#334 Zuul: doesn't start right now
Closed by fbo. Opened by jaruga.

See https://src.fedoraproject.org/rpms/ruby/pull-request/110#comment-101517 .
What's wrong?

Here is the current time.

$ date -u
Wed Mar 30 05:24:23 PM UTC 2022

@fbo PTAL

Metadata Update from @mvadkert:
- Issue priority set to: High (was: Medium)
- Issue tagged with: Zuul CI

This has been the case for the entire day. The latest koji build in https://koji.fedoraproject.org/koji/tasks?owner=zuul&state=all is from yesterday.

@msrb told me this might have to do with an internal outage of something called PSI.

@msrb woot? how can that be related?

The service which listens for messages and triggers Zuul is running on the internal OCP4 now.

🤦 that is very bad idea :( added to OSCI-2908

Thanks @msrb I confirm that since OCP4 is down Fedora message bus events could not be processed by Zuul. The related service is: https://pagure.io/fm-gateway

The service could be run elsewhere obviously. Or, probably, could be removed if we could configure Pagure, automatically to send events to Zuul. Today AFAIK this would require a manual configuration from the package maintainer.

I am doing a summary in OSCI-2908, this is one of the thing there. Let me finish that and schedule a mtg for next week to discuss how to move forward.

I think that make sense to migrate the service on: https://fedora.softwarefactory-project.io/
So I propose that patch https://softwarefactory-project.io/r/c/software-factory/sf-infra/+/24468 then I need to wait for my teammates to review. Then after as we'll have to update the source IP authorized list then a Zuul restart will be needed.

@fbo ty, makes sense, it is great it is this easy ... the outage is still ongoing ... for second full day

The fm-gateway has been moved to the new host, zuul-web service has been restarted and jobs are starting now.

I close here. Please re-open if needed.

Metadata Update from @fbo:
- Issue status updated to: Closed (was: Open)

Metadata