Machine #1:

NetBSD 1.0_BETA (SENTINEL) #0: Mon Oct 24 20:29:26 CDT 1994
    root@phoenix.dodds.vorpal.com:/usr/src/sys/arch/i386/compile/SENTINEL
CPU: i386DX (386-class CPU)
real mem  = 3801088
avail mem = 2375680

Machine #2:

NetBSD 1.0_BETA (WYVERN) #0: Mon Oct 24 20:37:11 CDT 1994
    root@phoenix.dodds.vorpal.com:/usr/src/sys/arch/i386/compile/WYVERN
CPU: i486DX (486-class CPU)
real mem  = 7995392
avail mem = 6328320

source tree is -current from Oct 14

#2 runs a slip connection (6 foot serial connection) to #1, speed has
been tried both 19200 and 38400

#2 nfs mounts /usr from #1.  #2 gets to the rc script and starts to run
the daemons.  both machines start to get silo overflows, (neither end
has a 16550 uart) and then the nfs mount seems to die.. the console of
#1 gives "/netbsd: nfsd send error 55"  The death is usually when trying
to start the cron daemon (that may not be important, but heck.. )

I tried running tcpdump on the connection, the only thing interesting
being some `truncated-ip` messages.
It seems that the nfsd has gotten tired of sending repeats and decides
that the connection isn't worth the trouble; so it gives up.

I can trace the origin of the error message in the source to
src/sys/nfs/nfs_socket.c
but I don't know enough to do anything more..

Another interesting point is that if I then reboot #2, I get 
NFS Portmap: RPC: Port mapper failure - RPC: Unable to send
when it tries to nfs mount /usr again

	If anyone can give me any help, it would be appreciated..

		Thanks -
			- Will Stoltenberg, griffon@phoenix.dodds.vorpal.com