We are facing the same problem.    What we though of doing is using
a seperate pool for snapshot backup using archive command and using copy
storage to move the data in the pool to a DR pool and move all the tape
in the DR pool to offsite.
    However, before we implemented, there are two major concerns i.e.
the ADSM's database will increased significantly as ADSM is doing file
level archive and not copy volume or disk image.  The second concern is
the archive command does't support exclude para, but not all files can
be backup online especially those system files, this causes lot of
errors in the log file and almost all the job status reflected as failed
and it is very difficult for us to check every error messages in the log
file to determine whether the backup set is a good copy.

Hilton Tina wrote:

> Archives are not incremental by nature.  You should look at the
> archive
> command and the output from it's execution.  Eyeball this with the
> directories and files on your system.  I bet there's a problem with
> the
> command and not all your files were included.
>
> Tina Hilton
> Thomson Consumer Electronics
> hiltont AT indy.tce DOT com
>
> > ----------
> > From:         Wayne
> > Merchant[SMTP:Wayne_Merchant/FALLSCHURCH/VBHNotes AT VBH DOT COM]
> > Reply To:     ADSM: Dist Stor Manager
> > Sent:         Friday, October 24, 1997 1:26 PM
> > To:   ADSM-L AT VM.MARIST DOT EDU
> > Subject:      Monthly, Quarterly, and Yearly Historical Backups
> >
> > We are required to maintain historical information for numerous
> > entities
> > and reasons. We also need this to be stored off site.
> >
> > We have been trying to get a month end " snapshot" of our data to no
>
> > avail.
> > ADSM support has advised us to do an archive.
> > The first time run we captured 300gb. After our second monthly
> archive
> > was
> > only 150gb, support indicated the archives were probably
> incrementals
> > and
> > suggested we run an export node. They also advised us that we could
> > only
> > run this once a month. We have tried to complete an export node
> > several
> > times only to have some problem occur that failed the job. (System
> > crashes,
> > and loss of power and such.)
> >
> > It would seem reasonable that backup requirements and processes that
>
> > have
> > occured for decades should be easily handled by ADSM.
> >
> > Is anyone generating these types of backups for historical off site
> > storage??
> >
> > Wayne
> >