Portal instance restarts on SCS failover - why?

[Former Member]

Hi Folks,

We have implemented EP on NW04S SP14, AIX 5.3 & Oracle 10g. The installation consists of CI, DI, SCS & DB on four LPARS across two physical hosts as shown below:

Physical Host 1:

LPAR 1 - CI

LPAR 2 - SCS

Physical Host 2:

LPAR 3 - DI

LPAR 4 - DB

The DB and SCS are highly available and failover on each other.

We've noticed that when SCS failsover on DB, the CI & DI instance restart. I'm looking for information on parameters/settings needed to keep the server instance alive until SCS is available again.

Thanks.

[Former Member]

Once ERS is configured, the CI / DI processes do not restart.

Edited by: SAP Portal Project on Oct 1, 2009 10:04 AM

[Former Member]

Hi Manoj,

We use HACMP for AIX.

Check these:

/thread/132860 [original link is broken] - Enque Server was restarted without restarting the J2EE engine (MSCS)

/thread/1329634 [original link is broken] - Critical shutdown was invoked. Reason is: Enque Server was restarted ...

Note 897583 - SAP WebAS Java 6.40 SP15 - List of corrections

Note 804078 - Replicated Enqueue Configuration on Windows

The behaviour of SAP WebAS Java was changed as of 6.40 SP15 as mentioned in SAP note 897583 - "Loss of enqueue table enforces system wide restart to assure system consistency (counter messure, install replicated enqueue server)".

Regards.

[Former Member]

Last night we observed with our XI system. Are you using windows/cluster?