cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Chip Childers <chip.child...@sungard.com>
Subject Re: Reviewboard and jenkins
Date Fri, 26 Jul 2013 00:28:02 GMT
Hugo,

Any chance you could add some logic to pick the right branch for the build
test?


On Mon, Jul 22, 2013 at 6:06 AM, Hugo Trippaers <hugo@trippaers.nl> wrote:

> The job is running on jenkins.cloudstack.org
>
> This is de job doing the work:
>
> http://jenkins.cloudstack.org/view/management/job/mgmt-build-reviewboard-requests/
>
> This is the job being executed with the patch:
>
> http://jenkins.cloudstack.org/view/master/job/cloudstack-master-with-patch/
>
> I really like the pipeline ideas, but we need to be careful with the
> timing and the external factors influencing builds. If we run complex
> chains there is a possibility that jobs will fail due to external factors.
> This would reflect badly on reviewboard as that review would be marked as
> "bad" for a reason that has nothing to do with the patch itself.
>
> For now i would like to stick with just the unit tests (master build) and
> when this proves stable we can add other builds in the chain?
>
> Cheers,
>
> Hugo
>
>
> On Jul 20, 2013, at 6:58 PM, Prasanna Santhanam <tsp@apache.org> wrote:
>
> > On Sat, Jul 20, 2013 at 09:46:59AM -0400, David Nalley wrote:
> >> On Sat, Jul 20, 2013 at 5:53 AM, Hugo Trippaers <trippie@gmail.com>
> wrote:
> >>>
> >>>
> >>> Sent from my iPhone
> >>>
> >>> On 20 jul. 2013, at 10:09, Prasanna Santhanam <tsp@apache.org> wrote:
> >>>
> >>>> On Sat, Jul 20, 2013 at 12:20:11AM +0200, Hugo Trippaers wrote:
> >>>>>
> >>>>> On Jul 19, 2013, at 7:15 PM, Daan Hoogland <daan.hoogland@gmail.com>
> wrote:
> >>>>>
> >>>>>> good stuff, does it run the risk of running while another instance
> is busy?
> >>>>>
> >>>>> Probably not, it's a single threaded script. I intend to use jenkins
> >>>>> to schedule it and i can tell jenkins not to start another instance
> >>>>> of a job once one is running.
> >>>> should set a "quiet period" for the job so that it doesn't kick up
> >>>> jobs for each checkin but a group instead. we'll run out of executors
> >>>> fast if that happens.
> >>>
> >>> Every time the script runs it will look at the newest 20 submissions
> and process them in order. That should be ok for now.
> >>>
> >>
> >> I actually wonder if we shouldn't pipeline this - e.g. - does the
> >> patch apply
> >> If it applies does it build (and have unit tests pass)
> >> If it builds does it package
> >> If it packages, do the integration tests pass (or given the long
> >> running nature, perhaps we have stages of that as well)
> >>
> >> The first three are pretty easy to scale using on-demand resources.
> >
> > Yup that's what I'd like to do with this as well. The simulator
> > pipeline that Ian is helping build will tie in very well with this.
> >
> >>
> >>
> >>>>
> >>>>> There are probably some bug still in the script, but hey we got
to
> >>>>> start somewhere. ;-)
> >>>>
> >>>> Absolutely, could you put it up on cso-infra/github so we can hack it
> >>>> and customize?
> >>>
> >>> Absolutely, will put it there this weekend if I have some time. I
> don't have enough permissions to setup a new repo in CloudStack extras yet.
> >>>
> >>
> >>
> >> You do now.
> >
> > --
> > Prasanna.,
> >
> > ------------------------
> > Powered by BigRock.com
> >
>
>
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message