1. Start a 3 broker cluster, do not create any replicated queues
2. Promote a broker to primary, wait for it to be "active".
3. Kill the primary, promote another broker to primary
Expected result: The new primary broker becomes "active"
Actual result: The new primary remains "recovering"