Strange NFS messages: "lockd: cannot monitor..."

[Terje Fåberg <terje_fb@yahoo.no>]

Hello everyone,

Short problem: When mounting an nfs filesystem
I get dozens of error messages, although the mount
seems to work flawlessly:

   lockd: cannot monitor 192.168.20.1

Long description: I'm mounting a nfs file system
out of a minimal system in an initrd.  In the initrd
i configures the network card, runs dhclient and then
mounts the nfs file system. "rpcinfo" shows me, that
portmapper, lockd and statd are running on the
client:
 
   $ rpcinfo -p <client-ip>
   Program Vers Proto   Port
    100000    2   tcp    111  portmapper
    100000    2   udp    111  portmapper
    100024    1   udp    607  status
    100024    1   tcp    610  status
    100021    1   udp  32768  nlockmgr
    100021    3   udp  32768  nlockmgr
    100021    4   udp  32768  nlockmgr

What did I miss? I manually start portmap and
rpc.statd before trying the mount. The kernel
starts lockd automatically. So what else do I
need to do?  I'm thankful for any hints.

Both client and server run Debian Sarge. Client
runs vanilla 2.6.9, server runs vanilla 2.4.27.

The background: I have a few diskless X-terminals
that currently mount their root file systems via nfs.
  
Of course, they cannot cope with a ro root file
system, so I am forced to export all filesystems in
writable mode on the server. This makes me feel a
_little_ uncomfortable in a network in which I do
not control every host. 

The terminals run a stripped-down version of debian
sarge with a total size of ~250MB of which ~210 MB
are in /usr.

So I planned to implement this boot sheme:  

 (1) initrd creates a new tmpfs
 (2) initrd copies the image of the root file-system
     minus /usr into that tmpfs
 (3) mount /usr from server in read-only mode
 (4) initrd pivots to the tmpfs

This way no data on the server needs to be writable
at all. 


Regards, 
Terje