Troubleshooting IMAP GroupWise issues

  • 7004663
  • 13-Oct-2009
  • 27-Apr-2012


Novell GroupWise 8
Novell GroupWise 7


Currently, most GroupWise IMAP problems fall under two categories, ABENDs and performance.

An ABEND will occur in either the GWIA, or POA.

A performance issue will have possible symptoms including any of the following:
High CPU utilization.
Agent consumes a large amount of memory.
Slow end user performance on the Post Office.
High number of IMAP threads consumed and hung.


Troubleshooting steps to resolve both problems:

1.  Patch to the latest GroupWise support pack. There are fixes in how messages are read in almost every patch. A piece of SPAM could have malformed data that can cause IMAP problems.
2.  On GWIA look at the agent logs.  The logs can show you a connection that maybe left open.  The POA logs are not verbose for IMAP information.
3.  For the POA, look at the HTTP monitor. IMAP connections will display connections which ones are  opened the longest and what they are doing.
4.  Consider getting a coredump. A coredump will show who is logged in through IMAP and what they are doing at the time.
5.  Have GWIA process the IMAP requests.  Users can be limited through the access control. 
6.  Limit the number of messages to be read in a folder.  The agent can be set to as little as 1000 messages per folder.  (Folder is read oldest to newest) .  A end user with lots of folders and 1000's of message in each folder can dramatically effect the agent.  You may find that 2000 messages or higher the problem occurs.  Next, find out who the offending user is.
6.  Limit the users that have access to IMAP through access control in GWIA and add the users back until the problem returns.

Additional Information

Performance issues with IMAP usually occur when too many agents or data is being accessed at a given time. The agents cache IMAP data in memory, causing agent to consume more memory. While the user is logged in the data stays in memory.  This is not a memory leak.  Novell has seen users cause this massive memory consumption because of either a client that never terminates their connection, or fetching a folders with lots of messages or bad data.