Hi,
  That's a odd-sounding story and doesn't seem quite right.  My site has a
DRM license, but we only use some of its volume-tracking features: we
developed our disaster recovery procedures before there was a DRM.  In our
case, we have two primary tape pools (one for archives, the other for
backups) and two copypools which backup up all our primary pools, both tape
and disk based.  One of the copypools we keep on site (we wait for tapes to
fill and then put them in a rack), the second we keep off site (if a single
byte is written to a tape in this pool we take it out of the library and
send it off site).  These offsite  copypool volumes, along with our database
backups and some other stuff (backups of our volume history file, device
configuration file, and filesystem backups and mksysb backups of our ADSM
server) are what we would use to recover the system in the event of a
disaster.  When we do a "query DRM" we see volumes from both copypools and
database backup volumes (of course DRM doesn't indicate which is what). In
that sense, what the IBM rep told you is right - DRM is for recovering the
whole library.

  If you have a copypool and if the tape in your primary pool was backed up
to the copypool, you should be able to do a restore volume (though you might
have to do some work to get the right copypool tapes loaded into your
library - if you set them to read/write or read only ADSM will ask for them
if they aren't in your library).

  If you don't have copypools then you aren't taking advantage of the
recovery capabilities of ADSM.  In that case, I don't know what is in your
DRM - maybe just database backup volumes.

  Of course, even following all the rules sometimes you can run into a
situation where you have a bad tape in a primary pool that didn't or
couldn't get copied and so you've lost data.  That's why a feature which
would allow us to force ADSM to write to multiple tape volumes
simultaneously (some people call this dual-copy mode) when storing data in
primary tape-based pools would be nice.

So to sum up - if you have copypools and use them, you ought to be able to
recover volumes in the primary pool unless you've had really bad luck.
> ----------
> From:         Scott Fluegge[SMTP:sfluegge AT VENATORGROUP DOT COM]
> Reply To:     ADSM: Dist Stor Manager
> Sent:         Monday, July 26, 1999 10:35 AM
> To:   ADSM-L AT VM.MARIST DOT EDU
> Subject:      Disaster-Recovery, Any value?
>
> OK,  I have had a minor disaster here which is making me fear the worst.
>
> We lost one of our Novell Servers and had to rebuild it.  This meant
> restoring
> it from ADSM.  The restore was started and was chugging along fine until
> it
> called for a tape that it couldn't find.  The restore died at that point.
> It
> turns out that the volume in question has disappeared.  Our librarians
> have no
> idea what happened to it.  No problem I thought.  I would just rebuild the
> tape
> using my disaster-recovery pool (We use ADSM's DR manager).
>
> I used the restore volume command but it came back saying the volume
> contained
> files that could not be restored.  After many calls to IBM (I finally got
> level
> 3 help), I was told that there was no way to restore a single volume, or
> even a
> single node from disaster-recovery media.  My data was simply lost and
> un-recoverable (even though there was a copy of the data in DR).  I was
> told
> that DR was only meant to rebuild an entire library and had no ties to
> files or
> nodes.   I was told that if we lost our on-site location and needed to
> rely on
> our DR tapes, the first thing we would have to do is re-create our entire
> library.  That we could not use the DR tapes directly, they would first
> have to
> rebuild the entire library.  I have more than 15 terrabytes of data being
> held.
> If I had to wait to build that it would take weeks before I could even
> start
> restoring nodes!!!
>
> The support person continued to tell me that for the future I should
> create a
> second storage pool, an on-site copy pool.  That would provide the only
> means
> for rebuilding tapes.  But I would need to purchase more than 1000 3590c
> tapes
> at around $100 each!  Not to mention the storage consideration.  We have 8
> frames in our 3494 and it is FULL! and keeping that much additional data
> would
> pound our database and dramatically reduce its response time.
>
> What are you all doing?  Does anybody use the DR manager and expect it to
> work
> the way I used to?  Am I missing something?  We are re-thinking our whole
> philosophy here and I could use some help and support!
>
> Thanks!!!!!
>
> Scott
>