It could be a bug !

I had it on 64 bit Oracle on 64 bit HPUX 11.00 which was cured by
patching to 8174. There is a fault where one (random) process runs off
with 100% of one of the CPUs and everything else comes grinding to a
halt (even if the other 3 CPUs are idle!). WHen you kill the process
that is running amok, everything returns to normal - for about 5 mins,
then another process takes over the mill.

Eventually, one of the Oracle background processes picks up the
processor, and that's it for today I'm afraid. Shutdown abort is the
only way to get out of that state. 8174 fixes the fault.

I don't know if this will be the cause of your problem, but it is
something to consider.


Cheers,
Norman.

--
-----------------------------------------------------------------
Norman Dunbar (at home on Linux)
Oracle (AT) mssqlBountifulSolutions (DOT) co.uk

(Delete a Microsoft database name to reply - clue, mssql !)
-----------------------------------------------------------------