Novell Open Enterprise Server 2 (OES 2) Linux Support Pack 2Novell GroupWise 7 Support Pack 3
Novell GroupWise Post Office Agent
Novell Open Enterprise Server (Linux based)
Novell GroupWise 32 bit Client in Caching Mode
GroupWise Post Office Agent (POA) Client Server (C/S) threads hang.
POA C/S threads hand randomly for up to weeks at a time before releasing.
C/S threads are hung in a Destroy Cursor state when viewed through the C/S Handler Threads detail screen on the POA's HTTP monitor.
The POA's C/S threads randomly hang in a Destroy Cursor state.
POA functionality is not adversely effected.
C/S threads do not release for up to thousands of minutes according to the POA's HTTP Monitoring Service.
The users associated with the hung thread do not experience any adverse side effects of the hung thread.
This defect has been fixed in GroupWise 7.0.3hp Post Office Agent (POA) code dated after 10/22/2008 and in GroupWise 8
This issue has only been duplicated in a GroupWise 7.0.3 post office running on OES sp1 for Linux. There are over 4,000 users all running in forced Caching Mode. The issue was discovered when it was noticed that the C/S Threads on this POA were not returning to 0 when the main load ended for the day. Upon further investigation, it was found that the several user connections, as viewed through HTTP Monitor | C/S Thread Details screen, were not releasing for up to thousands of minutes. Upon further investigation, it was discovered that the users associated with these hung threads were not experiencing any negative side effects.
At this time, it is assumed that the threads are left in this Destroy Cursor state when a TCP/IP communication is interrupted. The client is able to reestablish connection but the thread stays open. Development is looking for a root cause to this as well as a way to automatically clean up these threads.