Two of my servers are experiencing abnormal memory usage by xmgr.nlm. All of my servers have been recently upgraded from NW65sp7 to NW65sp8 and from Edirectory v8.8.2 to Edirectory v8.8.5-ftf2. All my other NW65sp8 servers (35 total) are not experiencing this problem.

I read various Tids on this issue for example:

XMGR.NLM appears to have a memory leak-->
XMGR.NLM appears to have a memory leak

OpenSSH patch for NW65SP8-->
OpenSSH patch for NW65SP8

But in my opinion I don't believe these apply to my issue because we are not using SSH or SFTP on either server. Actually, I checked all the NLM's that are updated in the OPenSSH patch and none of them are in my sys:system folder. Also the SS patch v2.06 says that I should not install SSv2.06 on Edirectory 8.8.3, so I didn't install it (assuming it means v8.8.3 or greater-you know what they say about assuming!!).

The two servers having the issue are running Groupwise v7.03 HP3. both hold Post offices and domains.

What started me looking at the issue was that I found my cache buffers dropping on both these servers and cache memory allocator error messages in the system log. Both the servers have 4GB of RAM and I have set file cache maximum size to 2147483648, but the server's auto-tune feature keeps adjusting memory and logging messages like this:

Server logical address space is running low. Increase the available logical space by restarting the server with the -u402075648 switch

3-01-2010 11:27:55 am: SERVER-5.70-3728
Severity = 0 Locus = 17 Class = 21
Server logical address space is running low. File Cache Maximum Size has been set to 1073741824.

I have not set the -U switch setting as I have had problems following these recommendations in the past. However I do reset the file cache maximum size to 2147483648 once the cache buffers drop.

I believe my issue must be related to the Links above but I'm not sure of the fix.

Has anybody out there heard of anything or know what I can do to resolve this issue?