#12561 Fedora Atomic Desktops Bootable Containers are not updated on the registry
Closed: Fixed by siosm. Opened by siosm.

  • Describe the issue

The Fedora Atomic Desktops Bootable Containers images haven't been updated in the registry for a while (15 days for F40 & F41, a month for Rawhide):
- https://quay.io/repository/fedora/fedora-silverblue?tab=tags
- https://quay.io/repository/fedora/fedora-kinoite?tab=tags
- https://quay.io/repository/fedora/fedora-sway-atomic?tab=tags
- https://quay.io/repository/fedora/fedora-budgie-atomic?tab=tags

The composes are successful:
- https://pagure.io/releng/failed-composes/issue/7802
- https://pagure.io/releng/failed-composes/issue/7803
- https://kojipkgs.fedoraproject.org/compose/rawhide/latest-Fedora-Rawhide/compose/Kinoite/x86_64/images/
- https://kojipkgs.fedoraproject.org/compose/updates/Fedora-41-updates-20250203.0/compose/Kinoite/x86_64/images/

So it's likely that it's the sync to the registry that is failing somewhere.

  • When do you need this? (YYYY/MM/DD)

No urgency as those are still experimental images and not the default.

  • When is this no longer needed or useful? (YYYY/MM/DD)

N/A

  • If we cannot complete your request, what is the impact?

This delays the move to those images for the future of the Atomic Desktops. See: https://pagure.io/releng/issue/10399


Huh. cloud-image-uploader should be doing these now (as far as I remember)...

CC: @jcline

Hmph. There's not much in the logs to indicate any problem, I have a suspicion something is filtering them out. I'm going to make some logging improvements and stare at this a bit more.

So, the last build of the uploader was in December. It definitely was handling them both because they're updated past that and I've got an alert for a failure:

[2025-01-14 01:31:55,917 fedora_image_uploader.handler ERROR] stderr: time="2025-01-14T01:31:55Z" level=fatal msg="writing blob: checking whether a blob sha256:5450689b19511141c881ae4f6fec7e60c1f9a4bc7b56eaba55b6ef8a2cca388e exists in quay.io/fedora/fedora-silverblue: received unexpected HTTP status: 502 Bad Gateway"

The currently-running pod was deployed Jan 22, 2025, 7:04 PM which is suspiciously close to when publishes stopped happening. Interestingly, it's happily handling the bootc images still...

Ah. I see what it is. The images are filtered by type, and the current compose reports all these images as the "bootable-container" type, not "ociarchive".

Metadata Update from @phsmoura:
- Issue tagged with: high-gain, medium-trouble

I believe https://pagure.io/cloud-image-uploader/pull-request/48 should fix it, although I'm assuming nothing else changed about the image (e.g. compression and such)

This appears to fixed now as far as I can see. Thanks!

Metadata Update from @siosm:
- Issue close_status updated to: Fixed
- Issue status updated to: Closed (was: Open)

yes, sorry, that was my fault. somehow forget to update c-i-u for the metadata change :(

Metadata