Re: be aware before thinking about deploy 11gR2 RAC
Date: Wed, 24 Feb 2010 07:44:37 -0600
Message-ID: <d4beff361002240544p2296b5f8q8f626b5ee9c3fe8d_at_mail.gmail.com>
Thanx for the info! This sounds very much like something we encountered recently on an 11gR2 Exadata machine... didn't pursue it at the time, because we were chasing down other major issues (both hardware and software as it turned out) and assumed it to be related.
In our case CRS would attempt to restart the database instances, but they would sometimes get stuck shortly after the mount phase completed... the alertlog would show the ALTER DATABASE OPEN command, but no matter how long we waited there was no corresponding "Completed" message. The affected instances would, of course, report that they weren't open if you tried to do pretty much anything.
We've got some Oracle ACS folks onsite this week... maybe I can get them to provide a bit more information.
On Wed, Feb 24, 2010 at 00:10, amonte <ax.mount_at_gmail.com> wrote:
> Hi
>
> We just hit a serious bug in 11gR2 RAC would like to let people know before
> thinking about deploy this new version.
>
> The bug seems to be 8809758 (dont know the bug description and details) and
> looks like when ASM is killed or dies when it restarts sometimes it leaves
> the database in mount state. We suspect this happens sometimes only with the
> node who is OCR master. The workaround is starting manually the instance....
>
> No fix until 12g due to clusterware design issues :-(
>
>
> Alex
>
>
-- "I'm too sexy for my code." -Awk Sed Fred. -- http://www.freelists.org/webpage/oracle-lReceived on Wed Feb 24 2010 - 07:44:37 CST