[Gluster-devel] Brick-Mux tests failing for over 11+ weeks

Shyam Ranganathan srangana at redhat.com
Fri May 25 14:25:41 UTC 2018


Updates from tests:

Last 5 runs on 4.1 have passed.

Run 55 and 58 failed on bug-1363721.t which was not merged before the
tests were kicked off, hence still considering it as passing.

Started 2 more runs on 4.1 [1] and possibly more during the day, to call
an all clear on this blocker for the release.

`master` branch still has issues, jFYI.

Shyam

[1] New patch to test mux only: https://review.gluster.org/#/c/20087/1
On 05/24/2018 03:08 PM, Shyam Ranganathan wrote:
> After various analysis and fixes here is the current state,
> 
> - Reverted 3 patches aimed at proper cleanup sequence when a mux'd brick
> is detached [2]
> - Fixed a core within the same patch, for a lookup before brick is ready
> case
> - Fixed an replicate test case, that was partly failing due to cleanup
> sequence and partly due to replicate issues [3]
> 
> Current status is that we are still tracking some failures using this
> [1] bug and running on demand brick-mux regressions on Jenkins.
> 
> State of 4.0 release is almost even with 4.1, with the fixes/changes
> above, but it is not yet in the clear to continue with 4.1 release.
> 
> Request any help that can be provided to Mohit, Du and Milind who are
> currently looking at this actively.
> 
> Shyam
> 
> [1] Mux failure bug: https://bugzilla.redhat.com/show_bug.cgi?id=1582286
> 
> [2] Mux patches reverted: https://review.gluster.org/#/c/20060/4 (has 2
> other patches that are dependent on this)
> 
> [3] Replicate fix: https://review.gluster.org/#/c/20066/
> _______________________________________________
> Gluster-devel mailing list
> Gluster-devel at gluster.org
> http://lists.gluster.org/mailman/listinfo/gluster-devel
> 


More information about the Gluster-devel mailing list