[Buildbot-devel] Problem with Triggerable Schedulers
Gareth Armstrong
gareth.armstrong at hp.com
Thu Sep 3 10:59:38 UTC 2009
Hello again,
just to close the thread and apologize for the noise, I found my problem.
One of the triggered schedulers was firing a builder on an offline
buildslave, and because I had used "waitForFinish=True", that scheduler
was waiting obediently to do its job.
Apologies again,
Gareth
On 08/31/2009 05:05 PM, Armstrong, Gareth wrote:
> Hello Folks,
>
> I would like to pick your brains on a problem that I have. I have
> recently changed projects and am in the process of putting in place a
> buildbot for my new group. Having a little time to experiment, I
> decided to get a little more elaborate with my buildbot setup.
>
> Throwing KISS to the wind, I set up some Triggerable schedulers and
> Triggers steps similar to the example given for Triggerable Scheduler in
> Section 4.5.7 of the current buildbot manual
> (http://djmitche.github.com/buildbot/docs/0.7.11/#Triggerable-Scheduler). My
> set master and slaves are run on Rhel5 with Twisted 8.20 and buildbot
> 0.7.11p3.
>
> I have a base scheduler with a builder which contains a buildfactory
> something like this:
>
> bf = BuildFactory()
> whenever_build_scheduler_name = ('%s-whenever-build-scheduler' %
> self.proj_info['name'])
> bf.addStep(Trigger(schedulerNames=[whenever_build_scheduler_name],
> waitForFinish=True,
> updateSourceStamp=True,
> haltOnFailure=True))
>
> pkg_scheduler_name = ('%s-package-scheduler' %
> self.proj_info['name'])
> bf.addStep(Trigger(schedulerNames=[pkg_scheduler_name],
> waitForFinish=True,
> updateSourceStamp=True,
> haltOnFailure=True))
>
> test_scheduler_name = ('%s-test-scheduler' %
> self.proj_info['name'])
> bf.addStep(Trigger(schedulerNames=[test_scheduler_name],
> waitForFinish=True,
> updateSourceStamp=True,
> haltOnFailure=True))
>
> Firing the main scheduler works great and the three triggered schedulers
> are queued. However, whether or not the buildfactory of the first of
> these triggered schedulers passes or fails, the upstream scheduler (the
> one that launched the others) considering itself to be still building.
>
> Are there are gotchas or is this a bug?
>
> All the best,
>
> Gareth
>
>
--
More information about the devel
mailing list