[ale] nfs broke

[Cor van Dijk]

Hi all,

I have a small network (just two machines) with an "eth0" interface and "nfs" under
redhat 6.0. This worked without a problem for a long time. Out of the blue I
get  problems during bootup. Looking in my logs, the problem seems
to start as follows:

......

Dec 14 23:05:07 tower nfs: Starting NFS services:  succeeded
Dec 14 23:05:07 tower kernel: neighbour table overflow
Dec 14 23:05:07 tower kernel: neighbour table overflow
Dec 14 23:05:07 tower rpc.statd: Cannot register service: RPC: Unable
to send; errno = No buffer space available
Dec 14 23:05:07 tower rpc.statd[401]: unable to register (SM_PROG,
SM_VERS, udp).
Dec 14 23:05:07 tower nfs: rpc.statd startup succeeded
Dec 14 23:05:07 tower kernel: neighbour table overflow
Dec 14 23:05:07 tower last message repeated 2 times
Dec 14 23:05:07 tower rpc.rquotad: Cannot register service: RPC: Unable
to send; errno = No buffer space available
Dec 14 23:05:07 tower rquota[411]: unable to register (RQUOTAPROG,
RQUOTAVERS, udp).
.............

There are many more similar messages. Eventually (after several minutes)
the various processes time-out and then the system finishes the boot process.

When I issue a nfs mount command, the command does not return (hangs?),
but seems to do its job anyway, because I am then able to access files
in the other machine from the mountpoint in the mount command.

One of the bad things about this situation is that the system does not

cleanly shutdown, buts leaves filesystems mounted. Has anyone seen something

like this? Any help is appreciated!

Cor





--
To unsubscribe: mail majordomo at ale.org with "unsubscribe ale" in message body.