Chris Beagle asks:
> Our site is the process of installing ADSM on a Sun Solaris system
> and in reading the Administrator's Guide on the subject of Reclamation
> of tapes there is no mention of Scheduling. Can you control, after a
> volume has reached the Reclamation Threshold, when the reclamation
> process is to begin, so to control the use of the tape drives. We
> only have two tape drives and don't want reclamation to occur during
> our backup or archive window.
ADSM's automatic functions, including (but not limited to) migration and
reclamation are kicked off by events, rather than by time of day. That is,
they kick off when they please, as soon as the high migration or reclamation
thresholds are reached.
Currently there is no way to specify a time-frame within which these
functions occur. However, you can sort of "kludge" it yourself. Just set
the threshold (migration or reclamation) to 100. This effectively disables
migration (or reclamation). Then, at the time of day you wish, kick off
a job or process to issue the UPDATE STGPOOL command, setting the migration
or reclamation threshold to the desired value. For example, if you reclaim
at 75% free, set the threshold to 75. This will immediately cause the
server to start performing migration or reclamation if the thresholds have
been exceeded. Later, run another job or process to reset the threshold
back to 100.
I use a variation on this type of scheduling to perform scheduled migration
of my disk backup storage pool. It can be inconvenient to have this start
whenever the high threshold is met, so I do the following:
* Maintain enough disk space to house one night's worth of backups.
* Set the thresholds to 10 and 90 (I don't use 100 in this case simply
because I don't want to wait for it to hit 100% in the middle of
backups. Normally it goes up to 70%, at the most.)
* At a convenient time of day (4PM for me) I run a batch job to set
the high threshold to 11. This immediately causes migration to
start.
* At 4:30PM I run another job to reset the threshold back to 90.
This prepares the disk backup pool to receive that night's backups without
migration running unexpectedly.
Andy Raibeck
Connecticut Mutual
|
