : +45 28 19 60 54
>
|