to clarify, what exactly was your repair command, and in reference to a ring did you mean the DC or the cluster, and has the repair been running for 2 weeks or is that in reference to the "ring"?
It would be helpful if you provided the relevant logs as well, also, the cassandra version you are running.
