Hey guys, for security and speed reasons I tried to bind glusterd and glusterfsd internal interfaces of my 2 storage nodes. Binding glusterd to internal interface works without problems. After starting glusterd I find out, that glusterfsd does not consider any interface parameters. I checked the source and found this snippet in xlators/mgmt/glusterd/src/glusterd-utils.c snprintf (cmd_str, 8192, "%s/sbin/glusterfsd --xlator-option %s-server.listen-port=%d " "-s localhost --volfile-id %s -p %s --brick-name %s " "--brick-port %d -l %s", GFS_PREFIX, volinfo->volname, port, volfile, pidfile, brickinfo->path, port, brickinfo->logfile); It seems like glusterfsd binds the mount always to localhost. That's the reason, why my internal interface constellation wont work. I adjusted this part using the bricks hostname instead localhost, as you can see here: snprintf (cmd_str, 8192, "%s/sbin/glusterfsd --xlator-option %s-server.listen-port=%d " "-s %s --volfile-id %s -p %s --brick-name %s " "--brick-port %d -l %s", GFS_PREFIX, volinfo->volname, port, brickinfo->hostname, volfile, pidfile, brickinfo->path, port, brickinfo->logfile); After restarting glusterd and the volume mount, everything works nice and I can mount the volume properly. I think the solution is okay, but better would be if glusterd considers the optionally --remote-host= parameter instead the bricks host. Maybe anybody can tell me how I have to adjust it. Regards, Sven