I have two machines, imaginatively called "one" and "nine"
nine exported / to one, with no_root_squash.
one mounted nine under directory nine.
mount -t nfs nine:/ /nine
nine was powered down.
one's load went up to over 16, as users ran "df"
occasionally. df never exited. I tried to unmount /nine as
root. That process hung forever. I was forced to reboot
the machine. It took quite a while for the nfs services on
one to shut down.
This is not a bug, but typical behavior when an NFS server goes down.