hi,
i’m using rpm-ostree native containers for customizing fc-silverblue but it seems the fedora-silverblue:39 container was last updated 16 days ago. so i’m not receiving any updates. is this because the fc40 release is near?
the contanierfile i’m using is located here. i’m using quay.io/fedora/fedora-silverblue:39 as the base image.
or should i switch to another image and the one i’m using is unsupported?
thanks for any help
toni
after reading this thread, should i move to using https://quay.io/repository/fedora-ostree-desktops/silverblue as a base? i’m i little bit confused right know which base image to use for my customizations. but i understand that is a moving target.
what’s is acutally the difference between quay.io/fedora/fedora-silverblue and quay.io/fedora-ostree-desktops/silverblue?
thanks for any hints
toni
I wouldn’t change yet until the infra is settled and working. So realistically, this is likely a couple of releases away for the initial compose/build system changes, then probably quicker to getting the OCI containers. But as you not, it’s a moving target right now.
quay.io/fedora/fedora-silverblue is a container image converted from the official Fedora ostree commits. It’s pushed by Fedora infra.
quay.io/fedora-ostree-desktops/silverblue is an unofficial build from fedora / Fedora Atomic Desktops / ci-test · GitLab done via GitLab CI (i.e. not strictly speaking official infra).
They both use the same sources and packages. This is all in flux until we complete: https://pagure.io/releng/issue/10399
I currently recommend using quay.io/fedora-ostree-desktops/silverblue.
This might be related to PR#1267: f39: fix container-nightly.sh script to sync the right thing - pungi-fedora - Pagure.io
hm, it seems there aren’t any tags available and i only get a blank page at https://registry.fedoraproject.org/repo/fedora-silverblue/tags/
thanks for clearing this up for me. i’m now using quay.io/fedora-ostree-desktops/silverblue and get updates so far. i’ll stick to this image for know and will monitor the issue you mentioned.
thanks for getting back.
toni
oh, yeah, that seems possible - I guess even though running that script there was obviously the wrong thing to do, it still did its job when it was called, and possibly we aren’t calling it from anywhere where it’s the right thing to do. @kevin wdyt?
Yep. It did. 
we forgot oci images…
We need to enhance the script bodhi uses to sync ostree commits for updates probibly to also generate/push the oci images. ;(
In the mean time I will manually do a sync now…
Oh, side note, the script never had Onyx in it… should that be synced too? Or were no oci images made of Onyx in f39?
Note that this scripts converts the current ostree commits into OCI images and then pushes them, which is not what we want to do in the future as we want to directly push the OCI images that we now build in another job instead.
This is related to:
Yeah. Agreed. Making the same thing over and over is bad. 
Perhaps after the f40 dust finally settles we can all meet up and come
up with actions and timeline and assign things, etc.
yeah, every time I try and come at this overall mess from one angle to make a little improvement, all the other angles show up