Bug 1327052

Summary: Glusterfs write meta data to normal file which cause file crashed
Product: [Community] GlusterFS Reporter: coyang <congxueyang>
Component: glusterdAssignee: bugs <bugs>
Status: CLOSED NOTABUG QA Contact:
Severity: urgent Docs Contact:
Priority: unspecified    
Version: 3.6.9CC: bugs, congxueyang, rkavunga
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2016-04-27 01:18:53 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:
Attachments:
Description Flags
detailed log when problem happen none

Description coyang 2016-04-14 07:43:13 UTC
Created attachment 1147059 [details]
detailed log when problem happen

Description of problem:
We found file is crashed in glusterfs brick, some glusterfs specific meta data is written to file. 

Version-Release number of selected component (if applicable):
we found the problem exist in glustfs 3.6.9, 3.7.2 and 3.7.6

How reproducible:
Env config description:
2 VM for glusterfs client + 3 storage nodes.

Steps to Reproduce:
1.in host use "nova stop" shutdown VMs in sequance
2.use "nova start " to startup VM
3. check system startup status
4. repeat step 1~ 3 

Actual results:
After some rounds, we found some file is crashed and file is replaced by glusterfs metadata (include some glusterfs gid)

Expected results:
file is not changed

Additional info:

in attachment.

Comment 1 Soumya Koduri 2016-04-19 12:33:34 UTC
Do you mean brick process crashed? I am unable to open the attached log file. Could you please provide core (bt).

Comment 2 coyang 2016-04-27 01:18:15 UTC
Hi,
Sorry for late reply.
Please close this bug, it's not a bug in glusterfs.
The root cause is in guest VM kernel config where the graceful shutdown of VMs by openstack does not work properly. 
With correction of kernel, the problem is fixed.
Thank you very much for your kindly check.



br,
Yang Congxue