[lug] Linux NFS server, Solaris client

[Nate Duehr]

Lots of things to check.

Also check and make sure it's not something silly like iptables blocking 
it (iptables -L) on the linux side of things, and put the machines in 
each other's host files as a test also... (and on the Solaris box, make 
sure /etc/nsswitch.conf isn't dorked up... that's pretty common if 
someone didn't check it).  If I remember correctly something in one of 
the RedHat implementations needed the hosts to reverse resolve as well 
as forward too... hmm, that was a while ago -- they say the first thing 
to go is memory... maybe I can get some error-corrected RAM for my head 
someday.

I also recently ran into some interesting weird errors with soft mounts 
between a Linux box and a NetApp and found some rather enlightening 
stuff on kernel.org about how little the current NFS maintainer cares 
about soft mounts working properly in later 2.4 kernels... sigh... 
something along the lines of "I told you guys to just use hard mounts!" 
in one of his posts.  Lovely.  Ohhhhhh well... the errors I had were 
after the mount was up and working, though... I/O errors.  Ick.

Make sure you're restarting both the portmapper and the nfsd on the 
linux side of things when you make changes if your distro's nfs scripts 
don't do that for you... not always necessary, but have seen it help in 
certain cases where things are just generally confused on the linux box.

I'm just adding to your pile of things to check, pretty much agreed with 
all the other posts... lots of stuff could be wrong.  You need a hint 
from the log files at this point.

Nate, nate at natetech.com