Create stable stream

[jlebon]

As we approach the stable release, we need to be ready to actually perform the release mechanically. There are a few things that need to happen for this:

• create the stable branch
• add the stable stream to the pipeline
• add instructions for making a stable release to the streams repo
• verify tools don't need to be adapted (plume release, fedora-coreos-stream-generator)
• verify Cincinnati is adapted to handle the stable stream (scraper: support multiple graphs (streams) fedora-coreos-cincinnati#4)
• update website to have a drop-down selector between stable and testing streams, defaulting to stable (though in a pinch, s/testing/stable/ would work too short-term) (Adapt FCOS download page for stable release #332)

(Chime in if there's anything else missing from there).

Seems like we should just go through all the steps of making a release to test this out, up to but not including actually updating the website?

[jlebon]

add instructions for making a stable release to the streams repo

coreos/fedora-coreos-streams#39

[lucab]

Cross-reference: coreos/coreos-installer#106

[jlebon]

verify tools don't need to be adapted (plume release, fedora-coreos-stream-generator)

I've had a quick look at them and I don't see anything offhand that would require adapting. Both of them parameterize on the stream to operate on.

[lucab]

Cincinnati is also already parametrized, but the backend will need a small update (possibly just its configuration) to start consuming stable updates metadata and release index.

EDIT: this actually needs patching the scraping and caching logic, see dedicated ticket.

[jlebon]

Cincinnati is also already parametrized, but the backend will need a small update (possibly just its configuration) to start consuming stable updates metadata and release index.

Good call, thanks. I added a checkbox for Cincinnati!

[lucab]

@jlebon do you plan to have the JSON metadata URL created in advance (and empty) to align the tools earlier, or do you prefer to use this event to exercise metadata creation logic?

[jlebon]

@jlebon do you plan to have the JSON metadata URL created in advance (and empty) to align the tools earlier, or do you prefer to use this event to exercise metadata creation logic?

Hadn't thought about it, though probably a good opportunity to test metadata creation logic? We still have quite a few more streams (mostly mechanical) that we'll want to set up eventually.

[jlebon]

Going to close this one out. Thanks all!