Bug 1474188 - kernel/tcmu-runner: fix uio_poll crash during uio device removal
kernel/tcmu-runner: fix uio_poll crash during uio device removal
Status: VERIFIED
Product: Red Hat Gluster Storage
Classification: Red Hat
Component: tcmu-runner (Show other bugs)
3.3
Unspecified Unspecified
unspecified Severity high
: ---
: RHGS 3.3.0
Assigned To: Prasanna Kumar Kalever
Sweta Anandpara
:
Depends On: 1473162 1465619
Blocks: 1417151
  Show dependency treegraph
 
Reported: 2017-07-24 01:49 EDT by Prasanna Kumar Kalever
Modified: 2017-09-19 06:40 EDT (History)
8 users (show)

See Also:
Fixed In Version: tcmu-runner-1.2.0-11.el7rhgs
Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: 1465619
Environment:
Last Closed:
Type: Bug
Regression: ---
Mount Type: ---
Documentation: ---
CRM:
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---


Attachments (Terms of Use)

  None (edit)
Comment 7 Sweta Anandpara 2017-08-03 07:07:58 EDT
Having discussed it with Prasanna, it will be difficult to hit this code path with the fix for uio_read() not yet in.

Hence, will be verifying this bug only after bz 1473162 is fixed.
Comment 8 krishnaram Karthick 2017-09-18 22:30:06 EDT
This issue is not seen at CNS (builds above cns-deploy-5.0.0-38.el7rhgs.x86_64). i.e., tcmu-runner-1.2.0-14.el7rhgs.x86_64

Logging and metrics using gluster block in ocp environment is up for close to 4 week now without any issues.
Comment 9 krishnaram Karthick 2017-09-19 01:21:17 EDT
(In reply to krishnaram Karthick from comment #8)
> This issue is not seen at CNS (builds above
> cns-deploy-5.0.0-38.el7rhgs.x86_64). i.e.,
> tcmu-runner-1.2.0-14.el7rhgs.x86_64
> 
> Logging and metrics using gluster block in ocp environment is up for close
> to 4 week now without any issues.

Please ignore this comment. This comment was not meant for this bug.
Comment 10 Sweta Anandpara 2017-09-19 06:40:52 EDT
Tested and verified this on the build tcmu-runner-1.2.0-15 and gluster-block-0.2.1-13 and gluster-3.8.4-44

Multiple block creates and deletes were given in parallel and in a sequence from different peer nodes, on a 2*3 volume. No crash (mentioned in description) was seen, this time or multiple other times that block deletes have been executed.

Moving this to verified in rhgs 3.3.0.

Note You need to log in before you can comment on or make changes to this bug.