Previously happy boxes started to consume massive amounts of
packet receive buffers as of this past weekend... I have two
boxes identically configured, problems started for both around
same time. Basic connectivity to Internet is not the problem...

I have patched to NW 5.1 SP 6 + TCPIP 585vrev2 + Proxy BM37ft4A
and still same issue. Unload proxy and number of packet receive
buffers, as shown in MONITOR-->Resource Utilization, in use drops
back to normal. Buffers are allocated to TCPIP.NLM. I have proxy
-cc both
boxes. Been using transparent HTTP proxy ( for past 4 years )
without incident.

Jumps in packet receive buffers seeming not proportional to the
number of requests... will jump up by thousands at a time then
back off... but eventually it will use > 20,000 and then die.
Normal usage
for these boxes is perhaps a couple hundred under severe load.

Any ideas appreciated...

-- Bob

^`∑.,,.∑'^`∑.,,.∑'^`∑.,,.∑'^`∑.,,.∑'^`∑.,,.∑'^`∑., ,.∑'^`∑.,,.∑'^

Robert Charles Mahar : "Programming is like teaching a
http://www.trafficshaper.com : jellyfish to build a house."
^`∑.,,.∑'^`∑.,,.∑'^`∑.,,.∑'^`∑.,,.∑'^`∑.,,.∑'^`∑., ,.∑'^`∑.,,.∑'^