Red Hat Bugzilla – Bug 922667
the cronned 'dnf makecache' runs should give up quickly if there's network/checksumming issues
Last modified: 2014-09-30 19:41:16 EDT
see Reindl Harald's email at: http://lists.fedoraproject.org/pipermail/devel/2013-March/180304.html
Fixed by d48bc7c. When 'dnf makecache' is started via the systemd timer unit, dnf only tries to synchronize the metadata from at most one mirror in the mirrorlist and does not try all of them in case the first one fails.