On 4 Feb 2003 05:55:01 -0600, Daniel Eggert wrote:
> For a while I've suspected this, and now with NFS it's really 
> bad.Traffic from one end of my network gets 'chocked' every now and 
> then.
> 
> I have clients that mount a nfs share from my file server. While moving 
> data to or from the fileserver all traffic (on all clients) hangs for 
> 10-15 secs every 5-10 minutes, then it is back to normal. Besides the 
> fact that this is very annoying, I don't know how to debug it.
> 
> Here's my setup
> 
> Client ------ switch ------ file server
> 
> I though, that maybe that file servers NIC is re-genotiating the 
> ehternet speed (or something like that) every now and then, and this 
> hangs the connection.
> 
> Any ideas? Thanks in advance.

I had what looks like a similar problem with the vr0 driver dropping
packets on my server's nic :
 http://www.NetBSD.org/cgi-bin/query-pr-single.pl?number=19373
tcpdump indicated what was happening. I haven't heard any response
about the vr0 driver and don't know if it has been fixed.

The typical problems (bad cable, hub or switch) didn't hang in this
cyclical manner that this vr0 bug did, where it ran great for a
while and then just hung due to dropping large chunks of packets
for several minutes (usually just long enough for cvs to give up).