You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@beam.apache.org by Valentyn Tymofieiev <va...@google.com> on 2020/02/06 15:33:30 UTC
Re: Jenkins trigger logic
>
> although sometimes no status checks appear, the job is actually submitted
> to Jenkins.
This indeed may be happening sometimes, but definitely not every time. The
earliest precommit command on the PR 10346 for commit 2bf1a2b was
triggered[1] on 10:26 AM PST or 6:26 PM UTC (Jenkins uses UTC). The
earliest test run triggered by jenkins on commit 2bf1a2b that I see in the
builds was at 3:21 PM PDT or 11:21 PM UTC [2], possibly triggered by the
command on the PR at 3 PM PDT [3]. At least 4 commands to trigger the suite
were ignored...
Hopefully whatever recent change that caused this will be reverted once we
> move to the new Beam Jenkins master. (https://ci-beam.apache.org).
Do we have a JIRA that tracks this migration?
Thanks
[1] https://github.com/apache/beam/pull/10346#issuecomment-578885566
[2] https://builds.apache.org/job/beam_PreCommit_Python_Phrase/1414/
[3] https://github.com/apache/beam/pull/10346#issuecomment-578996626
On Tue, Jan 28, 2020 at 9:37 AM Udi Meiri <eh...@google.com> wrote:
> Hopefully whatever recent change that caused this will be reverted once we
> move to the new Beam Jenkins master. (https://ci-beam.apache.org).
>
> On Tue, Jan 28, 2020, 00:39 Ismaël Mejía <ie...@gmail.com> wrote:
>
>> I also noticed that after the changes on Jenkins by Infra in multiple PRs
>> we need to trigger the message twice (or 3 times) and then the run starts.
>>
>> On Tue, Jan 28, 2020 at 9:36 AM Kamil Wasilewski <
>> kamil.wasilewski@polidea.com> wrote:
>>
>>> I observe that although sometimes no status checks appear, the job is
>>> actually submitted to Jenkins. It seems this happened in your case as well:
>>> https://builds.apache.org/job/beam_PreCommit_Python_Phrase/1417
>>>
>>> Also, does anyone know a reliable way to trigger all precommit jobs for
>>>> a PR, instead of calling out each precommit job separately?
>>>>
>>>
>>> I believe `retest this please` does the trick.
>>>
>>> On Tue, Jan 28, 2020 at 3:16 AM Valentyn Tymofieiev <va...@google.com>
>>> wrote:
>>>
>>>> It appears that Jenkins sometimes does not respond to the trigger
>>>> commands for extended periods of time (see image below from
>>>> https://github.com/apache/beam/pull/10346). Does anyone know why this
>>>> happens?
>>>>
>>>> Also, does anyone know a reliable way to trigger all precommit jobs for
>>>> a PR, instead of calling out each precommit job separately?
>>>>
>>>>
>>>> [image: image.png]
>>>>
>>>
Re: Jenkins trigger logic
Posted by Shoaib Zafar <sh...@venturedive.com>.
Hi Everyone.
Similar issue with the PR https://github.com/apache/beam/pull/10712 as well.
I've seen the tests are triggered and passed on Jenkins but for some reason
Github not showing any activity!
https://builds.apache.org/job/beam_PreCommit_Python_Commit/11080/
Regards,
*Shoaib Zafar*
Software Engineering Lead
Mobile: +92 333 274 6242
Skype: live:shoaibzafar_1
<http://venturedive.com/>
On Sat, Feb 8, 2020 at 5:55 AM Mark Liu <ma...@google.com> wrote:
> I noticed that a similar instance happened in
> https://github.com/apache/beam/pull/10762.
>
> To me it's more like Jenkins configuration issue probably related to ghprd
> plugin. I filed following jira to INFRA team and ask for the insights:
> https://issues.apache.org/jira/browse/INFRA-19836
>
> On Thu, Feb 6, 2020 at 7:34 AM Valentyn Tymofieiev <va...@google.com>
> wrote:
>
>> although sometimes no status checks appear, the job is actually submitted
>>> to Jenkins.
>>
>>
>> This indeed may be happening sometimes, but definitely not every time.
>> The earliest precommit command on the PR 10346 for commit 2bf1a2b was
>> triggered[1] on 10:26 AM PST or 6:26 PM UTC (Jenkins uses UTC). The
>> earliest test run triggered by jenkins on commit 2bf1a2b that I see in the
>> builds was at 3:21 PM PDT or 11:21 PM UTC [2], possibly triggered by the
>> command on the PR at 3 PM PDT [3]. At least 4 commands to trigger the suite
>> were ignored...
>>
>> Hopefully whatever recent change that caused this will be reverted once
>>> we move to the new Beam Jenkins master. (https://ci-beam.apache.org).
>>
>>
>> Do we have a JIRA that tracks this migration?
>>
>> Thanks
>>
>> [1] https://github.com/apache/beam/pull/10346#issuecomment-578885566
>> [2] https://builds.apache.org/job/beam_PreCommit_Python_Phrase/1414/
>> [3] https://github.com/apache/beam/pull/10346#issuecomment-578996626
>>
>>
>> On Tue, Jan 28, 2020 at 9:37 AM Udi Meiri <eh...@google.com> wrote:
>>
>>> Hopefully whatever recent change that caused this will be reverted once
>>> we move to the new Beam Jenkins master. (https://ci-beam.apache.org).
