Larry,
You may want to take some additional steps to reign in the volume usage by
the Groupwise nodes. Because the transient files are all uniquely named
(at least as far as we could tell), they are retained according to the
RETONLY parameter of the MC that they are bound to. And since they aren't
on the node any more, they can't be rebound to a new MC with a different
setting (0 would expire them immediately).
In the case of the central dedicated GW server that was threatening to take
over the tape library, we wound up moving it to a policy domain we had
created for this purpose. The MC that these files were backed under had
RETONLY set to 5, all other MCs were identical to the original domain.
We then ran an expire inventory manually:
ANR0812I Inventory file expiration process 682 completed: examined 534319
objects, deleting 533910 backup objects, 0 archive objects, 0 DB backup
volumes, and 0 recovery plan files. 0 errors were encountered.
After that, we moved the node back to its "normal" domain. In the case of
the smaller remote multiple-use GW servers, we elected to just let the
transient files expire off "naturally" since there was much more danger of
expiring something that might be needed later (like 10 minutes after it was
gone).
An added benefit from this work was that the db usage went from close to
80% to below 50%. (For a while, anyway...)
Good luck!
Ted
