The first job to complete will probably attempt to bring the DB back up,
too.

You have to write a stream-aware set of scripts for DB startup & shutdown.

I do mine with a series of lock files & counters.  The number of streams is
an environment variable & so I have a lockfile.  For startup, each stream
checks for the lock file, creating it & starting the DB preparation if it is
the first to start.  The loser of the race starts to watch for a completion
flag that the first stream will write when the DB is all the way down - then
they stop waiting & end to allow the backup to proceed.

The end-job is pretty much the same in reverse.  The last stream to complete
has to restart the DB.

One quick way around this, since you've got RMAN is drop the Netbackup
scheduling & run the job from  cron on the DB client.  It's all driven from
RMAN script that way.

HTH - Mark

-----Original Message-----
From: Steven L. Sesar [mailto:ssesar AT mitre DOT org]
Sent: Wednesday, June 11, 2003 8:24 AM
To: List, VERITAS NBU
Subject: [Veritas-bu] NBU/RMAN Cold Backup question


Hi,

When I kick off a cold backup, NBU breaks the backup up into several 
jobs. The problem with this, is that each new job that starts, tries to 
shutdown/startup mount the SID that it is backing up.
 
Has anyone else encountered this, and what have you done about it?

NBU 4.5MP4, Solaris 8.

--Steve