Hey, I have to report the same issue again, the scratch build should run on some PRs but nothing happens today: https://src.fedoraproject.org/rpms/python-sphinxcontrib-bibtex/pull-request/3 - here, commit was added an hour ago (atm), or here: https://src.fedoraproject.org/rpms/python-django/pull-request/28 Triggering through [citest] doesn't do anything visible to me.
Also:
https://src.fedoraproject.org/rpms/ansible-core/pull-request/10 (18 hours ago) https://src.fedoraproject.org/rpms/python3.11/pull-request/63 (today)
Zuul works.
See also https://lists.fedoraproject.org/archives/list/ci@lists.fedoraproject.org/message/PFBM32NGUAXCS3NSICD3D7ODTR2KJ7NH/
This is not only about tests for PRs. I wait for gating tests for normal builds since yesterday. An example https://bodhi.fedoraproject.org/updates/FEDORA-2022-88b9abd432. I guess something died in the clouds and there is no human to supervise it.
I confirm that Zuul CI is working w/o issue at the moment. python-django and python-sphinxcontrib-bibtex are not part of the Zuul CI config [1], [2] thus Zuul does not run on those repositories.
[1] https://fedora.softwarefactory-project.io/zuul/projects [2] https://pagure.io/fedora-project-config/blob/master/f/resources/fedora-distgits.yaml
@fbo: Do we need to add projects to Zuul CI config, to have fedora-ci do scratch builds and run the gating tests when new PRs are submitted? Is this a new change?
We have two systems: Zuul and Fedora CI Jenkins https://osci-jenkins-1.ci.fedoraproject.org/
The second one picks events automatically from message bus. Zuul needs to be configured to recognize the project: https://fedoraproject.org/wiki/Zuul-based-ci#Add_the_repository_into_the_Zuul_configuration
Gating tests which happen after the merge are handled by Jenkins.
I see that we had a large queue there, and Jenkins has been restarted. Let's see if it works now.
All the static nodes connected to the instance got disconnected and did not auto-reconnect. Since Jenkins was slow, I killed the pod and that was a mistake -- this instance struggles to recover after a non-clean restart. For some reason, it runs on 2.5 CPUs(!) and a slow disk.
If we are not going to get rid of this instance anytime soon, we should definitely fix the problem with resources and add monitoring.
Has the hub recovered? I can see that fedora-ci.koji-build.rpmdeplint.functional and fedora-ci.koji-build.rpminspect.static-analysis results are collected in a timely manner, but results for fedora-ci.koji-build.tier0.functional never appears. E.g. here https://bodhi.fedoraproject.org/updates/FEDORA-2022-b03c209f64 I had to waive the gating after retriggering tests and waiting for 5 hours for a result.
Now I got the very first result for fedora-ci.koji-build.tier0.functional https://bodhi.fedoraproject.org/updates/FEDORA-2022-c36997f342.
Log in to comment on this ticket.