Bug 1448469

Summary: [GSS] Facing GFID mismatch on 1x3 replica volume
Product: [Red Hat Storage] Red Hat Gluster Storage Reporter: Abhishek Kumar <abhishku>
Component: replicateAssignee: Ravishankar N <ravishankar>
Status: CLOSED INSUFFICIENT_DATA QA Contact: Nag Pavan Chilakam <nchilaka>
Severity: high Docs Contact:
Priority: high    
Version: rhgs-3.1CC: abhishku, amukherj, asrivast, bkunal, olim, ravishankar, rhs-bugs, sankarshan, storage-qa-internal
Target Milestone: ---Keywords: Triaged, ZStream
Target Release: ---   
Hardware: Unspecified   
OS: Unspecified   
Whiteboard: waiting_on_gss
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2017-12-15 10:33:03 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 1474007    

Description Abhishek Kumar 2017-05-05 13:36:41 UTC 
Description of problem:

Facing GFID mismatch on 1x3 replica volumes

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

glusterfs-3.7.9-12.el7rhgs

How reproducible:

Customer Environment

Actual results:

cluster.quorum-type is set to default,still facing GFID mismatch issue in the cluster

Expected results:

When cluster.quorum-type is set to default, volume accessibility should become read-only and no new writes should be happening,thus preventing from GFID mismatch issue 

Additional info:

Customer has already acknowledge that there is some network issue between peers that why many disconnection is happening.
Comment 3 Ravishankar N 2017-05-05 16:59:02 UTC 
Abhishek, please provide the following:
1. sosreports of the servers and clients accessing the volume.
2. For any one file in gfid-split-brain, please provide the file name and `getfattr -d -m . -e hex` output for both the file *and* the parent directory.
Comment 7 Ravishankar N 2017-05-10 11:51:19 UTC 
Hi Abhishek, please also provide the fuse client logs from the clients.

Note: Abhishek told me over IRC that the workload is running dovecot on the gluster volume and pointed me to the gfid split-brain messages in the shd logs.