[Gluster-devel] smoke test fails for netbsd6

Atin Mukherjee amukherj at redhat.com
Mon Apr 18 15:44:00 UTC 2016



On 04/18/2016 09:03 PM, Michael Scherer wrote:
> Le lundi 18 avril 2016 à 21:00 +0530, Atin Mukherjee a écrit :
>>
>> On 04/18/2016 08:32 PM, Michael Scherer wrote:
>>> Le lundi 18 avril 2016 à 19:15 +0530, Atin Mukherjee a écrit :
>>>>
>>>> On 04/16/2016 10:00 PM, Polakis Vaggelis wrote:
>>>>> Hi,
>>>>>
>>>>> tried to backport fix for BUG 1269470
>>>>> https://github.com/gluster/glusterfs/commit/ebe40c5047e7501a5bd1747cb4d62277ae0db6e9
>>>>>
>>>>> to 3.6 branch.
>>>>>
>>>>> review link:
>>>>> http://review.gluster.org/#/c/14007/
>>>>>
>>>>> but smoke test failed
>>>>> --- glusterd_la-glusterd-op-sm.lo ---
>>>>> /home/jenkins/root/workspace/netbsd6-smoke/xlators/mgmt/glusterd/src/glusterd-op-sm.c:
>>>>> In function 'glusterd_op_start_rb_timer':
>>>>> /home/jenkins/root/workspace/netbsd6-smoke/xlators/mgmt/glusterd/src/glusterd-op-sm.c:3692:19:
>>>>> error: 'GF_REPLACE_OP_START' undeclared (first use in this function)
>>>>>
>>>>> full console log:
>>>>> https://build.gluster.org/job/netbsd6-smoke/12665/consoleText
>>>>>
>>>>> commit is not related to glusterd or should have any relation to the
>>>>> above error.
>>>>>
>>>> This is a known issue and patch [1] tries to solve it, but still
>>>> doesn't. IIRC, in release-3.6 we were just running a sub set of smoke
>>>> test and skip bsd smokes, did that get overridden by any chance?
>>>>
>>>> [1] http://review.gluster.org/#/c/12710/
>>>>
>>>> Adding misc in loop.
>>>
>>> so, does it happen in a reproducible way ?
>>> (cause if not, it might be some kind of timing issue, ie missing deps in
>>> a makefile that would be triggered with bsd make and not gnumake)
>> Looks like reproducible as it failed both the times.
> 
> Both, so for more than just 1 job ?
http://build.gluster.org/job/netbsd6-smoke/12665/
http://build.gluster.org/job/netbsd6-smoke/12671/
> 
> Do we have a list of the various jobs that failed ?
> 
>>  Does it have
>> anything to do with the recent gerrit migration?
> 
> Unlikely. Gerrit didn't changed at all, it was a copy of the disk.
> 
> And I do not see how gerrit would impact the build on jenkins, and
> specifically on netbsd.
> 


More information about the Gluster-devel mailing list