I have a NW65SP5 cluster that shows an odd trend with SLP.NLM memory
usage. Once they're up for a while, they get over 400MB of RAM consumed.
That looks like a bug to me, and I see that SP6 has the same date of
SLP.NLM that I'm running right now.

Alloc Memory (Bytes) 430011988 SLP Service Agent Memory

A freshly (within 18 hours) rebooted machine looks like this:

Alloc Memory (Bytes) 2203048 SLP Service Agent Memory

Our SLP scope has 293 items in it at the time of this posting, so I doubt
it's trying to cache h-u-g-e amounts of objects.

Any idea what could be driving this up? Of my longer uptime servers (more
than 2 days) all of them have SLP Service Agent Memory between 400M and
450M bytes. That's remarkably uniform.

Novell, it does a network good