Comment 3 for bug 1691570

After updating to a version of the charm that includes the ability to tune the rsync timeout, we upped the timeout to >12 hours, and replication is able to complete on many hosts, but some rsync processes are hanging or timing out on the server side and causing swift-object-replicator hangups polling filehandles for the rsync pipes which interrupts the every 5 minute log-dump of current replication status. This shows up after "lockup detected, killing coros..." shows up in the logs.