Mysterious failure!

Hi, all.

We had an event the other night that's baffling, and I'm hoping someone on
here can help me figure out how to shed more light on it.

Specifically, our Oracle (10G, 10.2.0.1.0, SLES 9 with Linux 2.6.5-7.267-bigsmp)
became unresponsive. Tomcat couldn't talk to it. sqlplus and rman were both
unable to connect to it.

After poking at it for a bit, we killed the Oracle processes. This is
something I should probably be able to figure out, but after this, I was
still unable to run rman to open the database back up. I'm guessing there's
some sort of state information somewhere that became obsolete when we killed
the processes... Killing them required a KILL signal - TERM didn't do it.

Finally we rebooted the server, and with that done, I was able to launch
rman and use it to recover and restart the database.

My big question is, is there any way I can find some trace of what happened?
I'd love to know what Oracle thinks was going on while it was unresponsive.
In digging through admin files (files in adump, bdump, cdump, udump from the
data in question) I'm not seeing anything obvious, but I'm not 100% sure what
to look for. I see some files that were last updated around the time we
believe the failure occurred, but I see nothing suspicious in them.

Thanks in advance!

-- 
    Mason Loring Bliss     mason@xxxxxxxxxxx       They also surf who
awake ? sleep : dream;     http://blisses.org/     only stand on waves.
--
http://www.freelists.org/webpage/oracle-l

Other related posts: