On Wed, 8 Mar 1995 14:36:11 -0700 88130 said:
>I have a question regarding collocation.  We incrementally backup 20+
>machines nightly that are running a flavor of Unix (CLIX) not natively
>supported by ADSM.  This is done by NFS mounting their drives, one
>machine at a time, to an AIX machine (AIXA), and then running a backup
>of them from AIXA.  AIXA then owns the backup information for the CLIX
>machines.  The backups of AIXA are run by userid root to an MVS server
>and collocation is enabled.  Since the backups are all run from AIXA
>using one userid, collocation causes backups for the 20+ machines
>to be grouped together on tape.  This is messy when a restore to one
>of the CLIX machines is required due to lots of tape mounts.  The
>question is this:  if we used a unique userid in backing up each of
>the CLIX machines from AIXA, would collocation then spread the files
>across different tapes based on userid, or will it keep them together
>because they are all from AIXA?
>
>Jim White
>Salt River Project
>Tempe, Arizona
>jwwhite AT srp DOT gov

Jim;  I'm pretty sure that using unique userids will not help you out.
However, with the current version of the ADSM client, it is possible to
specify the nodename of the client, either in dsm.sys, dsm.opt, or on the
command line.  Therefore, you could actually register each of your 20+
systems as individual ADSM nodes.  When backing up each individual system,
specify the nodename for that system when running the ADSM client.  This
will keep the backup data for each node separate from each other, and will
allow collocation to work as you want.  When restoring or retrieving files,
you will now have to make sure to specify the correct nodename when invoking
the ADSM client.
..Paul

Paul Zarnowski                     Phone:   607/255-4757
Cornell Information Technologies   Fax:     607/255-6523
Cornell University                 US Mail: 315 CCC, Ithaca, NY 14853-2601