Re: Database backups with ADSM

1997-12-16 09:08:38
Subject: Re: Database backups with ADSM
From: Daniel Thompson <thompsod AT USAA DOT COM>
Date: Tue, 16 Dec 1997 08:08:38 -0600

  We have tried similar things in the past to allow applications to kick
off backups of their files during a batch processing cycle.

  If you use the same node name with different exclude/include statements
specified in different dsm.opt files, you will not get good backups.  When
an incremental runs it will expire those files that have backups but now
have exclude statements.  The good news is that you use less space on the
ADSM server (ha ha, sorry).

We got around this by using different node names on the host.  In my
opinion, this is not ideal as it adds steps to full server or full drive
recoveries.  I don't really mind the extra work, just the possibility that
a step may be missed.  It would be nice if ADSM would support a new type of
exclude statement that would not kick off expiration of the data, but just
not back the files up.

To specify the use of a different .opt file, the answer is slightly
different depending on the platform and how you are starting the backup.
For NT, the following answers apply:

If using multiple ADSM services, each with a different options file,
install the services and then edit the entry pointing to the DSM.OPT for
appropriate service(s).

If running on NT (or OS/2) in batch files, set the DSM_CONFIG and DSM_DIR
environment variables at the beginning of the batch file to point to the
appropriate locations.  For example: DSM_DIR=D:\ADSM

I hope this helps,
Dan T.

> From: Buechler, Becky <buechler AT KCC DOT COM>
> From: Buechler, Becky <buechler AT KCC DOT COM>
> Subject: Database backups with ADSM
> Date: Monday, December 15, 1997 12:35 PM
> I am running the ADSM V2.l.013 server on MVS.  I want to start using
> ADSM to back up databases directly to ADSM using scripts.  Today we
> schedule all backups on the ADSM server and guess at a time when the
> actual database backups are done. We then exclude the actual database
> files but back up the "backups" of the databases which are done to disk.
> An example of this is our HPUX server running Oracle.
> One of my questions is how do you run an incremental backup to point to
> another dsm.inclexcl file? I want to continue to run my normal backup of
> this system excluding all database files so I need to put the excludes
> in the dsm.inclexcl file. If I do this and then try to do the backup
> script for the database files it won't look at those excluded files.
> Also I would like to know if anyone decided to use two separate nodes
> for this kind of backup and why? I am interested in anyone's experience
> with doing these types of backups for databases not yet fully supported
> by ADSM directly.
> TIA,
> Becky
<Prev in Thread] Current Thread [Next in Thread>