Policy Server crashing - smservmon process crashing and restarting every minute.

Products

• CA Single Sign On Secure Proxy Server (SiteMinder)
• CA Single Sign On SOA Security Manager (SiteMinder)
• CA Single Sign-On

Issue/Introduction

Description:

The policy server service keeps crashing quite frequently at irregular intervals with no real observed pattern.

In smexec.log, smservmon process crashing and restarting. The stopping of smservmon has no effect on crashing of policy server.

*************************************************************************

Thu Apr 22 16:39:26 2010: Service smservmon died - restart #0, pid: 6673        
Thu Apr 22 16:40:27 2010: Service smservmon died - restart #0, pid: 7743 
Thu Apr 22 16:41:28 2010: Service smservmon died - restart #1, pid: 8855
Thu Apr 22 16:42:28 2010: Service smservmon died - restart #2, pid: 10066
Thu Apr 22 16:43:29 2010: Service smservmon died - restart #0, pid: 11121
Thu Apr 22 16:44:30 2010: Service smservmon died - restart #0, pid: 12175
Thu Apr 22 16:45:31 2010: Service smservmon died - restart #0, pid: 13302
Thu Apr 22 16:46:32 2010: Service smservmon died - restart #0, pid: 14349
Thu Apr 22 16:47:32 2010: Service smservmon died - restart #1, pid: 15540
Thu Apr 22 16:48:33 2010: Service smservmon died - restart #0, pid: 16578
Thu Apr 22 16:49:34 2010: Service smservmon died - restart #0, pid: 17635
Thu Apr 22 16:50:35 2010: Service smservmon died - restart #0, pid: 18778
Thu Apr 22 16:51:36 2010: Service smservmon died - restart #0, pid: 19924

*************************************************************************

Solution:

This issue is fixed in release 6.0 SP6 CR01

Cause of the issue:

1. The transport layer of the monitoring API was shutting down when it detected that the connection to the monitor service is broken, terminating the instance. The problem here that the PS can be sending a message to the monitor at the same time using the instance that was being destroyed.
   
   
2. Zombie threads were created on unix machines when the connection was broken.

Environment

Release:
Component: SMPLC