fedora-archive.repo: add archives mirrors

[jbtrystram]

Add another set of mirrors for archives in case dl.fedoraproject.org/pub/archive timeouts or flake

We've had the pipeline fail for older releases on that so a set of extra mirrors can help

[dustymabe]

Hmm. didn't know about a second site. Are there any docs for it?

[travier]

I don't think that's the same thing. Those archives are for EOL Fedora releases, not for archived updates for current releases.

[dustymabe]

I don't think that's the same thing. Those archives are for EOL Fedora releases, not for archived updates for current releases.

I think we are using archives for EOL releases here so he is in the right context. From the file:

fedora-coreos-config/fedora-archive.repo

Lines 1 to 9 in 4b6d1fc

	# Certain kola tests on older RHCOS branches rely on EOL Fedora containers
	# to set up and run the test environment. The ITUP cluster, being used by
	# the RHCOS pipeline, requires all outbound connections to be explicitly
	# specified in a Firewall Egress file. By using these archive repos on
	# the older RHCOS branches, we can ensure that those kola tests will
	# always download archived content from `https://dl.fedoraproject.org`.
	# We include both EOL and non-EOL repository locations in the baseurl list
	# to simplify maintenance, avoiding the need to remove the fedora.repo file
	# as versions reach EOL.

[jbtrystram]

Hmm. didn't know about a second site. Are there any docs for it?

I discovered it by following the archive link that was in this .repo file for secondary arch and found this readme
If you look 1 level above there is nothing there, but only for secondary arches.

[dustymabe]

right, but it feels a bit awkward to have a link to both

• https://dl.fedoraproject.org/pub/archive/fedora-secondary/
• https://archives.fedoraproject.org/pub/archive/fedora-secondary/

i.e. I can almost guarantee they are just the same exact thing. In other words, if one of them fails they will probably both fail?

In the cases where you saw a failure did a rerun succeed (i.e. the automatic rerun that we do not a manually initiated rerun)?