From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On 
Behalf Of Gill, Geoffrey L.
>>From looking at the log below it looks like this backup failed 
>because the
>backup could not mount a tape. The server logs do show migrations and
>reclamation running at this time. The disk pool itself is set to a size
>limit of 8GB.
>
>Could the reason be because, either the tape was in use during 
>a reclamation
>process or because all drives were in use at the time the 
>backup ran? Is
>there a way for me to see if all drives were in use at that 
>time? It looks
>like TSM tries 3 times and fails instead of bumping a process, 
>is this a correct assumption?

You could determine which drives had tapes mounted in them, but it would
be a matter of piecing through the activity log looking for mounts and
dismounts, which can be laborious. (This is one of those times that
running DSMADMC -MOUNTMODE comes in handy.)

Backups are the bottom of the job hierarchy with TSM; in other words,
all other tape-related operations will take precedence over a backup if
the backup is going to tape. (Restores, OTOH, are the top of the
hierarchy and will preempt enough processes to get tape drives to
perform its job.)

A database deadlock is almost always caused by the same entry in the
database being accessed by two (or more) different processes. (This is a
reason why you don't want to run multiple concurrent MOVE NODEDATA or
EXPORT NODE processes.) You probably had this situation going on--one
more good reason not to run anything else during reclamations and
inventory expirations.

--
Mark Stapleton