Bug#637874: nfs mounts unreliable

[J <TeMail@gmx.com>]

Hello Ben,

I have been watching this bug because I have similar behaviour - but I 
do not know that it is the same.

I have two exports from the nfs server, which are mounted by one client.

After boot, it is common for one of the exports to mount correctly, 
while the other one is nowhere to be seen.

/etc/network/interfaces is set up like this, straight out of the squeeze 
setup (6.0.2 release):

allow-hotplug eth0
iface eth0 inet static
        address 10.10.0.218
        netmask 255.255.255.0
        network 10.10.0.0
        broadcast 10.10.0.255
        gateway 10.10.0.231
        # dns-* options are implemented by the resolvconf package, if 
installed
        dns-nameservers 10.10.0.231
        dns-search my.domain

The nfs-server is listed in /etc/hosts, so I doubt that it is a DNS 
problem.  I am thinking of moving from "allow-hotplug" to "auto" and 
seeing if that fixes the problem.  For now, I have put the following 
into /etc/init.d/mount_local:

/bin/ping -q -i 0.2 -c 1 -w 10 nfs-server
/bin/mount -v /mount/a
/bin/mount -v /mount/b
/bin/mount

With the ping in place, all exports are correctly mounted.
Without the ping, only one export will be mounted.

By the way, I was hoping that you (or somebody) could explain what the 
following snippet is supposed to do:


case "$FSTYPE" in
          nfs|nfs4|smbfs|cifs|coda|ncp|ncpfs|ocfs2|gfs)
            ;;


It looks like a no-op to me.

bfn,

J