Re: RAC Failing after O/S patch

Hi Bill

Assuming there are no leftover shared memory segments (ipcs) and the
linux kernel parameters are as per 10gR1 recommendations, you can try
reducing the db_cache_size and see if helps ?

-Rajeev

On 2/28/07, William Wagman <wjwagman@xxxxxxxxxxx> wrote:


Greetings,

I am running a 2 node RAC cluster on Oracle Standard Edition 10.1.0.4.0,
database is on ASM and CRS is 10.1.0.4.0 also. The O/S is RHEL3. Last
week we upgraded from 2.4.21-32.0.1.Elsmp to 2.4.21-47.0.1.Elsmp and now
in one of the two instances am seeing the following errors in the
database instance (not the ASM instance) alert log and the trace files
being generated...

Wed Feb 28 07:46:46 2007
Trace dumping is performing id=[cdmp_20070228074646]
Wed Feb 28 07:47:29 2007
Timed out trying to start process PZ99.
Wed Feb 28 07:47:29 2007
Errors in file /opt/pkg/oracle/admin/mothra/bdump/mothr1_ora_24179.trc:
ORA-00603: ORACLE server session terminated by fatal error
ORA-27504: IPC error creating OSD context
ORA-27300: OS system dependent operation:sendmsg failed with status: 12
ORA-27301: OS failure message: Cannot allocate memory
ORA-27302: failure occurred at: sskgxpsnd1
Wed Feb 28 07:47:30 2007
Trace dumping is performing id=[cdmp_20070228074730]

The errors are only appearing on node 1. I have searched Metalink and
the web and found references to these errors but nothing which matches
this situation. I am still waiting to hear from the sysadmin as to
whether or not he has seen any reports of the O/S encountering memory
problems.

I would appreciate any thoughts as to what to look at or suggestions as
to how to track down the error. So far Oracle has been unresponsive.

Thanks.

Bill Wagman
Univ. of California at Davis
IET Campus Data Center
wjwagman@xxxxxxxxxxx
(530) 754-6208
--
http://www.freelists.org/webpage/oracle-l

Other related posts: