Return to BSD News archive
Path: sserve!newshost.anu.edu.au!munnari.oz.au!spool.mu.edu!uunet!enterpoop.mit.edu!cambridge-news.cygnus.com!athena.mit.edu!raeburn
From: raeburn@athena.mit.edu (Ken Raeburn)
Newsgroups: comp.os.386bsd.questions
Subject: Re: BUG IN NFS CLIENT ON NETBSD
Date: 16 May 1993 23:34:42 GMT
Organization: Massachusetts Institute of Technology
Lines: 28
Distribution: world
Message-ID: <RAEBURN.93May16193441@cambridge.mit.edu>
References: <1svfi2INNekn@darkstar.UCSC.EDU>
NNTP-Posting-Host: cambridge.cygnus.com
In-reply-to: buhrow@cats.ucsc.edu's message of 14 May 1993 06:51:46 GMT
In article <1svfi2INNekn@darkstar.UCSC.EDU> buhrow@cats.ucsc.edu (Brian Buhrow) writes:
Problem: When I try to nfs mount a filesystem from a remote fileserver, it
mounts fine, but any sizable reads from that filesystem subsequent to the
mount hang forever. The message, "nfs server host.foo.bar: not responding"
is generated once,
a NE-1000 or 2000 clone,
I haven't had a chance to try NFS since installing NetBSD, but this
sounds like the problem I had under 386BSD-0.1. Basically, for large
packet sizes over ethernet, packets get fragmented, and the recieving
host must reassemble them. For whatever reason (hardware or software,
I don't know), the PC reliably fails to get some of the fragments.
And when the request is retransmitted, the response is a new
fragmented packet, which the PC again fails to receive intact.
(I did get an improved ne2000 driver from someone on the net, which
improved ftp performance considerably, but unfortunately I don't have
it any more, and don't remember where I ftp'ed it from, or whether it
helped the NFS problems.)
Try using `rsize=1024,wsize=1024' in the mount options, and see if
that fixes it. Packets that small shouldn't get fragmented.
Good luck.
Ken