Description of problem: The brick op of inode status commands fails on a distributed volumes as it fails to acquire lock. Version-Release number of selected component (if applicable): 3.12.2-8 How reproducible: 100 Steps to Reproduce: 1. create a distributed replicated volume. 2. mount it and create a lot of files. 3. issue gluster volume status volname inodes Actual results: Commit fails on the remote nodes. Expected results: should print the inodes status output with the connections. without the commit failing (prints it as commit failed but its a brickop failure.) Additional info:
REVIEW: https://review.gluster.org/20035 (Core: The lock contention on gf_client_dump_inodes_to_dict) posted (#1) for review on master by hari gowtham
COMMIT: https://review.gluster.org/20035 committed in master by "Amar Tumballi" <amarts> with a commit message- Core: The lock contention on gf_client_dump_inodes_to_dict Problem: For a distributed replicated volume, in the inode status command the lock on gf_client_dump_inodes_to_dict is held by the first brickop. while this is being processed, if the second brickop comes. It fails to get the lock and the whole brick op fails. Fix: Instead of using a TRY_LOCK which errors out if the lock is busy, Use LOCK which will wait till the lock is acquired. Change-Id: I768a0a1b60f28c4f7f94549e18ee3765b69cc528 fixes: bz#1579769 Signed-off-by: hari gowtham <hgowtham>
This bug is getting closed because a release has been made available that should address the reported issue. In case the problem is still not fixed with glusterfs-5.0, please open a new bug report. glusterfs-5.0 has been announced on the Gluster mailinglists [1], packages for several distributions should become available in the near future. Keep an eye on the Gluster Users mailinglist [2] and the update infrastructure for your distribution. [1] https://lists.gluster.org/pipermail/announce/2018-October/000115.html [2] https://www.gluster.org/pipermail/gluster-users/