Hello All,

We are running BorderManager 3.6 SP2A on NetWare 5.1 SP5. On Monday 3/10 we suddenly started seeing the server's Packet Receive Buffers
being rapidly consumed ... this behaviour continued throughout the
through a couple of reboots of the server, then stopped late in the

For a long time we've had Minimum Packet Receive Buffers = 2000 and
Maximum Packet Receive Buffers = 4000. The server would run for
at a time without adding buffers above the initial minimum of 2000.
Once in a rare while something unknown would cause it to spike up to
maximum of 4000 (generating the "low on event control blocks"
But after a reboot it would go back to 2000 and stay there.

Today after the server started consuming all the receive buffers and becoming unresponsive, I ran Craig Johnson's latest TUNEUP.NCF to
increase the minimum to 5000 and the maximum to 10000, and rebooted. After a reboot, I watched in MONITOR as the number of allocated
rose from 5000 to 10000 in less than 30 seconds. I then turned to the

Proxy Console, "Display Current Activity" screen, where I could see
number of allocated receive buffers actually in use. It would rise up

to near 10000, then drop down to below 10 for a while, then climb
rapidly back up to 10000. Late in the day it seemed to stabilize at a

low number.

Does anyone have any ideas as to what might be going on here, or
suggestions on how to proceed to identify the cause of this problem?


Doug Percival, CNE
LAN Administrator
World Resources Institute
Washington DC