I have an NFS file server that exports mounts to 55 workstations. Occasionally the DHCP lease on the workstation IP addresses expire and get reissued new IP addresses. Unfortunately I think this causes the file server to hang (suddenly it takes 60 seconds to log in, another 60 seconds to load 'top', etc.) When I examine the CPU load or top I don't see any load on the server. Rebooting everything doesn't help.
When I encountered this problem before, my solution involved shutting off all the workstations, purging the NFS server (I'm running ubuntu server 10.04), rebooting the server, reinstalling NFS, then bringing the workstations back online.
I am not 100% sure this is the problem, but my only hint is that 'showmount' shows more IP addresses than there are workstations. I'd like know if there is some way to determine if there is a hanging NFS request, or to purge all existing requests, or hear suggestions on what else might be the problem.