Description of problem: In case a directory rename fails on a brick, a DHT lookup tries to create a directory with the new name as part of the heal.This means that there will be two directories with different names but the same gfid on the brick. This causes a lot of issues wrt file access as some files will not be visible on the mount point. If the code is changed so that the directory is not created if a dir with the same gfid already exists, there will be different dirs with the same gfid across bricks but file listing still works and it is easier to recover from the scenario. Version-Release number of selected component (if applicable): How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: Additional info:
REVIEW: http://review.gluster.org/15660 (storage/posix: Skip mkdir if gfid exists) posted (#1) for review on master by N Balachandran (nbalacha)
REVIEW: http://review.gluster.org/15660 (storage/posix: Skip mkdir if gfid exists) posted (#2) for review on master by N Balachandran (nbalacha)
REVIEW: http://review.gluster.org/15660 (storage/posix: Skip mkdir if gfid exists) posted (#3) for review on master by N Balachandran (nbalacha)
REVIEW: https://review.gluster.org/23245 (storage/posix: Skip mkdir if gfid exists) posted (#1) for review on master by Sheetal Pamecha
This bug is moved to https://github.com/gluster/glusterfs/issues/1059, and will be tracked there from now on. Visit GitHub issues URL for further details