Subject: Re: NFS crashes over PPP
To: Rich Van Gaasbeck <richv@hpindsl.cup.hp.com>
From: Jason Thorpe <thorpej@cs.orst.edu>
List: netbsd-users
Date: 10/05/1994 17:01:14
On Wed, 05 Oct 1994 16:16:11 -0700 
 Rich Van Gaasbeck <richv@hpindsl.cup.hp.com> wrote:

 > What nfs options are you giving to mount?  You might try setting timeo
 > really high to prevent a meltdown of retransmissions queueing up
 > behind each other and using up memory.  You could also limit rsize and
 > wsize to reduce the transmission time for each udp segment and also
 > give non-nfs traffic a chance at the link.

These are the kind of steps I took once when I compiled an entire kernel 
for an Amiga on an A3000, connected via 14.4k PPP, and the source was on an old 
Sequent running Dynix 3.1.2NFS.  I forget what I set the timeout to, but 
I had wsize and rsize at 1024...Seemed to work OK...

It took over 2 days, but the connection was reliable :-)

--------------------------------------------------------------------------
Jason R. Thorpe               thorpej@cs.orst.edu                 758-2003
Systems Administrator            CSWest Room 5                    737-2229
CS Dept, Oregon State University           http://www.cs.orst.edu/~thorpej
               "I brought my BOWLING BALL -- and some DRUGS!"
                      -- ztp