NFS client doesn't reconnect to server - processes all hang

[Olivier Sessink <olivier@lx.student.wau.nl>]

Hi all,

accidently a NFS server (linux 2.4.8, Debian Woody, NFS-kernel server) was
shut down for a night, and one client (linux 2.2.19, Debian Potato) had
processes using a mounted export from that server.

Those processes are in State:  D (disk sleep) (according to
/proc/7434/status) and should be running again after the server is up
again. It was mounted without initr or soft, so the processes can't be
killed.

The server is up already for two days, and I can mount new exports from the
server, but that specific mount is still not alive, and 9 processes are
still waiting for it. How do I force the NFS client to bring that mount up?
How do I debug the current state to find out why the client thinks the
server is not up yet?

dmesg shows: nfs: task 21156 can't get a request slot 
so the client thinks the server is not responding.., but I can mount other
exports from that same server...........

any help/info/links is very much appreciated

regards,
	Olivier