[Gluster-devel] Transport endpoint not connected - Why?

[Andreas Mayer]

Hi, I'm using a simple replication solution with 2 servers which are
both Web servers with GlusterFS running both as a server and client
for each other.

Debian 6.0.4, GlusterFS 3.0.5

----- /etc/glusterfs/glusterfsd.vol -----
volume www-posix
  type storage/posix
  option directory /srv/data/www
end-volume

volume www-locks
  type features/locks
  subvolumes www-posix
end-volume

volume www-io-threads
  type performance/io-threads
  subvolumes www-locks
end-volume


volume server
  type protocol/server
  option transport-type tcp

  subvolumes www-io-threads
  option auth.addr.www-io-threads.allow *
end-volume
-----

Now, when I want to mount www via this client configuration (I took
out everything which was possible):

----- www.vol -----
volume www-server2
  type protocol/client
  option transport-type tcp
  option ping-timeout 10
  option remote-host www-server2         # local IP - defined in /etc/hosts
  option remote-subvolume www
end-volume

volume www
  type cluster/replicate
  #subvolumes www-local www-server2
  subvolumes www-server2      # access the LOCAL glusterfds, just to show
                                           # problem (doesn't work
with remote, too)
end-volume
-----

I get
ls: cannot access www: Transport endpoint is not connected

The client log shows:
[2012-04-24 18:01:28] N [glusterfsd.c:1409:main] glusterfs: Successfully started
[2012-04-24 18:01:28] N [fuse-bridge.c:2953:fuse_init] glusterfs-fuse:
FUSE inited with protocol versions: glusterfs 7.13 kernel 7.13
[2012-04-24 18:01:32] W [fuse-bridge.c:725:fuse_attr_cbk]
glusterfs-fuse: 2: LOOKUP() / => -1 (Transport endpoint is not
connected)

What does this mean?

The server log doesn't show anything interesting.

The configuration worked some time ago, and I don't know what it has
caused to stop working :(

Any help appreciated,
Best regards,
Andreas