Several of the Coconut failures for Alpha TC5 seem to be to do with
mirror issues. It might be best to have Coconut use
dl.fedoraproject.org ?
I'll let someone more knowledgeable respond to this, but I assume you're referring
to those HTTP 404/550 issues during anaconda package download, which we see heavily on the
Boston machine, and less frequently on the Brno machine.
In general, yes, I think it would be worth to use a specific mirror, for several reasons.
In Boston, it might be reasonable to use dl.fp.o just to ensure we use the very latest
package versions (otherwise the repo we hit might be up to 2 days old). In Brno, it might
be reasonable to use our local mirror, even in case it's 1-2 days old, just because of
the download speedups. However, this can't be used in all test cases. If we want to
test the absolutely default installation, we can't set a custom mirror. If we want to
test "closest mirror" feature, we can't set a custom mirror. In many cases,
yes, it would help either speed or reliability.
Please note, though, that we think that the mirror issue is a legitimate bug, either
caused by dnf stack, or anaconda. Jan reported it yesterday here:
https://bugzilla.redhat.com/show_bug.cgi?id=1196164
If this turns out to be true, it would also mean that OpenQA helped us uncover quite an
interesting bug, which would be much harder to spot without the massive number of
installations executed regularly. So there is also some value in using closest mirror as
installation source.