Re: Semaphores on W2K

drak0nian_at_yahoo.com (Paul Drake) wrote in message news:<1ac7c7b3.0406012100.17539be5_at_posting.google.com>...
> joel-garry_at_home.com (Joel Garry) wrote in message news:<91884734.0406011340.3b774ef4_at_posting.google.com>...
> > drak0nian_at_yahoo.com (Paul Drake) wrote in message news:<1ac7c7b3.0406011042.436a6f2c_at_posting.google.com>...
> > > srivenu_at_hotmail.com (srivenu) wrote in message news:<1a68177.0405312139.5ede6ce7_at_posting.google.com>...
> > > > Have a look at this note
> > > > http://metalink.oracle.com/metalink/plsql/ml2_documents.showDocument?p_database_id=FOR&p_id=411387.999
> > > >
> > > > regards
> > > > Srivenu
> > >
> > > Srivenu,
> > >
> > > thanks for the link. I didn't see any cases of this issue for 9.2.x.
> > >
> > > Certainly, "rebooting the server" is not a fix imo.
> > > Seems to me link another reason to get off of 8.1.7.4.x and onto
> > > 9.2.0.
> > >
> > > We haven't moved off of 9.2.0.4 (testing 9.2.0.5) due to some wrong
> > > results (bug) in complex queries. Seems to be a fairly common problem,
> > > as there is already a patchset out that has a fix for a similar case.
> > >
> > > I wonder how many people are thinking about skipping the migration
> > > from 8.1.7 to 9.2 and going straight to 10.x after the first patchset
> > > is out.
> > >
> > > Paul
> >
> > Thanks for the interesting info, guys. Although none really addresses
> > the case of a single instance being rebooted, I can make wild guesses
> > that Oracle hadn't quite figured out mutex timing issues.
> >
> > jg
>
> Joel,
>
> are you absolutely certain that the oracle process was killed?
> One of the metalink forums articles listed having to stop the listener
> also, that a semaphore could have hung around off of the listener, not
> just off of the oracle process.
>
> if you're not running MTS^Wshared server then stopping/restarting the
> listener might not be so bad.

Well, I can't imagine them not being killed on a server reboot, but then again, Windows does things beyond imagining.

jg

--
@home.com is bogus.
http://www.stirlingengine.com/