[heketi-devel] Periodic (nightly) centos-ci runs

Niels de Vos ndevos at redhat.com
Fri Feb 16 10:44:15 UTC 2018


On Tue, Feb 06, 2018 at 09:43:45AM -0500, John Mulligan wrote:
> As we discussed this morning I think it would be nice to have regular runs
> of the centos-ci scripts on the master branches of our major projects
> (starting with heketi and gluster-kubernetes), not just on our PRs. I think
> this would help us find issues with our "upstreams" like when vagrant,
> kubernetes, gluster, etc. changes something. It may also help us tell the
> difference between common test failures vs. a new test failure introduced
> by a PR.

Changes in the CI have been made to run the tests on daily basis (@daily
in the Jenkins trigger). The following pull requests should get merged
to get everything in sync in Jenkins:

- centos-ci/gluster-kubernetes: run the job once a day
  https://github.com/gluster/glusterfs-patch-acceptance-tests/pull/123
  This PR also contains a fix for 'Vagrant fails to download box', which
  currently still makes the job fail.
  -> https://ci.centos.org/view/Gluster/job/gluster_kubernetes/

- centos-ci/heketi-functional: run the job once a day
  https://github.com/gluster/glusterfs-patch-acceptance-tests/pull/124
  -> https://ci.centos.org/view/Gluster/job/gluster_heketi-functional/


I am not sure how monitoring of the jobs is done best. At the moment the
same jobs are used for GitHub PRs and daily runs. That means the
"Embeddable Build Status" will fluctuate with the results of the PR
testing.

Niels


More information about the heketi-devel mailing list