Here's an odd problem which has been plaguing me since the days of 1.3 that I 
am now fed up with and am looking to get resolved...

I have a series of systems that mount various servers for easy access of log 
information, and these servers get restarted every now and then. Every now and 
then the box gets into an incoherent state and needs to be restarted due to 
unresolvable NFS issues. All in all there are over 500 mounts in total on the 
client system.

The client was originally configured using amd, but this seemed to be 
particularly unstable, with the instability increasing from 1.3 to 1.4, but 
seems a little better under 1.5 and -current. Trying to get to the bottom of 
this, the amd maps were configured to fixed mounts, and the only mounts that 
seem to be causing a real issue are those to a handful of Solaris/x86 boxes, 
and in particular after the Solaris box has been rebooted.

Trying to see what traffic was being sent between the machines resulted in 
seeing null rpc procedures being called every minute or so from the BSD to the 
Solaris box with a successful return, but no luck figuring out where things 
were wedged.

Does anyone have any suggestions on how to go troubleshooting this particular 
setup? I'm willing to go digging into code, but would like a couple of 
pointers before starting to hunt around...

Thanks.

-pjl

-------------------------------------------------------------------------------
Paul J. Lavoie		pjl@ilx.com		(212) 510-3029
ILX Systems, Inc.	111 Fulton St		New York, NY  10038