[SunHELP] I'm at a loss...

[Shawn Tagseth]

I've been fighting with a broken NFS setup for a couple days and nothing
seems to help.

Situation:  3 out of 4 solaris client machines (2 solaris 7, 1 solaris 8)
with the /home directory nfs mounted from another solaris 8 machine can't
obtain locks on files or receive various errors like "can't close file:
resource temporarily unavailable".  The AIX machines and linux machines we
have mounting to this home directory don't seem to have a problem :(

I thought it might be some stale locks so I nfs.client stop|start on the
three machines.  I've increased the kernel threads for the nfs server.  Out
of desperation we even (gasp, the horror!) rebooted all the machines,  to no
avail (I know they aren't windows machines but we had to try something, I'm
out of answers).

There are no network problems, netstat -i and the cisco switch show no
errors.  Iostat -e does have a couple errors on that drive but no more than
other machines that have been up for the same amount of time.  The rpc
services are all running and there is nothing in the /var/log/messages.  It
seems to work for about an hour if we nfs.server stop|start.  I am about to
start nfslogd but having never used it before I don't know if it will give
me more information.  Iowait and kernel memory usage as reported by top seem
to be higher than usual, but I can't be 100% sure of that. 

Anyone have any idea where else I can look?

Thanks,
Shawn