when replace-brick start fails-it should all terminate the new process started by replace-brick operation on source and target bricks. -------------------------------- # gluster volume replace-brick QA8 172.24.21.134:/mnt/311qa8 172.24.21.132:/mnt/311qa8 start Replacing brick from Volume QA8 failed [root@compute-241-04 ~]# ps aux | grep gluster root 2237 0.0 0.3 78064 14696 ? Ssl 23:33 0:00 glusterd root 2321 5.4 0.4 218308 16764 ? Ssl 23:36 0:28 /usr/local/sbin/glusterfsd --xlator-option QA8-server.listen-port=24009 -s localhost --volfile-id QA8.172.24.21.133.mnt-311qa8 -p /etc/glusterd/vols/QA8/run/172.24.21.133-mnt-311qa8.pid --brick-name /mnt/311qa8 --brick-port 24009 -l /usr/local/var/log/glusterfs/bricks/mnt-311qa8.log root 2326 6.8 1.2 180912 51288 ? Ssl 23:36 0:35 /usr/local/sbin/glusterfs -f /etc/glusterd/nfs/nfs-server.vol -p /etc/glusterd/nfs/run/nfs.pid -l /usr/local/var/log/glusterfs/nfs.log ~~~~~~~~~~>>>>root 2406 0.0 0.2 62584 9052 ? Ssl 23:40 0:00 /usr/local/sbin/glusterfs -f /etc/glusterd/vols/QA8/rb_client.vol /etc/glusterd/vols/QA8/rb_mount ============================== target brick # ps aux | grep gluster root 2039 0.0 0.3 78076 14580 ? Ssl 22:38 0:00 glusterd ~~~~~~~~~~>>>> root 2085 0.0 0.1 60316 6792 ? Ssl 22:40 0:00 /usr/local/sbin/glusterfs -f /etc/glusterd/vols/QA8/rb_dst_brick.vol -p /etc/glusterd/vols/QA8/rb_dst_brick.pid --xlator-option src-server.listen-port=24009 --------------------------------
Glusterd doesn't have a mechanism to rollback effects of a failed commit operation. This is true for all the CLI commands. This is not specific to replace-brick operation.
Closing this as won't fix for the reason mentioned in #comment1.