Re: RMAN on Windows Server

  • From: "Rich Jesse" <rjoralist@xxxxxxxxxxxxxxxxxxxxx>
  • To: "Oracle-L@xxxxxxxxxxxxx" <oracle-l@xxxxxxxxxxxxx>
  • Date: Tue, 1 Dec 2009 10:22:19 -0600 (CST)

Hey Robert,

> Are you using Database control or grid control? They will automate this for
> you using the Oracle Scheduler. Of course, if your shutting down your
> database from time to time then the database scheduler would not be able to
> run the backups.

I'd say this option is "USE AT YOUR OWN RISK".  I chose it to get RMAN
going, but there is a severe issue with Grid Control jobs.  Unfortunately
all my SR information is now lost since the upgrade to MOS, so I can't give
any specifics, although patch 759579 is what I applied to hopefully fix it. 
The problem is that the scheduled jobs in Grid Control can silently
disappear.  If I hadn't noticed my archive_dest suspiciously filling up on
the weekend, it could have been major trouble.

I'm in the delicate process of replacing all RMAN backups in Grid Control
with a POSIX shell script run locally on each server.  Much more robust and
reliable, IMNSHO.  One can't modify *any* aspect of the RMAN script once
it's scheduled.  The job must be dropped and recreated.  How absurd!

Unfortunately I have one Winders XE DB that I'd still like backed up.  And
since I'm a solo DBA, I can't take the time to rewrite a backup script
specific for this ancillary DB, so Grid Control it is for this guy.  I've
supplemented the job with yet ANOTHER "Check to make sure Grid Control is
running" script utilizing the output from a "repvfy verify jobs" command of
the EMDiag kit.  And even the version of EMDiag somehow broke a few weeks
ago, where I needed to upgrade it.  Granted, I'm using this diag tool well
beyond what it was meant to do, but at this point I don't another viable
option.

Really really really (REALLY!) try to avoid Grid Control for production
database backups.

Really.

My $.02,
Rich


--
//www.freelists.org/webpage/oracle-l


Other related posts: