>I thought about that, but would that help? If TSM still has to scan every
>object for a match, it wouldn't help much... That's the problem, I don't know
>how SQL works...

Eric - Your perception is correct: if you scan a table, it will traverse the
       whole thing.  Whereas the Backups table is the predominant (=huge)
table in a TSM system, it will take a long time.  Some optimization can be
had through well-formulated queries, but the opportunities for doing that are
rather rare.  The only thing that really helps SQL performance is indexing,
where short, key columns are also kept in a hash.  Whereas TSM SQL is an
overlay on a B-tree database, I don't believe there is any indexing
opportunity, and so SQL scans are painful.

Sometimes, the best thing to do is perform Query Backup from the client side,
where dedicated logic gets results faster.  It is often possible to accomplish
that by masquerading as each defined TSM node, via VIRTUALNodename.
Another approach to finding flotsam, of course, is to inspect the last backup
time in filespaces, which helps narrow down the search arena.

   Richard Sims