>>>
>>> On Tue, Jan 28, 2020, 00:39 Ismaël Mejía <ie...@gmail.com> wrote:
>>>
>>>> I also noticed that after the changes on Jenkins by Infra in multiple
>>>> PRs we need to trigger the message twice (or 3 times) and then the run
>>>> starts.
>>>>
>>>> On Tue, Jan 28, 2020 at 9:36 AM Kamil Wasilewski <
>>>> kamil.wasilewski@polidea.com> wrote:
>>>>
>>>>> I observe that although sometimes no status checks appear, the job is
>>>>> actually submitted to Jenkins. It seems this happened in your case as well:
>>>>> https://builds.apache.org/job/beam_PreCommit_Python_Phrase/1417
>>>>>
>>>>> Also, does anyone know a reliable way to trigger all precommit jobs
>>>>>> for a PR, instead of calling out each precommit job separately?
>>>>>>
>>>>>
>>>>> I believe `retest this please` does the trick.
>>>>>
>>>>> On Tue, Jan 28, 2020 at 3:16 AM Valentyn Tymofieiev <
>>>>> valentyn@google.com> wrote:
>>>>>
>>>>>> It appears that Jenkins sometimes does not respond to the trigger
>>>>>> commands for extended periods of time (see image below from
>>>>>> https://github.com/apache/beam/pull/10346). Does anyone know why
>>>>>> this happens?
>>>>>>
>>>>>> Also, does anyone know a reliable way to trigger all precommit jobs
>>>>>> for a PR, instead of calling out each precommit job separately?
>>>>>>
>>>>>>
>>>>>> [image: image.png]
>>>>>>
>>>>>
Re: Jenkins trigger logic
Posted by Mark Liu <ma...@google.com>.
I noticed that a similar instance happened in
https://github.com/apache/beam/pull/10762.
To me it's more like Jenkins configuration issue probably related to ghprd
plugin. I filed following jira to INFRA team and ask for the insights:
https://issues.apache.org/jira/browse/INFRA-19836
On Thu, Feb 6, 2020 at 7:34 AM Valentyn Tymofieiev <va...@google.com>
wrote:
> although sometimes no status checks appear, the job is actually submitted
>> to Jenkins.
>
>
> This indeed may be happening sometimes, but definitely not every time. The
> earliest precommit command on the PR 10346 for commit 2bf1a2b was
> triggered[1] on 10:26 AM PST or 6:26 PM UTC (Jenkins uses UTC). The
> earliest test run triggered by jenkins on commit 2bf1a2b that I see in the
> builds was at 3:21 PM PDT or 11:21 PM UTC [2], possibly triggered by the
> command on the PR at 3 PM PDT [3]. At least 4 commands to trigger the suite
> were ignored...
>
> Hopefully whatever recent change that caused this will be reverted once we
>> move to the new Beam Jenkins master. (https://ci-beam.apache.org).
>
>
> Do we have a JIRA that tracks this migration?
>
> Thanks
>
> [1] https://github.com/apache/beam/pull/10346#issuecomment-578885566
> [2] https://builds.apache.org/job/beam_PreCommit_Python_Phrase/1414/
> [3] https://github.com/apache/beam/pull/10346#issuecomment-578996626
>
>
> On Tue, Jan 28, 2020 at 9:37 AM Udi Meiri <eh...@google.com> wrote:
>
>> Hopefully whatever recent change that caused this will be reverted once
>> we move to the new Beam Jenkins master. (https://ci-beam.apache.org).
>>
>> On Tue, Jan 28, 2020, 00:39 Ismaël Mejía <ie...@gmail.com> wrote:
>>
>>> I also noticed that after the changes on Jenkins by Infra in multiple
>>> PRs we need to trigger the message twice (or 3 times) and then the run
>>> starts.
>>>
>>> On Tue, Jan 28, 2020 at 9:36 AM Kamil Wasilewski <
>>> kamil.wasilewski@polidea.com> wrote:
>>>
>>>> I observe that although sometimes no status checks appear, the job is
>>>> actually submitted to Jenkins. It seems this happened in your case as well:
>>>> https://builds.apache.org/job/beam_PreCommit_Python_Phrase/1417
>>>>
>>>> Also, does anyone know a reliable way to trigger all precommit jobs for
>>>>> a PR, instead of calling out each precommit job separately?
>>>>>
>>>>
>>>> I believe `retest this please` does the trick.
>>>>
>>>> On Tue, Jan 28, 2020 at 3:16 AM Valentyn Tymofieiev <
>>>> valentyn@google.com> wrote:
>>>>
>>>>> It appears that Jenkins sometimes does not respond to the trigger
>>>>> commands for extended periods of time (see image below from
>>>>> https://github.com/apache/beam/pull/10346). Does anyone know why this
>>>>> happens?
>>>>>
>>>>> Also, does anyone know a reliable way to trigger all precommit jobs
>>>>> for a PR, instead of calling out each precommit job separately?
>>>>>
>>>>>
>>>>> [image: image.png]
>>>>>
>>>>