Description of problem: gluster snapshot delete volume <volname> (or) gluster snapshot delete all The above two commands might fail if some operation which acquires the lock comes in between. Explanation : snapshot delete volume <volname> does a multiple RPC when tried to do deletion of multiple snapshots in a volume. If some other operation which acquires the lock comes in between, then snapshot delete will fail with "another transaction in progress" error. Version-Release number of selected component (if applicable): How reproducible: 1/1 Steps to Reproduce: 1. issue a snapshot delete volume <volname> (say volname is vol1) 2. when step (1) is in progress, issue another command which takes a lock on volume vol1 Actual results: snapshot delete fails with another transaction in progress. Expected results: command should not fail. Additional info: Its better to give certain retries before failing.
REVIEW: http://review.gluster.org/8454 (glusterd/snapshot : Differentiating between various error scenario.) posted (#1) for review on master by Sachin Pandit (spandit)
IMO, we don't need to fix this in GD1, at least not a priority.