Re: win32: veritas embedded ms sql server huge memory hog

On 9/13/05, Branimir Petrovic <BranimirP_at_cpas.com> wrote:
>
> Paul,
> Just an idea - why not keep all Veritas services in stopped state all
> the time *except* for a period shortly before scheduled backup? This
> can easily be scripted and scheduled, you should also add post backup
> script (the very last thing Veritas does after the job gets done) that
> will perform hara-kiri (in form of small DOS batch or VBscript to stop
> all Veritas services).
> Not pretty, nor should be necessary (who on Earth wanted/requested
> database catalog of all backed up stuff - when flat file list stored
> on the tape itself would do it nicely) but the kludge should ensure
> forced release all the RAM Veritas would otherwise chew...
> Branimir
>

Branimir,

Great idea, thought of it myself.
Funny thing about backups is ... SysAdmins don't like to mess with them.

I can just see the conversation now between BOFHs:

"So this prima donna dba asked me if we could stop and start the backup software services as a function of time of day ... he won't be logging onto that server for awhile."

I'm guessing that wouldn't get very far.

The real resolution is for Veritas to ship a non-brain dead config file such that the embedded db inst doesn't grab the process limit of vm.

And by now poeple are probably very tired of me repeating that.

In other news, if you hit a bug using dbms_stats on win32 after upgrading from 9.2.0.5.3 to 10.1.0.4 <http://10.1.0.4> ... there is a fix included in 10.1.0.4 <http://10.1.0.4> patch 5.
I'm testing it now. (patch 4530022)

Paul  

-----Original Message-----
> *From:* Paul Drake [mailto:bdbafh_at_gmail.com]
> *Sent:* Tuesday, September 13, 2005 2:25 PM
> *To:* kevinc_at_polyserve.com
> *Cc:* Oracle-L
> *Subject:* Re: win32: veritas embedded ms sql server huge memory hog
>
>

-- 
#/etc/init.d/init.cssd stop
# f=ma, divide by 1, convert to moles.

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