Example: https://koji.fedoraproject.org/koji/taskinfo?taskID=34175660
LibrepoException: (19, u'Yum repo downloading error: Downloading error(s): repodata/b3fb7268879ba81d16872f95d2cc089796a7ecbfc2aca04cbe22ded9082bb427-primary.sqlite.bz2 - Cannot download, all mirrors were already tried without success', u'Usable URL not found')
It happens all the time in module building. See: https://koji.fedoraproject.org/koji/tasks?state=failed&owner=mbs/mbs.fedoraproject.org&view=tree&method=all&order=-id
Could you fix this?
Short answer: Probibly not.
I wonder if we are now doing too many mbs builds at once and swamping it.
I'll try and do what I can, but I'm not sure what that might be right now.
So how will we do mass-rebuild this time? If s390x builders can't handle load. Will it be same as last mass-rebuild where thousands of builds failed due to s390x builders.
I noticed one of the kojipkgs backends was down. It's back now.
Can you tell me if it's any better?
I think so because my builds succeeded.
Also I have submitted patch for Koji which should improve situation a lot -- https://pagure.io/koji/pull-request/1398.
ok, lets close this then until the next time it has some issue. I hope it will be smooth for a while at least and that koji patch should really help too.
Metadata Update from @kevin: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)