In some tests we perform removeShard/addShard operations, with the same shard being removed first and re-added later.
This can lead to issues, since a shard that was re-added can become "visible" to some pending operation, before the given addShard completes.
One such failure could be seen in BF-33741. 

The goal of this ticket is to ensure that when adding a shard to the cluster, we add a fresh shard with a unique hostname/port [preferred solution, other solutions tbd].