[Bugs] [Bug 1334566] New: Self heal shows different information for the same volume from each node

[bugzilla at redhat.com]

https://bugzilla.redhat.com/show_bug.cgi?id=1334566

            Bug ID: 1334566
           Summary: Self heal shows different information for the same
                    volume from each node
           Product: GlusterFS
           Version: 3.7.9
         Component: sharding
          Severity: medium
          Priority: medium
          Assignee: bugs at gluster.org
          Reporter: pcuzner at redhat.com
        QA Contact: bugs at gluster.org
                CC: bugs at gluster.org



Created attachment 1155512
  --> https://bugzilla.redhat.com/attachment.cgi?id=1155512&action=edit
vol heal output from each node

Description of problem:
I noticed that nagios was showing heal required in the BAGL test environment,
but when I checked on node gprfc085, self heal was 0.

However, I ran the following 

for node in gprfc085 gprfc086 gprfc087; do pssh -P -t 60 -H $node 'date;
gluster vol heal engine info ; sleep 1'; done

and could see that on node 85, self heal was 0 but the other two nodes show
shards listed.

Trying to understand why...I did note that for some reason
cluster.data-self-heal/entry-self-heal and meta

To date, this issue is ONLY against the 'engine' volume, which is sharded
volume and has the hosted_engine vm running on node '86


Version-Release number of selected component (if applicable):


How reproducible:
Each time.

Steps to Reproduce:
1. Run vol heal commands on each node at around the same time
2. 
3.

Actual results:
1 node shows the volume is clean, the other 2 invariably report shards in the
heal list.

Expected results:
I would expect all nodes to have the same view of heal state

Additional info:
output attached
glusterfs-3.7.9-3 build

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are on the CC list for the bug.
You are the assignee for the bug.