rogues java process.... OEM very strange

Linux RH3, 10.1.2 OEM

We had our box crash... I thought it went all the way
down.... I am checking with the Unix team.

I know they had to bring the linux box up in
interactive mode.

The question is we found a java process still running
timestamped before the crash.. 

It was a oc4j for OEM.  We were having HUGE DB spikes
when the AWR ran each hour... 

I stopped the process tried to restart it and it
failed.

Once I killed the java process I could bring it up
just fine and things ran better.  There was no Oracle
process associated with it.

I have also had an issue with altering a system
parameter via OEM specifically PMON target.  the OS
process and oracle sql just hung and caused
performance problems.  I did a kill -9 to the Unix
process and killed the oracle process.  The Oracle
process was marked for kill but never died.. I tried
to wake up SMON several times but the process did not
died until I 
bounced the database.  

I opened up a tar with Oracle regarding the PMON/Smon
issue but they told me to bounce the db and try to
upgrade (which we are going to do 10.2 )

Anybody see anything like this?

Thanks as always.

Kathy






                
__________________________________ 
Yahoo! Mail - PC Magazine Editors' Choice 2005 
http://mail.yahoo.com
--
http://www.freelists.org/webpage/oracle-l

Other related posts:

  • » rogues java process.... OEM very strange