[Gluster-users] georeplication woes

Maarten van Baarsel mrten at ii.nl
Fri Jul 20 13:06:12 UTC 2018


I've upgraded my gluster 3.13 cluster to 4.0 on an Ubuntu 
server and it broke my geo-replication because of this missing 


However, between geo-replication dying and me finding that bugreport
I've tried multiple other things including setting up geo-rep again
so I've still got a Faulty georeplication now.

Removed it, created it again with the suggested fix in place but at 
the moment I see this error on the georep slave, in 

E [MSGID: 106581] [glusterd-mountbroker.c:579:glusterd_do_mount] 0-management: Missing mspec: Check the corresponding option in glusterd vol file for mountbroker user: georep [No such file or directory]

These messages also seem relevant from /var/log/glusterfs/geo-replication-slaves/.../gsyncd.log:

[2018-07-20 12:32:11.452244] I [resource(slave gluster-1.glstr/var/lib/gluster):1064:connect] GLUSTER: Mounting gluster volume locally...
[2018-07-20 12:32:11.503766] E [resource(slave gluster-1.glstr/var/lib/gluster):973:handle_mounter] MountbrokerMounter: glusterd answered       mnt=
[2018-07-20 12:32:11.504313] E [syncdutils(slave gluster-1.glstr/var/lib/gluster):747:errlog] Popen: command returned error     cmd=/usr/sbin/gluster --remote-host=localhost system:: mount georep user-map-root=georep aux-gfid-mount acl log-level=INFO log-file=/var/log/glusterfs/geo-replication-slaves/gl0_gluster-4.glstr_glbackup/mnt-gluster-1.glstr-var-lib-gluster.log volfile-server=localhost volfile-id=glbackup client-pid=-1   error=1
[2018-07-20 12:32:11.504446] E [syncdutils(slave gluster-1.glstr/var/lib/gluster):751:logerr] Popen: /usr/sbin/gluster> 2 : failed with this errno (No such file or directory)
[2018-07-20 12:32:13.404048] W [gsyncd(slave gluster-2.glstr/var/lib/gluster):293:main] <top>: Session config file not exists, using the default config path=/var/lib/glusterd/geo-replication/gl0_gluster-4.glstr_glbackup/gsyncd.conf

Is there anyone out there that can tell me what to do with these 
error messages? I've resorted to checking out the source code 
but that road did not lead to enlightment either :(

thanks in advance,

[1] The gluster master is a 3-replica; the slave is a single 
volume at another physical location, geo-replication user is 
georep. main volume is called 'gl0', backup 'glbackup'.

root at gluster-4:/var/log/glusterfs# gluster-mountbroker status
|    NODE   | NODE STATUS |            MOUNT ROOT           |   GROUP    |       USERS       |
| localhost |          UP | /var/lib/georep-mountbroker(OK) | georep(OK) | georep(glbackup)  |

More information about the Gluster-users mailing list