Open Stack

To avoid zuul (v2.1.0) posting comments with no logs we need to backport
the patch mentioned by Jeremy (https://review.openstack.org/188383).

Clone zuul project and cherry-pick the mentioned patch:

- git clone https://review.openstack.org/openstack-infra/zuul
- cd zuul && git checkout tags/2.1.0
- git fetch https://review.openstack.org/openstack-infra/zuul
refs/changes/83/188383/3 && git cherry-pick FETCH_HEAD


Install patched zuul:
- python setup.py install


In zuul/layout.yaml file defines a filter for a job to be run on
openstack-dev/ci-sandbox patchset creation

jobs:
  - name: test-job
    files:
      - '^test-1$'
    voting: true

Run zuul-server as usual
Upload a patch for review modifying test-1 and verify that CI correctly
posts a comment with link to logs
Upload a patch for review modifying test-2 and verify that CI does not post
any comment


On Wed, Oct 21, 2015 at 5:02 PM, Anita Kuno <anteaya at anteaya.info> wrote:

> On 10/21/2015 10:58 AM, Francesco Santoro wrote:
> > Hi Anita,
> > thanks for your answer.
> >
> > Since we noticed that our CI posted comments with no logs we stopped
> > commenting on patches.
>
> Oh how heartening, thank you. Your level of responsibility regarding
> your third party ci's effect on the community makes me happy for the
> day. Thank you, Francesco!
>
> >
> > We were thinking to backport the patch mentioned by Jeremy in order to
> > avoid waiting for the next zuul release.
>
> That certainly is an option, yes.
>
> > The idea is to correctly test the feature on the openstack-dev/ci-sandbox
> > before starting commenting on other projects.
> > What do you think?
>
> Oh my goodness, two wonderful things you acknowledge in your email.
> Thank you so much. I look forward to meeting you some day and shaking
> your hand. Thank you Francesco. Yes, the ci-sandbox repo is the place to
> test third party ci operations. My irc nickname is anteaya, if you want
> me to look at any of your output while you are testing do ping me in
> #openstack-infra.
>
> Thank you so much, Francesco,
> Anita.
>
> >
> > Best regards,
> > Francesco
> >
> > On Wed, Oct 21, 2015 at 4:04 PM, Anita Kuno <anteaya at anteaya.info>
> wrote:
> >
> >> On 10/21/2015 09:17 AM, Jeremy Stanley wrote:
> >>> On 2015-10-21 13:36:41 +0200 (+0200), Francesco Santoro wrote:
> >>> [...]
> >>>> This setup is correctly working but zuul publishes comments also on
> >>>> other patches that don't trigger any job.
> >>> [...]
> >>>
> >>> It's because file and branch filters affect which jobs run, but
> >>> comments are tied to whether pipelines are triggered (not to
> >>> individual jobs).
> >>>
> >>> What you probably want is https://review.openstack.org/188383 which
> >>> will cause Zuul not to report for a pipeline at all if it has no
> >>> matching jobs for a change triggering it. This merged to master a
> >>> few weeks after 2.1.0 was released so isn't in an officially
> >>> released version yet. You could try running from the tip of the
> >>> master branch to get this improvement, or backport the patch to your
> >>> installed version, or wait for the next release.
> >>>
> >>
> >> Please note that should you choose to wait for the next release, having
> >> your third party ci system post comments to gerrit that contain no logs
> >> is contrary to the requirements of third party ci systems:
> >>
> http://docs.openstack.org/infra/system-config/third_party.html#requirements
> >>
> >> Jeremy was answering your technical inquiry about zuul and its versions.
> >> I am addressing the expectations of every third party ci operator. You
> >> are responsible for knowing and adhering to requirements.
> >>
> >> Thank you,
> >> Anita.
> >>
> >> _______________________________________________
> >> OpenStack-Infra mailing list
> >> OpenStack-Infra at lists.openstack.org
> >> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
> >>
> >
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-infra/attachments/20151021/a07bebc1/attachment.html>