Blocker bug process is serial in nature and every week it takes a lot of effort and time to go through each bug and co-ordinate with the package maintainer to look into. The idea is to have a blocker bug application that people can work on blocker bugs in an async manner.
Metadata Update from @kparal: - Issue assigned to kparal
I have an unfinished proposal in my queue for using pagure (or other systems) instead of live meetings. I'll try to finish it soon and send it out.
The proposal has been submitted here: https://lists.fedoraproject.org/archives/list/test@lists.fedoraproject.org/thread/KWO7L3FDEJW2OC2YFUDYZTQUDGAROX3B/
PoC implementation tracker ticket is here: https://pagure.io/fedora-qa/blockerbugs/issue/94
Commit 9bc2ee91 relates to this ticket
Commit cd60ff43 relates to this ticket
The current status is that the PoC has been fully implemented. BlockerBugsApp can now interface with a Pagure repo, create tickets for each new proposed blocker/FE, and we have a discussion watcher that looks for votes (in a defined format) and updates the ticket description with voting summary. It has been tested in our development environment. It hasn't been deployed to staging/production yet, because we rely on new Pagure API that hasn't been deployed yet (at least into production, it should arrive in staging very soon or perhaps arrived already, we'll check). So there might be some bugs, but overall the PoC is completely and we're blocked on Pagure deploying its new version into production.
Okay, blockerbugs for testing async process has been deployed: http://blockerbugs-blockerbugs.apps.os.fedorainfracloud.org/
stg.pagure.io already supports everything that's necessary and repo for testing the process has been set up: https://stg.pagure.io/fedora-qa/blocker-review
I did basic sanity check of deployment set, webhooks, sync, counting, API access... all seem to work just fine.
Everybody from Fedora QA should have access there. @sdharane , you don't seem to have an account on stg.pagure.io, so I didn't add you there.
@kparal It's using production BZ, I don't thinks it's an issue, but that can be changed through OpenShift console in a minute or so, if partner BZ is desired.
We found some bugs, so @lbrabec will work to address them. Current completion status is 90% (5% to fix bugs, 5% to deploy to production once we can).
Status update: We can't deploy the new BBA update (supporting async discussions) to production, because the VM runs on RHEL7 and we need Fedora 32 because of new dependencies. Upgrading the VM requires time and effort. And the VM will need to go away anyway, so deploying there would mean some downtime in the future (which would be more inconvenient if async discussions were involved). The idea was that we deploy the new BBA to communishift once the infra move is over. However, communishift future looks uncertain at the moment. We can also deploy it to Infra's openshift, and @frantisekz is learning to work with it. There is no time estimate atm, though. Frantisek might have more information in a week or so.
The review discussions feature is now LIVE at https://pagure.io/fedora-qa/blocker-review . I believe we can close this :-)
Metadata Update from @kparal: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
Log in to comment on this ticket.