we have a setup with a distribute volume with two bricks, the issue happens is that glusterd fails to restart, the data set contains, a dir, under which we have the nested dirs and leaf is having set the quota limit set for 1MB , altogether it's around 400-500 dirs. the log files says this: [2011-07-11 20:41:10.132789] E [glusterd-store.c:990:glusterd_store_handle_retrieve] 0-glusterd: Unable to retrieve store handle for /etc/glusterd/vols/dist1/bricks/(null), error: No such file or directory [2011-07-11 20:41:10.132800] E [glusterd-store.c:1634:glusterd_store_retrieve_volumes] 0-: Unable to restore volume: dist1 [2011-07-11 20:41:10.132816] E [xlator.c:1390:xlator_init] 0-management: Initialization of volume 'management' failed, review your volfile again [2011-07-11 20:41:10.132826] E [graph.c:331:glusterfs_graph_init] 0-management: initializing translator failed [2011-07-11 20:41:10.132833] E [graph.c:503:glusterfs_graph_activate] 0-graph: init failed [2011-07-11 20:41:10.132966] W [glusterfsd.c:712:cleanup_and_exit] (-->/root/quota/inst/sbin/glusterd(main+0x3c2) [0x405262] (-->/root/quota/inst/sbin/glusterd(glusterfs_volumes_init+0x18b) [0x40426b] (-->/root/quota/inst/sbin/glusterd(glusterfs_process_volfp+0x17a) [0x4040ca]))) 0-: received signum (0), shutting down
PATCH: http://patches.gluster.com/patch/7854 in release-3.2 (mgmt/glusterd: write complete string contained in value in to the volinfo file.)
Re-opening as this requires more changes.
CHANGE: http://review.gluster.com/71 (Change-Id: I64d7e143eb4b0fda76a9b97134d0233763a1679a) merged in release-3.2 by Anand Avati (avati)
CHANGE: http://review.gluster.com/8 (Change-Id: I06173a4cf22e12bc543f8ff2d151078333b500e1) merged in master by Anand Avati (avati)
Created around 600 directories by name dir0/dir1/dir2/dir3...... And on the leaf directory set the quota limit. Killed glusterd and restarted it. glusterd started and was able execute further gluster commands such as quota remove etc.