Found the problem - brick on the other server had attributes set. Looks like once a brick has been used in the volume, after replacing once the brick comes back online, it cannot be re-used in the volume as it has the information left over.
From: [email protected] [mailto:[email protected]] On Behalf Of Prasad, Nirmal Sent: Monday, July 28, 2014 9:16 PM To: [email protected] Subject: Re: [Gluster-users] Replace brick of failed server - 3.5.1 Based on the logs - this is what it gets, can someone shed some light on this. [2014-07-28 17:06:43.467045] D [glusterd-op-sm.c:3008:glusterd_op_ac_send_stage_op] 0-management: Sent stage op request for 'Volume Replace brick' to 44 peers [2014-07-28 17:06:43.467062] D [glusterd-op-sm.c:3013:glusterd_op_ac_send_stage_op] 0-management: Returning with 0 [2014-07-28 17:06:43.467075] D [glusterd-utils.c:6111:glusterd_sm_tr_log_transition_add] 0-management: Transitioning from 'Lock sent' to 'Stage op sent' due to event 'GD_OP_EVENT_ALL_ACC' [2014-07-28 17:06:43.467105] D [glusterd-utils.c:6113:glusterd_sm_tr_log_transition_add] 0-management: returning 0 [2014-07-28 17:06:43.467157] E [glusterd-rpc-ops.c:773:__glusterd_stage_op_cbk] 0-management: Received stage RJT from uuid: 9edaaf4c-2f74-4242-b7cf-8320edf61f58 [2014-07-28 17:06:43.467178] D [glusterd-utils.c:5439:glusterd_friend_find_by_uuid] 0-management: Friend found... state: Peer in Cluster [2014-07-28 17:06:43.467207] D [glusterd-op-sm.c:5640:glusterd_op_sm_inject_event] 0-management: Enqueue event: 'GD_OP_EVENT_RCVD_RJT' [2014-07-28 17:06:43.467222] D [glusterd-op-sm.c:5717:glusterd_op_sm] 0-management: Dequeued event of type: 'GD_OP_EVENT_RCVD_RJT' [2014-07-28 17:06:43.467236] D [glusterd-op-sm.c:3602:glusterd_op_ac_stage_op_failed] 0-management: Returning 0 [2014-07-28 17:06:43.467248] D [glusterd-utils.c:6111:glusterd_sm_tr_log_transition_add] 0-management: Transitioning from 'Stage op sent' to 'Stage op failed' due to event 'GD_OP_EVENT_RCVD_RJT' From: [email protected]<mailto:[email protected]> [mailto:[email protected]] On Behalf Of Prasad, Nirmal Sent: Monday, July 28, 2014 7:35 PM To: [email protected]<mailto:[email protected]> Subject: Re: [Gluster-users] Replace brick of failed server - 3.5.1 Error: volume replace-brick: failed: brick: 192.168.24.57:/var/gluster/data does not exist in volume: gl_disk From: Prasad, Nirmal Sent: Monday, July 28, 2014 7:35 PM To: [email protected]<mailto:[email protected]> Subject: Replace brick of failed server - 3.5.1 Used to work in 3.5.0 - Have a completely failed server in a replica set of 3 - attempting to remove it and replace with another node fails. How do I achieve the same result? gluster volume replace-brick gl_disk 192.168.24.57:/var/gluster/data 192.168.24.12:/var/gluster/data commit force volume replace-brick: failed: brick: 10.238.24.57:/var/gluster/data does not exist in volume: gl_disk It is not practical to assume that the dead node will always be available - if it has completely crashed - it may never come back. Thanks Regards Nirmal
_______________________________________________ Gluster-users mailing list [email protected] http://supercolony.gluster.org/mailman/listinfo/gluster-users
