Bear with me here yall, this is pretty interesting(to me anyway). I
already found a fix to the problem but don't really know *why* it
happens. So here goes:

All 3 BM servers are NW51sp6, latest tcpcp, BM36sp2.

Master site - T1 Internet connection
Slave 1 site - Cable Internet (Cox)
Slave 2 site - SBC DSL

I have an AIX accounting server at the master site that folks at the
slave sites access with a term emulation program, basically telnet onsteroids. Users at both Sites 1 and 2 started complaining after the
initial installation of the VPN back in 2001(Cat may remember this)
that they couldn't access a certain report module on the AIX server.
The screen would just freeze. Reducing the MTU size on the
workstations resolved this problem and has been working great ever
since, until now. After keeping up with patches, sp's, and even
upgrading to new server hardware(identical at all sites), users at
*only* Site 1 started having the same symptom again. So I update
everything I can on that BM server but still no luck. I eventually
updated the .lan drivers on the Master BM server(just copied the new
files over and unloaded them and reinitialized system) and this fixedthe problem. Hmm...still don't understand why it didn't affect Site 2

but who cares. I got it working. This was probably a month ago.

Monday evening I reboot the Master BM server and Tuesday morning a
at Site 1 again complains of the problem. Huh? No way! So updatingthe .lan drivers didn't actually fix the problem. The ultimate
solution is unloading the .lan drivers and reinit system. I can
this over and over again: Reboot the server, user's screen freezes;
reset the nic drivers, problem goes away.

Looking at a trace before resetting the .lan drivers, the data is
getting fragmented here and there. No fragmentation afterwards. Anythoughts on this?

Lance Reynolds, CNE