There are also cases where more than 1 policy per client makes sense.
I've got 1 client where I have one policy per filesystem, and I'll
explain why.

NetBackup takes snapshots at the beginning of the *parent* job, not at
the beginning of each child job.  If you do a lot of FlashBackups,
you'll find that taking the snapshot right away but not starting the
backup until hours (or potentially days) later is really a dumb idea.
You tend to have a lot of snapshots lying around wasting disk and cpu
resources for absolutely no good reason.  You'll have to allocate
snapshot disk space that you don't even care about since your backup
hasn't started yet and that backup could subsequently fail with a 196
if you have too many backups queued up (yes, this is acceptable on
this client).

I logged a call with Symantec and they said that they would not change
this behavior.  So....1 policy per filesystem (for a subset of the
filesystems no less!) with coordinated access through a DSSU to
control how many can actually run.

I have other cases where different file systems may have different
scheduling requirements and that also forces you to split things up
between policies.

In general, we do not have a single policy per client though.
Databases and wannabe clusters without cluster-wide filesystems are
the common exception to this rule.

-- 
Ed Wilts, Mounds View, MN, USA
mailto:ewilts AT ewilts DOT org
_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu