On Mon, Dec 12, 2016 at 4:47 PM, Barak Korren <bkorren(a)redhat.com> wrote:
You could've asked me, I explained exactly this on one of the threads
that discuss these failures.
Its not using the default configs though - just the none-proxies ones,
those conf files are managed by us as well.
That's exactly why I am writing this e-mail. So thanks for your reply and
sorry for missing that mail thread.
Neah, its good to have around so we're resilient to proxy server failures.
Since the check is based on the URL in the *-proxed.conf file, it
means that when you change it to not use repoproxy, this check will
not use it as well.
I see. Maybe the best thing we can do is to remove metalink from non
proxied configs and put there the same set of mirrors configuration we use
in proxied, but just without proxy. This should remove that point of
failure on Fedora side while keeping this mitigation for our proxy to be
down.
--
Anton Marchukov
Senior Software Engineer - RHEV CI - Red Hat