Arnaud Brion wrote:
> I'm facing an embarasting situation since I changed the way our daily job  
> runs : I start a reclamation job, by decreasing the rec parameter to 50 % on 
> all storage pools every morning at 3:00.
> Two hours later I start a backup stgpool job, on all storage pools too, and 
> what happens is that most of the backup stgpool processes are waiting for a 
> free drive to begin.
> This should never happen, because of the preemption system included in TSM, 
> that avoids this kind of painfull sitation : backup of stgpools has a higher 
> priority than reclamation, so why do I have this problem ? ( nopreempt 
> parameter is set to no)
> Did anybody face the same situation, and was it solved ?

There is documentation in the admin guides that show the priorities that
tape operations follow. Restores are at the top of the food chain,
followed by several others (including reclamation), with all backups
(including client and storage pool) at the bottom of the list. Whatever
is higher in the priority list will take precedence over those
operations that are lower.

This makes sense when you consider what is most important in a
backup/recovery system. Restores are the primary function, while cleanup
and bookkeeping jobs are not as important. As for backups, they can be
done anytime there is a drive available.

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