+++ This bug was initially created as a clone of Bug #1473636 +++ Description of problem: It was found that when I/O was going on in a rep 2 volume and an add-brick was done to convert it to rep-3, mkdir failed on the newly added brick because the parent dir on that brick did not contain dht layout xattr due to which posix failed the mkdir (see BZ 1470599 if you have view permissions). So fuse guarantees that a lookup (named OR gfid lookup) is issued before the mkdir. In afr, metadata heals (which heal the xattrs including dht xattrs) happen only as a part of named lookups. Fix it to include gfid lookups as well. Note that named lookups can do name heals (if entry doesn't exist on the 3rd brick), gfid lookups can't do that. The most we can do is perform metadata heals if the entry is already present in the brick. --- Additional comment from Worker Ant on 2017-07-21 06:53:11 EDT --- REVIEW: https://review.gluster.org/17850 (afr: heal metadata in discover code path) posted (#1) for review on master by Ravishankar N (ravishankar) --- Additional comment from Worker Ant on 2017-07-23 07:16:43 EDT --- REVIEW: https://review.gluster.org/17850 (afr: heal metadata in discover code path) posted (#2) for review on master by Ravishankar N (ravishankar) --- Additional comment from Worker Ant on 2017-08-16 07:46:51 EDT --- COMMIT: https://review.gluster.org/17850 committed in master by Ravishankar N (ravishankar) ------ commit 468ca877807625817b72921d1e9585036687b640 Author: Ravishankar N <ravishankar> Date: Fri Jul 21 15:21:20 2017 +0530 afr: heal metadata in discover code path During graph switch, if fuse sends nameless (gfid) lookups, afr takes the discover code path to serve it. If there are pending metadata heals, they do not happen unless an inode refresh happens as a part of discover (which is not guaranteed to happen always). This patch fixes it by attempting metadata heal as a part of discover, just like how it is done in lookup code path. Also removed creating superfluous heal frames when launching heal. Change-Id: I49868649361ebe5d70b6ea150f4686169b6c3070 BUG: 1473636 Signed-off-by: Ravishankar N <ravishankar> Reviewed-on: https://review.gluster.org/17850 Smoke: Gluster Build System <jenkins.org> CentOS-regression: Gluster Build System <jenkins.org> Reviewed-by: Karthik U S <ksubrahm> --- Additional comment from Ravishankar N on 2017-09-02 11:31:17 EDT --- Sending one more patch based on post-merge review comments for patch in comment #3. --- Additional comment from Worker Ant on 2017-09-02 11:39:01 EDT --- REVIEW: https://review.gluster.org/18187 (afr: discover/lookup heal fixes) posted (#1) for review on master by Ravishankar N (ravishankar) --- Additional comment from Worker Ant on 2017-09-04 01:31:35 EDT --- COMMIT: https://review.gluster.org/18187 committed in master by Pranith Kumar Karampuri (pkarampu) ------ commit 37e16afa4853dcb5df6d6d0c91a6afd960c3a901 Author: Ravishankar N <ravishankar> Date: Sat Sep 2 20:24:00 2017 +0530 afr: discover/lookup heal fixes Addresses review comments in commit 468ca877807625817b72921d1e9585036687b640 Change-Id: I04b1bd3b00abfd6758798d6272954e36a24249a9 BUG: 1473636 Signed-off-by: Ravishankar N <ravishankar> Reviewed-on: https://review.gluster.org/18187 Smoke: Gluster Build System <jenkins.org> CentOS-regression: Gluster Build System <jenkins.org> Reviewed-by: Pranith Kumar Karampuri <pkarampu>
REVIEW: https://review.gluster.org/18304 (afr: heal metadata in discover code path) posted (#1) for review on release-3.10 by Ravishankar N (ravishankar)
COMMIT: https://review.gluster.org/18304 committed in release-3.10 by Shyamsundar Ranganathan (srangana) ------ commit d3b0205dbd299ff69883133ce22ca9d0623e87e2 Author: Ravishankar N <ravishankar> Date: Fri Sep 15 14:49:50 2017 +0530 afr: heal metadata in discover code path ****************************************************** Backport of: https://review.gluster.org/18202 Also added loc_is_nameless() to libglusterfs since the patch that introduced it in master was not backported to release-3.10. Note: 18202 is a squash of 17850 and 18187 in master. ****************************************************** During graph switch, if fuse sends nameless (gfid) lookups, afr takes the discover code path to serve it. If there are pending metadata heals, they do not happen unless an inode refresh happens as a part of discover (which is not guaranteed to happen always). This patch fixes it by attempting metadata heal as a part of discover, just like how it is done in lookup code path. Change-Id: I87c493045b9225741cad173bf3f645848697032e BUG: 1492010 Signed-off-by: Ravishankar N <ravishankar> Reviewed-on: https://review.gluster.org/18304 Smoke: Gluster Build System <jenkins.org> Reviewed-by: Pranith Kumar Karampuri <pkarampu> CentOS-regression: Gluster Build System <jenkins.org> Reviewed-by: Shyamsundar Ranganathan <srangana>
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-3.10.6, please open a new bug report. glusterfs-3.10.6 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] http://lists.gluster.org/pipermail/announce/2017-October/000084.html [2] https://www.gluster.org/pipermail/gluster-users/