Problem
When SCO NFS server is going into multiuser mode, or SCO NFS is restarted, the following error appears:
statd: cannot talk to statd at <client_machine>The client machine has been disconnected from the network (e.g., a DOS PC client) without un-mounting the resource, now the server still gives this message after a reboot.
Cause
The client should always un-mount a resource before disconnecting.
Since NFS is based on a stateless protocol, the server provides
crash and recovery functions for the locking services using the
network status monitor (/etc/statd). The statd preserves
crash/recovery state in the /etc/sm directory.
The failure of a client (e.g., client rebooted) is only detected
upon its recovery, hence when the server is rebooted it still
looks for the client machine - and the error above indicates that
the client cannot be contacted. Solution
In order to remove this message use the following procedure:
(1) Stop NFS services to safeguard other resources:
nfs stop(2) Edit the following file and remove entries identifying the disconnected client.
/etc/sm/recoverThe 'recover' file records the hostname of all systems that have as yet not been notified of statd's failure.
NOTE for UnixWare7:
==================
Check also the contents of /etc/sm.bakNOTE for ODT3.0:
================
Refer to the STATD (NADM) and STATMON (NF) manual page for more
information on these files.The STATMON (NF) manual page incorrectly references the monitor
files as "current", "backup", and "state". The correct files
are, as listed in the actual text:/etc/sm/record
/etc/sm/recover
/etc/sm/state See also
statd(1Mnfs)
statd(NADM)