>> On Thu, 12 Jul 2007 09:50:25 -0500, Andrew Carlson <naclosagc AT GMAIL DOT
>> COM> said:
> Is this really a good answer? I know it's "working as designed", as
> IBM is sometimes fond of saying. TSM knows whats on what tapes. It
> seems like, unless the processing required to do it would be
> substantial, TSM could split the processes so this doesn't occur, at
> least as often. With non-collocated tapes, I can see that it could
> happen sometimes, but I would think TSM could be smart enough to
> minimize it.
A threshold would be nice: So I could express the idea
"If you've waited more than 10 minutes for a volume, then skip it, in
that case I prefer the 'skipped work' failure mode to the 'stalled
process' failure mode"
I've had as many as 4 tape drives idle, (2 processes, one reader one
writer each) waiting for an in-use volume, and sometimes it's been
hours before I noticed them.
- Allen S. Rout
|