I'm assuming the number of filling tapes isn't increasing, it's higher than
you want, but stable. yes?

As i'm sure you are aware, this is "normal" behavior.  I've also been there
and tried to fight it.  In that instance I ran a job after the migration
that did a move data on the smallest volumes.

You're fighting against something you're not going to win though,  they
will keep coming back.

If you need collocation groups and you are in this situation, the best
approach is to procure extra library capacity.  Easier said than done I
know, but a MUCH easier approach.

Steven

On 12 March 2012 23:32, Roger Deschner <rogerd AT uic DOT edu> wrote:

> I'm having a problem with FILLING tapes multiplying out of control,
> which have very little data on them.
>
> It appears that this happens at the end of migration, when that one last
> collocation group is being migrated, and all others have finished. TSM
> sees empty tape drives, and less than the defined number of migration
> processes running, and it decides it can use the drives, so it mounts
> fresh scratch tapes to fill up all the drives. This only happens when
> the remaining data to be migrated belongs to more than one node - but
> that's still fairly often. The result is a large number of FILLING tapes
> that contain almost no data. A rough formula for these almost-empty
> wasted filling tapes is:
>
>  (number of migration processes - 1) * number of collocation groups
>
> Is there a way, short of combining collocation groups, to deal with this
> problem? We've got a very full tape library, and I'm looking for any
> obvious ways to get more data onto the same number of tapes. Ideally,
> I'd like there to be only one FILLING tape per collocation group.
>
> Roger Deschner      University of Illinois at Chicago     rogerd AT uic DOT 
> edu
>               Academic Computing & Communications Center ======I have
> not lost my mind -- it is backed up on tape somewhere.=====
>