Re: Weird database hanging

Your data from ASH doesn't have anything between 10:44 and 10:53 - it was probably stuck as well.

"os thread startup" is rather strange - problems starting new processes?

And this is interesting - "cursor: pin S wait on X". Alex Fatkulin is hinting next to me that it *might* be new "cool" mutex that is enabled by default in 10.2.0.2. Disable - _kks_use_mutex_pin=false.

There was a discussion on Oracle-L about it.

On 10/24/07, Don Seiler <don_at_seiler.us> wrote:
> Well the mysterious 10-15 minutes instance hanging happened again
> today. It also happened two weeks ago. Much better than 1-3 times
> per day, but frustrating none-the-less. From our vantage point,
> Veritas *looks* to be in order this time. I really need to increase
> my familiarity with SAN/disk stuffs in general.
>
> dba_hist_active_sess_history shows a boatload "latch: library cache"
> and "latch: shared pool" wait events during the instance hanging. As
> before, nothing is printed to the alert log during the hanging, and
> immediately after a flurry of connection timeout warnings get printed.
>
> This morning's fun data is up at http://pastebin.ca/748150 . Hanging
> started ~ 10:43 AM.
>
> I did run a hanganalyze and systemstate dump, although the latter
> ended with a segfault. If anyone is interested in those, feel free to
> email me.
>
> I'm reluctant to open another SR and be pushed back with "change your
> application SQL" reponses. Any help would be much appreciated and
> thoroughly documented for everyone to enjoy.
>
> To recap:
> Oracle Database 10g Enterprise Edition Release 10.2.0.2.0 - 64bit Production
> Red Hat Enterprise Linux AS release 4 (Nahant Update 5)
>
> --
> Don Seiler
> oracle: http://ora.seiler.us
> ultimate: http://www.mufc.us
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>

-- 
Alex Gorbachev, Oracle DBA Brewer, The Pythian Group
http://www.pythian.com/blogs/author/alex http://www.oracloid.com
BAAG party - www.BattleAgainstAnyGuess.com
--
http://www.freelists.org/webpage/oracle-l