Description of problem: I have a distributed volume with 4 bricks. When i start replcace brick , the source brick goes offline. log says. ############################################################################### [2012-02-22 13:11:09.226290] I [glusterfsd.c:1578:main] 0-/usr/local/sbin/glusterfsd: Started running /usr/local/sbin/glusterfsd version 3git [2012-02-22 13:11:09.689069] W [socket.c:419:__socket_keepalive] 0-socket: failed to set keep idle on socket 8 [2012-02-22 13:11:09.689255] W [socket.c:1885:socket_server_event_handler] 0-socket.glusterfsd: Failed to set keep-alive: Operation not supported [2012-02-22 13:11:10.314013] W [xlator.c:211:xlator_dynload] 0-xlator: /usr/local/lib/glusterfs/3git/xlator/cluster/pump.so: undefined symbol: afr_fill_loc_info [2012-02-22 13:11:10.314090] E [graph.y:221:volume_type] 0-parser: Volume 'doa-pump', line 30: type 'cluster/pump' is not valid or not found on this machine [2012-02-22 13:11:10.314137] E [graph.y:330:volume_end] 0-parser: "type" not specified for volume doa-pump [2012-02-22 13:11:10.314226] E [glusterfsd.c:1466:glusterfs_process_volfp] 0-: failed to construct the graph [2012-02-22 13:11:10.314612] W [glusterfsd.c:783:cleanup_and_exit] (-->/usr/local/lib/libgfrpc.so.0(rpc_clnt_handle_reply+0x20e) [0x7ffe45740b4f] (-->/usr/local/sbin/glusterfsd(mgmt_getspec_cbk+0x354) [0x40beca] (-->/usr/local/sbin/glusterfsd(glusterfs_process_volfp+0x1cd) [0x407c46]))) 0-: received signum (0), shutting down ############################################################################# but the brick vol file has pump volume doa-pump type cluster/pump subvolumes doa-io-threads doa-replace-brick end-volume How reproducible : always with my automation script.
CHANGE: http://review.gluster.com/2796 (pump: Fixed undefined reference to fill_loc_info fn.) merged in master by Vijay Bellur (vijay)