On Thu, 17 Mar 2005 12:12:06 -0800
Jeff Rizzo <riz@tastylime.net> wrote:

> Reinoud Zandijk wrote:
> 
> >Hiya folks,
> >
> >i think there is a misunderstanding with my post; my main problem is not 
> >the checksums being wrong in one of the `tcpdump's.
> >
> >My problem is that the client machines lock up on a vnode while all other 
> >network trafic goes well.
> >
> >At times bringing the interface up & down will help but that might be 
> >coincidence. In the dmesg/syslog a note is made that the NFS server is down 
> >but it *never* comes up again...
> >
> >Cheers,
> >Reinoud
> >
> >  
> >
> 
> This looks an awful lot like a problem I ran into with a pair of 2.0
> machines when I was doing pkgsrc bulk builds on one with NFS storage on
> the other...  there was a thread on it several months ago, but I don't
> think a specific PR ever got opened because I was trying to see if it
> fit into an existing one.

You can partially workaround this locks with TCP transport. Please take
a look to kern/29528 and kern/26209.

--
Mishka.