Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Mailing Lists -> Oracle-L -> RE: SMON - Killed process

RE: SMON - Killed process

From: Weerd de E.C. Kirsten <Kirsten.deWeerd_at_Oranjewoud.nl>
Date: Fri, 30 Jun 2000 16:44:37 +0100
Message-Id: <10544.110928@fatcity.com>


Steve,

        I know the job that got killed, and there are plenty of commits, just about every second for that matter. So I couldn't understand that would be a problem. Anyhow, I did a shutdown abort this morning, whilst shutdown immediate was hanging, and started the DB again, with no significant waiting at all ... So it doesn't exaxtly look like it needed any rollback at all ?

> -----Original Message-----
> From: Steven Monaghan [SMTP:Steve.Monaghan_at_wcom.com]
> Sent: Friday, June 30, 2000 3:42 PM
> To: ORACLE-L_at_fatcity.com
> Cc: Kirsten.deWeerd_at_Oranjewoud.nl
> Subject: RE: SMON - Killed process
>
> I'm no expert either, but you are out of luck either way. If Oracle needs
> to roll back that transaction, there is no "better" way to get rid of it.
> Even if you shutdown abort, on recovery it will still need to do that.
> Your
> best solution would be to find out who cretaed the job that was rolling
> back
> and make sure they don't do any significant job like that without
> intermittent commits.
>
> Steve Monaghan
> WorldCom
>
> > -----Original Message-----
> > From: root_at_fatcity.com [mailto:root_at_fatcity.com]On Behalf Of Weerd de
> > E.C. Kirsten
> > Sent: Friday, June 30, 2000 8:08 AM
> > To: Multiple recipients of list ORACLE-L
> > Subject: SMON - Killed process
> >
> >
> > Hi,
> >
> > We run Oracle 7.2.3 on HP-UX 10.20
> >
> > Each night, we run a script which kills running sessions, if any, on
> > HP-level. After that an shutdown immediate is performed an backup
> starts.
> > Everything worked fairly well.
> >
> > Last night, a hugh Oracle procedure was running and get killed by
> > the script. However, about 8 hours later the immediate shutdown
> > was not yet
> > performed. As far as I could tell, it was waiting for the
> > ORA_SMON process,
> > which was running at 90% cpu. Is SMON cleaning up the killed procedure
> ???
> >
> > Best solution I figure would be killing the ORACLE-session, rather
> > than killing the process on HP-level. Would that prevent events similar
> to
> > last night ? But how to do it ???
> >
> > Any hints for this constantly learning - I only learned about Oracle
> > about 2 years ago - they're all looking at me - junior would be grately
> > appreciated, as ever !!!
> >
> > Greets,
> >
> >
> > Kirsten
> > Living : Near Antwerp - Belgium
Received on Fri Jun 30 2000 - 10:44:37 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US