Open Stack

On 06/09/2015 05:33 AM, Gareth wrote:
> Hi infra team,
> 
> The origin concept of jenkins is "one trigger one job". For example, the
> job "neutron-unit-test-py27" could respond a new change set from neutron
> upstream. But the truth is that a gerrit event trigger a set of jenkins
> jobs: py27-unittest, pep8-check, and many dsvm jobs...
> 
> How can I configure my jenkins like this?
> 
> 
> 
> __________________________________________________________________________
> OpenStack Development Mailing List (not for usage questions)
> Unsubscribe: OpenStack-dev-request at lists.openstack.org?subject:unsubscribe
> http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
> 
We use a tool called Zuul written expressly for our needs which has a
concept of pipelines. The status of the patch in Gerrit (Workflow 0,
Workflow +1, merged) gives Zuul the information it needs to put a
patchset in a given pipeline based on a new event.

A repo will have a cluster of jobs set based on pipeline.

Here is the zuul configuration file for the openstack/keystone repo:
http://git.openstack.org/cgit/openstack-infra/project-config/tree/zuul/layout.yaml#n1832

There are templates which are common across many repos, then specific
jobs for the check pipeline, gate pipeline, post pipeline (which means
post merge) and the experimental pipeline (which we created as a way of
graduating new jobs).

Here would be a good place to begin reading about zuul:
http://docs.openstack.org/infra/system-config/zuul.html

Thanks,
Anita.