Well I've been wracking my brains but I can't think of a great way to solve this :( I think the 'theoretically ideal' way would be to have openQA do the work between the POST request and the job being run, as it does for use of e.g. ISO_URL, but in practice it's hard to achieve. We'd have to get some extremely Fedora-specific code upstreamed, and I don't think that's really a good idea. We can't do it as a plugin, I don't think, as I don't think we can hook a plugin into the openQA workflow in the appropriate place (between the POST request and the job being unblocked and sent to workers).

We could implement some kinda little middleware service that sits in front of openQA, steals the POST requests, does this job, then forwards them to openQA, but ugh, that seems like a whole lot of engineering and I don't know if it's worth the effort.

On the whole I might just unpick this (which sadly isn't trivial because we did the work to support testing side tags since it landed, so it'll be a fairly manual "reversion") and go back to the previous approach of each test doing all this stuff. It's silly and wastes time, but I can't come up with anything better right now.

OK, I have a revert to the old way of doing things written and in testing on staging (on the revert-update-images branch). I'm also testing a change to then have the tests use an asyncio-based Python script to do the actual package downloads concurrently, instead of doing them one at a time, but it's taking a while (currently my attempt to limit the downloads to batches of 20 at a time isn't working, and I need to find a way to not have the test have to type the entire huge list of NVRs as part of launching the script...)

https://pagure.io/fedora-qa/os-autoinst-distri-fedora/pull-request/311 fixes this.

This issue has been migrated to Fedora Forge:
https://forge.fedoraproject.org/quality/fedora_openqa/issues/108

Please continue any further discussion there.