Fedora container uses fixed, currently broken, repository, maybe other solution possible?

[B5JohnSheridan]

Hi,

we are using the Fedora 37 LXC template (via Proxmox) and stumbled over the currently not working Fedory repository csclub.uwaterloo.ca (some dirs from the repo file apparently not existing, causing dnf to fail).

That made us wonder why it tries to connect especially to this repository until we tracked this down to here and the fedora.yaml, which explicitely disables the metalink and hardcodes this repository.

I'd at least like to point this out as this now leads to the issue with the broken repository and sheepishly comment that there is maybe a better choice than the hardcoded repository for the container?

Thanks!

[stgraber]

@monstermunchkin can we make sure we revert the repo config to the default values at the end of the image build?

[monstermunchkin]

Our current config does the following:

• backup all original repo files
• modify them (change mirror etc.)
• restore original repo files

I just built a fresh Fedora 37 image, and there are no traces of the csclub.uwaterloo.ca repo in the repo files.

@B5JohnSheridan are you using the most recent fedora template?

[stgraber]

Okay. I believe we changed that a few months ago, so it could be an old container or an old image being used here.

[B5JohnSheridan]

Thanks!

I tested with the most recent version of the container deployed with Proxmox VE (problem was still there), so I will ask them to consider this and update.

[stgraber]

Hmm, yeah, no idea how Proxmox handles image syncs or how they even consume our images in the first place.