I've got a client with a Netware machine that is intermittently (4
times last month) starting to behave badly. It stops taking new client
connections for anything, including AdRem sessions. Existing
connections do not appear to be affected. There are very few errors
(last time, there were no errors) on the sys$err.log file, it does not
appear to be a licensing problem.

I'm thinking it may be a memory issue - this server does a lot, and
only has 1 GB of RAM. They're also starting to see what may be the same
problem on a couple of other servers just this week, but we haven't had
any repeats so we don't know for sure.

I've run some graphs from our SNMP software, and we can see Alloc
Memory Pool climb and climb to over 800 MB followed by a huge drop
(indicating the reboot.) We see the inverse behavior with the Cache
Buffers - it goes down and down, and then jumps way high after a

The server is 5.60 rev 5. I think I've ruled this out:
They do have NSS volumes, and have Cache Balance Enable ON, 85%.

I also found something about Workstation Registration - they do run
ZENWorks, but they don't appear to have a registration policy at all.
Checking a few workstation objects in NDS shows that the last
registered time was in the morning today, so they're not re-registering
all day long and eating resources that way.

I downloaded seg.nlm and memcalc.exe, all that memcalc shows is NLM
Memory : 0 which is probably weird. The segstats.txt file is full of
info though, can post if anyone wants to see it.