We have currently upgraded 2 servers to Netware 6.5 SP5 (OES SP2). We've
applied nw65os5a to server 1 because we were having fragmentation issues.
Fragmentation issues have gone away but now we have intermittent episodes
where 3-4 thousands of not-logged-in connections happen within about 2
minutes. This happens several times a day and when they get to about
24,000 connmgr grows very large which pushes the available logical space
way down until logins don't function and we need to reboot the server.
This takes about 2-3 days. We've got the following applications running
on the server: Bordermanager 3.8sp4 ir2a running http and tcp proxies (on
a traditional netware volume), Groupwise 7.0 (mta/poa/gwia), Apache
2.0.54, 2Point (faxserve) fax gateway. We've got eTrust 7.0 and 8.0
clients on our desktops but have ensured they're not checking network
volumes. The only difference between this site and another server running
6.5 is nw65os5a was not applied to server 2, apache is not running on
server 2, and the faxserve gateway is not running on server 2. We're not
experiencing this problem with server 2. When people leave for the day,
this behavior ceases. Any help would be very much appreciated. If Craig
Johnson is listening, have you heard of any bordermanager issues that may
cause this?

Thanks in advance,