I've had a number of abends with AFPTCP and LIBNSS. The abend does not unload afptcp, and trying to manually unload afptcp realllly hangs the server. The only "fix" is to reset the server. Since it's abending about every 3 hours of access this isn't a real good plan.

Running NW6 sp1.1
LIBNSS ver 3.20
AFPTCP ver 2.01.10

Originally thought it might be MP issue so commented out acpidrv.psm in startup.ncf - which dropped us to one processor - didn't help.
Thought (through antidotal info) that the restore operation we we running with BackUp Exec 9.1 were the issue - nope.
We are also using this as out Groupwise server.

Sample abend log (partial)

Server MARLBOROUGH_2 halted Tuesday, February 3, 2004 10:27:18.877 am
Abend 1 on P00: Server-5.70.01: Page Fault Processor Exception (Error code 00000000)

CS = 0060 DS = 007B ES = 007B FS = 007B GS = 007B SS = 0068
EAX = 00000000 EBX = 00000000 ECX = FFFFFFFF EDX = E84F4048
ESI = 00000000 EDI = 0000001E EBP = 8C26F5EC ESP = 8C26F5E8
EIP = BE94418C FLAGS = 00010246
EIP in LIBNSS.NLM at code start +0000318Ch
Access Location: 0x0000001E

The violation occurred while processing the following instruction:
BE94418E F7D14989C88D NOT ECX, 8DC88949
BE944194 65FC GS:
BE944196 5F POP EDI
BE944197 5D POP EBP
BE944198 C3 RET
BE944199 0000 ADD [EAX], AL
BE94419B 0000 ADD [EAX], AL
BE94419D 0000 ADD [EAX], AL
BE94419F 005589 ADD [EBP-77], DL

Running process: AFP Work Thread (395) Process
Thread Owned by NLM: AFPTCP.NLM
Stack pointer: 8C271098
OS Stack limit: 8C26D200
Scheduling priority: 67371008
Wait state: 5050030 Blocked on Semaphore
Stack: --00000000 ?
--8C26F6B4 ?
81BCD0B9 (COMN.NSS|zZIDRename+B9)
--0000001E ?
--8C26F658 ?
8185D600 (NSS.NLM|cacheLookup+20)