Open Stack

Excerpts from Anita Kuno's message of 2015-05-28 17:35:06 -0400:
> On 05/28/2015 04:46 PM, Jeremy Stanley wrote:
> > On 2015-05-28 13:30:44 -0700 (-0700), Clint Byrum wrote:
> > [...]
> >> Do we already have this hidden somewhere, or would it make sense to
> >> maybe add this as something in openstack-infra/project-config along side
> >> the jjb definition that creates the job/class of job somehow?
> > [...]
> > 
> > We don't (yet anyway). It's a little tricky since a "job" isn't
> > necessarily a particular configuration element but rather often
> > arises by instantiating a parameterized template. So the assembly of
> > a particular template along with a specific set of parameters is
> > what we might want to associate with a given contact. It's possible
> > we could add a contact parameter to these so that it accompanies
> > each job configuration as metadata and transform that into a
> > responsible parties list somewhere easy to reference... but I expect
> > there are plenty of alternative options for this. Also we should get
> > input from the QA team as well (tag added) since they'd be one of
> > the more frequent consumers of this information.
> > 
> I tend to default to the ptl of the repo on which the main job is
> running (for example if the job runs on cinder, devstack-gate and
> tempest, and contains the word cinder, I would ask the cinder ptl for
> further direction on the job).
> 
> This works mostly and for me is preferable to a secondary file which
> would be hard to keep updated.

The jobs Clint is thinking of map more closely to the third-party CI
jobs, where someone with specialty expertise on a messaging driver may
need to get involved in debugging a functional test failure. IIRC,
there's a wiki page for third-party CI owners, so maybe we should do
something similar for these jobs.

Doug