On Mon, Feb 08, 2016 at 08:13:38AM +0900, fumihiko kakuma wrote:
> Hi Jeremy,
> 
> On Fri, 5 Feb 2016 14:45:17 +0000
> Jeremy Stanley <[email protected]> wrote:
> 
> > On 2016-02-05 09:00:24 +0100 (+0100), Andreas Jaeger wrote:
> > > One idea: Remove the branch filters you created.
> > [...]
> > 
> > Correct. From Zuul's perspective periodic jobs are triggered simply
> > by a changeless/refless timer, and thus branchless since there's no
> > associated Gerrit change or Git ref from which it can infer a
> > branch. Asking Zuul to limit a periodic job to specific branches
> > means it will never be run, as that precondition can never be
> > satisfied.
> > -- 
> > Jeremy Stanley
> 
> I see. Thank you for the explanation of the mechanism of periodic
> pipeline.
> 
> BTW can you also explain the non-indication of periodic pipeline data
> by openstack-helth and tell the resolution to it?

There is nothing wrong with openstack-health it's behaving as currently
expected. There is a known limitation with the dashboard right now where
results aren't counted if the job failure occurs before devstack starts. If your
jobs are running but never even getting to devstack openstack-health (well
really the subunit2sql db) will not have any data on those runs. Once you fix
these jobs to actually start running devstack (or anything else which generates
a subunit stream in the expected place) it'll appear on the dashboard.

I wrote a brief blog post a little while ago on how openstack-health works and
some of the features it has:

http://blog.kortar.org/?p=279

The current limitations section explains this issue in a bit more detail.

-Matt Treinish

Attachment: signature.asc
Description: PGP signature

_______________________________________________
OpenStack-Infra mailing list
[email protected]
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra

Reply via email to