> I have a x86 system running the current from this weekend's tar files.  Right
> now it's building open office.  On top of that I have two sparc 20s, each
> with multiple cpus, building different versions of 3.0 and current for sparc
> using the disk on the x86.  The x86 keeps hanging up every 12 hrs or so with
> nfsd in vnlock.  Is there an easy way to fix this?

 I had the same problem and I mostly fixed it, or hid it anyway.  If it helps
anyone diagnose this, my original setup was:
  a fileserver, one disk, with /export/src, /export/pkgsrc, etc etc each
    directory exported individually.
  on the server, I did a RO loopback/null mount to see those as /usr/src,
    /usr/pkgsrc, etc etc, and one RW mount of /export/distfiles as
    /usr/pkgsrc/distfiles.
  on the NFS client I did RO NFS mounts (RW for ../distfiles/)
  Running 'lintpkgsrc -i' at the same time on each machine was a very reliable
    way to exhibit the bug.
  rebooting the NFS server would unlock the client

 Now, I rescheduled my cron jobs to not run at the same time (which wasn't
enough, but helped a lot), and changed my NFS server to export just one
directory, /export.  I've had no lockups since then (Nov 8th I think).  I
upgrade my kernels regularly, so it's possible it was a change there and not my
NFS export change that did it.

 (Any replies to the list please, I'm more likely to see it that way)