1. How often are you doing a full Database backup? I think that is where
you need to look for the answers to yor questions. If you do a full backup
every day, then you probably can get by with three day retention. BTW, the
reuse delay of 3 days from TMS will not gain you anything, since you deleted
the backups from ADSM already. If you really want to keep backups a week,
then delete them in 6 days, not 3.
2. As for the rest of your tapes in the vol-hist file, I would think that
the same number of days as your DB backups. This file is used if you need to
recreate - so info prior to your oldest full backup is not of much value.
BTW - we see deadlocks on this file as well. **** IBM **** how big a problem
is this - do we need to open a PMR?
We have a setup very similar to yours, but we keep files for 3 weeks - I do a
full backup once a week; incrementals every day, and send the DB and
incremenals of site each week. This is a key point to remember - how does
your DR plan fit into the backup strategy?
Hope this helps.
Jerry Lawson
jlawson AT thehartford DOT com
______________________________ Forward Header __________________________________
Subject: VOLH Question
Author: INTERNET.OWNERAD at SNADGATE
Date: 5/13/97 9:18 AM
We are running ADSM V2 on MVS (level 12) using CA-TLMS for our tape
management. Since we want to let TLMS manage our tapes, and we want
colocation by node, we set our max scratch volumes to 10,000. Was
this a mistake?
Since ADSM is registered to TLMS as an External Data Manager, all
our tapes are created as permanent (no expiration). We have a reuse
delay of 3 days. We also issue a DEL VOLH TOD=TODAY-3 TYPE=DBB every
day.
While the DBB entries are being cleared in the sequential volume
history, STGD records are added every day. It looks as though ADSM is
trying to be my tape library manager and keeps a permanent record of
ALL tapes, even if it no longer owns them.
Can I safely issue a DEL VOLH TOD=TODAY-3 TYPE=STGD command every
day? Also, since ADSM has already signalled TLMS to scratch these
tapes at some prior time, will it delete the records without
signalling TLMS to scratch (as it does for DBB records)?
We are experiencing DATABASE LOCK CONFLICTs on the sequential
history file and I can't help but think that the extra entries are not
helping the situation. We have about 20,000 records in the sequential
history file, with only about 1500 of them reflecting an active tape.
Any help would be appreciated!
|