Red Hat Bugzilla – Bug 1404110
[Eventing]: POSIX_SAME_GFID event seen for .trashcan folder and .trashcan/internal_op
Last modified: 2017-03-23 01:56:27 EDT
Description of problem: ======================= In my 4-node cluster with eventing enabled and webhook configured as a listener, chanced upon EVENT_POSIX_SAME_GFID for .trashcan and .trashcan/internal_op folders, and not for any of the other paths, during my testing. Looks like this event is masked for all paths, but not for .trashcan. Raising a BZ after confirming it with Pranith. Version-Release number of selected component (if applicable): ============================================================= 3.8.4-7 How reproducible: ================= Many times Additional info: ================== [root@dhcp47-26 ~]# rpm -qa | grep gluster glusterfs-events-3.8.4-7.el7rhgs.x86_64 glusterfs-geo-replication-3.8.4-7.el7rhgs.x86_64 glusterfs-fuse-3.8.4-7.el7rhgs.x86_64 vdsm-gluster-4.17.33-1.el7rhgs.noarch glusterfs-3.8.4-7.el7rhgs.x86_64 glusterfs-cli-3.8.4-7.el7rhgs.x86_64 gluster-nagios-addons-0.2.8-1.el7rhgs.x86_64 gluster-nagios-common-0.2.4-1.el7rhgs.noarch glusterfs-client-xlators-3.8.4-7.el7rhgs.x86_64 glusterfs-server-3.8.4-7.el7rhgs.x86_64 glusterfs-api-3.8.4-7.el7rhgs.x86_64 glusterfs-libs-3.8.4-7.el7rhgs.x86_64 python-gluster-3.8.4-7.el7rhgs.noarch [root@dhcp47-26 ~]#
Missed pasting this in the description. "{u'message': {u'gfid': u'00000000-0000-0000-0000-000000000005', u'newpath': u'/.trashcan/', u'brick': u'dhcp41-210.lab.eng.blr.redhat.com:/bricks/brick0/a3', u'path': u'/bricks/brick0/a3/.glusterfs/00/00/00000000-0000-0000-0000-000000000001/.trashcan'}, u'event': u'POSIX_SAME_GFID', u'ts': 1477475560, u'nodeid': u'461896db-d8a7-4881-82fc-42d6042e80c4'}
This issue comes only when we do a brick stop and brick start which is not an event we do frequently. Shwetha is also trying to recreate to see if there are any other code paths where this appears. After discussing with Shwetha, we are of the opinion that we can defer this at the moment and if Shwetha finds that there are other code paths which can lead to this event showing up frequently then we can bring it back for 3.2.0.
mainline patch posted - http://review.gluster.org/#/c/16212/
Tested and verified this on the build 3.8.4-13. Steps to reproduce: ================== 1. Have a distribute replicate volume on a gluster cluster, with eventing enabled. 2. Kill one of the brick processes 3. Start the volume by force, thereby starting the brick process again. When the events were monitored on the webhook, no POSIX_SAME_GFID events were seen on .trashcan and entries under it. Moving this BZ to verified in 3.2.
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHSA-2017-0486.html