NFS mount(2): timeout on Ubuntu 12.04.1

By on 21 septembre 2012, in Boulot, Informatique, Journal, Libre Software, Licorn®, LXC, System administration, Ubuntu, Work

Searching on google, I did get many results for this kind of problem, involving many solutions to try. The thing to note is that my configuration worked a few days ago. I Tried without success:

– disabling explicitely NFSv4 (I already used v3 only, to support POSIX1e ACLs),
– trying other various NFS options,
– disabling my local bridged ethernet configuration (needed for all my LXCs) on both NFS server and client,
– disabling IPv6 on both, too.

Nothing worked, no more message was printed, no debug, nothing. Yikes!

Exact symptoms:

– all mounts from client fail with `mount(2): timeout`;
– `telnet server mountd_port` works !!
– all RPC services are up and running (`rpcinfo -p` OK, etc);
– everything else work like a charm, as far as I tested.

The problem was coming from my NETGEAR GS724T L2/L3 switch, on which I recently turned on « Denial of Service » features, just to test them.

The culprit was **Denial of Service TCP Flag**: with it enabled, the mount call fails. Without it, everything is back to normal and my NFS mounts are instantly mounted.

NOTE: I didn’t test other DoS features: I disabled them all like they were a few days ago and NFS mount succeeded, enabled **Denial of Service TCP Flag** only and noticed mount failed again, then disabled it and didn’t touch to these F….G features again.

Problem solved.

1
  • sorcier777

    Thanks for sharing, I just got the same issue and it’s solved now.

Trackbacks and Pingbacks

Reply