Use oslo_concurrency for craeterepo locking
I observed the deadlocked createrepo() method in Red Hat internal
copr instance, too. Seems like there's no chance to do something
wrong, so probably there's some bug in lockfile.FileLock class.
OpenStack moved from lockfile to oslo_concurrency, and it seems to
be more supported code nowadays. Small benefit (if anything) with
oslo_concurrency that it is much easier to keep track of the lock
files because we can specify the lock_path= parameter.