problems in RAC with MCSG/OPS

[Oved]

Hey,

I have a configuration of MCSG/OPS that runs RAC on it.

I have a weird problem: sometimes when I want to bring down the instances of the database then one instance goes down, and when I try to bring the other down it waits for the other instance to leave, although it is already down.

The only solution in this case is to perform a "shutdown abort" to the instance.

Afterwards, when I try to start the instances, I get an oracle error message, saying that the INSTANCE_NUMBER is already in use. I double checked to see that all the resources of the instance are down, but still the error occured.

The only solution now was to perform "cmhaltcl -f" and "cmruncl" and then to start the instances.

Down anyone know what is the problem ?
Oracle say it is the MCSG/OPS problem, but I want to make sure it is.

Thanx,
Oved

[Carsten Krege]

There is a known problem with patches PHSS_29096 and PHSS_29097 that causes a failure to shutdown Oracle 8i instances gracefully. "shutdown abort" needs to be used. Messages similar to the following can be observed:

kjzmdreg1: slos err[5 skgxndrg: gm_primary_detach NMAPI2:1434 0]
kjzmerr: NM service is not available
kjzmdreg1: slos err[4 skgxndrg: invalid ctx NMAPI2:1408 0]
kjzmerr: Invalid argument for kjzmleave1
error 29701 detected in background process
OPIRIP: Uncaught error 447. Error stack:
ORA-447: fatal error in background process
ORA-29701: unable to connect to Cluster Manager
ORA-27300: OS system dependent operation:skgxndrg: invalid ctx failed with
status: 0
ORA-27301: OS failure message: Error 0
ORA-27302: failure occurred at: NMAPI2:1408
ORA-29701: unable to connect to Cluster Manager
ORA-27300: OS system dependent operation:skgxndrg: gm_primary_detach failed
with status: 0
ORA-27301: OS failure message: Error 0
ORA-27302: failure occurred at: NMAPI2:1434

These error messages will also show up on Oracle 9i and 10g, however, the database will successfully shutdown with these Oracle versions.

The problem has been introduced by PHSS_29096 for HPUX 11iv1 (11.11) and PHSS_29097 for HPUX 11iv2 (11.23). A fix will be in the upcoming patches PHSS_30417 (HPUX 11.11) or PHSS_30418 (HPUX 11.23).

HP will release a non-critical patch warning for PHSS_29096 and PHSS_29097 soon. This is a pure HP problem.

Carsten