From owner-freebsd-questions  Wed Apr 19 14:07:15 1995
Return-Path: questions-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id OAA29151
          for questions-outgoing; Wed, 19 Apr 1995 14:07:15 -0700
Received: from ref.tfs.com (ref.tfs.com [140.145.254.251])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id OAA29113
          for <questions@freefall.cdrom.com>; Wed, 19 Apr 1995 14:05:52 -0700
Received: (from julian@localhost) by ref.tfs.com (8.6.8/8.6.6) id OAA10207; Wed, 19 Apr 1995 14:05:23 -0700
From: Julian Elischer <julian@ref.tfs.com>
Message-Id: <199504192105.OAA10207@ref.tfs.com>
Subject: Re: NFS hanging up
To: wardd@khis.Kodak.COM (Doug Ward)
Date: Wed, 19 Apr 1995 14:05:23 -0700 (PDT)
Cc: questions@freefall.cdrom.com
In-Reply-To: <9504191650.AA20030@dal.khis.kodak.com> from "Doug Ward" at Apr 19, 95 11:50:50 am
Content-Type: text
Content-Length: 1719      
Sender: questions-owner@FreeBSD.org
Precedence: bulk

> 

 I BELIEVE THE PROBLEM IS THE ETHERNET CARDS..
(oops.. cap-lock)
This MAY NOT be your problem, but has been the case for many people..

a lot of older enet cards (including that one I think) only have
8k of buffer space that has to be used for xmit and rcv.

NFS tries to do 8k operations by default
so every time a server sends it an 8K burst  of data,
your (rather slow 386) PC drops one of the packets
and the 8k UDP packet is never transferred completely..
the NFS code retries, but that just results in the same 8K burst
of data, and of course some is missed again.... ad-infinitum...

read the man-pages for fstab,  mount_nfs and mount, and set the read and
write size to 2K.
that way, each request is likely to succeed :)
adjust to suit your system.

if you can get a card with 16K buffers (or a faster processor)
you will find that 8k transfers start working again. :)


> I am networking two FreeBSD 2.0 systems using 3Com Etherlink (not II) 
> ethernet cards.  I am using the el0 interface.
> 
> RPC, ftp and telnet work great.  I can cd and ls within a remote mounted 
> file system but when I try to cp to or from the remote file system, the
> shell I am in hangs.  The shell also hangs when I try to execute a program
> on  the remote file system.  The results are the same no matter which 
> machine I mount from.
> 
> One of the machines is a 386DX40 with 8 MB of ram, an Adaptech 1542C SCSI 
> controller and a 1.1 GB Micronet SCSI hard drive.
> The other machine is a 386DX33 with 4 MB of ram, an 100 MB IDE hard drive.
> 
> Is there a patch I need?  Is there a parameter for mount I need to check?  What
> information provided by nfsstat is useful in looking at this?
> 
> 
> thanks,  William
>