alias quay.io/fedora/fedora-coreos:stable to :latest in quay

[cgwalters]

If you go to https://quay.io/repository/fedora/fedora-coreos quay says: podman pull quay.io/fedora/fedora-coreos but that doesn't work because the runtime assumes :latest.

(This is definitely a quay bug)

But that said, it would seem to make sense for us to point :latest at :stable, right?

[dustymabe]

We discussed this in the community meeting today.

13:12:32  dustymabe | #agreed While this would provide benefit to new users who just want to pull FCOS
                    | and see what's inside we could see some potential problems with it and would like
                    | to defer implementing something like this until our update story for CoreOS
                    | Layering (https://github.com/coreos/fedora-coreos-tracker/issues/1263) is figured out.

To summarize the conversation a bit more:

Obviously being able to just pull quay.io/fedora/fedora-coreos (which will pull the :latest tag by default) is attractive because users will get $something on their first try.

There are potential problems with this though:

• Forcing users to know about streams and what stream they are on could be a nice forcing function to them being more educated about what our streams are and why they are useful.
• If a user tries to rebase using rpm-ostree rebase --experimental ostree-unverified-registry:quay.io/fedora/fedora-coreos what actually shows up in rpm-ostree status output and what is the user actually following?
  • We could make some logical assumptions here and fixup the tag client side in rpm-ostree code, but that requires assumptions and could be (probably is) FCOS specific.
  • If registries properly supported redirects then we could redirect quay.io/fedora/fedora-coreos:latest to quay.io/fedora/fedora-coreos:stable and there would be much less concern here because the client tooling would get and reflect :stable.
• This may have impact (unkown at this point) on our yet to be worked out implementation of zincati updates for OSTree Native Containers.

So while we work out some details we're electing to defer decisions on implementing this to a later point in time.

[cgwalters]

In the interest of keeping our issue list small, I'm going to close this for now.