GlusterFS: "0-management: connection attempt failed"
Posted: Tue Mar 24, 2015 4:42 pm
[PROBLEM]
On a 3-node glusterfs v3.4.6 setup on Debian 7 (Wheezy), I got "/var/log/glusterfs/etc-glusterfs-glusterd.vol.log" flooded with the following error/warning about "Connection refused" ever 3 seconds (!):
The share can be mounted, the data is accessible without problems, but the logs say that something's not running smoothly 
[SOLUTION]
This seems to be caused by issue #976750, where turning off NFS support causes these connection attempt failures.
It should already be fixed in v3.4.1, but that doesn't seem to be true.
Workaround: Don't disable NFS support
On a 3-node glusterfs v3.4.6 setup on Debian 7 (Wheezy), I got "/var/log/glusterfs/etc-glusterfs-glusterd.vol.log" flooded with the following error/warning about "Connection refused" ever 3 seconds (!):
Code: Select all
[2015-03-23 12:54:47.965886] I [glusterd-handler.c:2172:__glusterd_handle_friend_update] 0-: Received my uuid as Friend
[2015-03-23 12:54:51.499622] E [socket.c:2872:socket_connect] 0-management: connection attempt failed (Connection refused)
[2015-03-23 12:54:54.500224] E [socket.c:2872:socket_connect] 0-management: connection attempt failed (Connection refused)
[2015-03-23 12:54:57.500839] E [socket.c:2872:socket_connect] 0-management: connection attempt failed (Connection refused)
[2015-03-23 12:55:00.501423] E [socket.c:2872:socket_connect] 0-management: connection attempt failed (Connection refused)
[2015-03-23 12:55:03.502059] E [socket.c:2872:socket_connect] 0-management: connection attempt failed (Connection refused)
...
...
[2015-03-24 11:26:51.831856] E [socket.c:2872:socket_connect] 0-management: connection attempt failed (Connection refused)
[2015-03-24 11:26:54.832377] E [socket.c:2872:socket_connect] 0-management: connection attempt failed (Connection refused)[SOLUTION]
This seems to be caused by issue #976750, where turning off NFS support causes these connection attempt failures.
It should already be fixed in v3.4.1, but that doesn't seem to be true.
Workaround: Don't disable NFS support
Code: Select all
$ gluster volume set $GLUSTER_VOL nfs.disable off