Greetings, all.  I'm hoping to engender a bit of discussion about how people
manage their TSM backups of DB/2; if folks feel this gets off-topic I'll try
to take it off the list.

At UF, we're in the process of deploying PeopleSoft, with DB/2 on AIX as our
database platform.  Peoplesoft implementations give rise to LOTS of databases,
and we've stuck each database in its own instance.  So here's what we've got
for backups:

Each database has a TSM node associated with it.

Full backups of the databases are kicked off to TSM on 'some periodic basis'.
We'll pretend it's weekly, but its actually more complex than that.

Transaction logs are stored (archived) in TSM by the DB/2 user exit.

>From time to time we might cut a different full backup, using a different TSM
node, if we want to retain it longer or shorter than our usual fulls.


We haven't gotten around to getting rid of any of the fulls yet. (heh.)

How do other DB/2 shops deal with (automatic?) deletion of old fulls?  Do you
have a DBA actually make a decision and run a command, do you roll something
locally that deletes 'old' versions periodically?

How do you handle restoring, and rolling forward, a database into a new
machine/instance?

What are some of the confusing tangles you've gotten into?

Backup/recovery war stories more than welcome.. :)


- Allen S. Rout