Description of problem: When running a delete process the system will error out reporting the directory is not empty. If you re-run the delete it will make it then actually delete. The issue seems to be related to files that have an existing self-heal issue. Version-Release number of selected component (if applicable): Gluster 3.3.2 Redhat 6.4 How reproducible: Delete a file that has an existing self heal issue. Steps to Reproduce: 1. Find a file with a self-heal issue 2. Run a delete - Error will be produced 3. Run Again - Will succeed Actual results: Doesn't delete the file Expected results: Should delete the file Additional info: It seems it might be advantageous to caused the self heal action to run and wait then resume the delete operation since the end goals is to actually remove the file. Not sure if this is related to bug# 966848
Could you please explain the self-heal issue which you are referring to in the bug description. Also Gluster 3.3.2 seem to be old release. Please test the same on the latest gluster build and check if you still see the issue.
Closing for the inactivity!