canonical-ci-engineering team mailing list archive

[Francis Ginther <francis.ginther@xxxxxxxxxxxxx>]

Sebastien,

We'll re-trigger any aborted builds that we find related to this shutdown
that haven't already been restarted. However, if you're still waiting for a
build 6 hours from now, please ping 'cihelp' or retrigger the build
manually as we may miss a few as there may not be a clear signature of the
builds that were aborted during the shutdown.

In the past this wasn't a problem because the jobs to comment on the MP
after a CI run were blocked from being run and destroyed when jenkins was
shutdown. This resulted in those MPs still being in the 'needs to be
tested' state and jenkins would pick them back up again after the restart.
The upgraded jenkins appears to be smarter now and is saving and
re-queueing these MP commenting jobs after the restart.

Apologies for the aborted results, we'll update our restart procedures to
avoid this in the future.

Francis

On Thu, Nov 19, 2015 at 12:57 AM, Sebastien Bacher <
sebastien.bacher@xxxxxxxxxxxxx> wrote:

> Le 18/11/2015 20:46, Francis Ginther a écrit :
> > Hello,
> >
> > http://s-jenkins.ubuntu-ci:8080/ requires some downtime to fix some
> > plugin issues brought about by the recent upgrade. This downtime
> > should last less then 30 minutes. I'm planning on performing the
> > downtime between 21:00 and 22:00 UTC today (about 1-2 hours from now).
> > No action is necessary after the downtime.
>
> Hey Francis,
>
> Was the issue you wanted to fix or the downtime leading to aborted CI
> reports on merge requests?
>
> examples
>
> http://s-jenkins.ubuntu-ci:8080/job/generic-click-autopilot-runner-touch/966/console
>
> http://s-jenkins.ubuntu-ci:8080/job/generic-deb-autopilot-vivid-touch/5245/console
>
>
> if so, is CI going to retry those or do we need to trigger a retry by
> some other way?
>
> Cheers,
> Sebastien Bacher
>



-- 
Francis Ginther
Canonical - Ubuntu Engineering - Continuous Integration Team