The cluster chunks resize routine should be resilient to collection UUID changes caused by renameCollection()

It has been observed that the execution of renameCollection with dropTarget=true may cause ClusterChunksResizePolicy to hang the config server if the target collection is an existing sharded namespace that is currently being processed by the policy.

The root cause of the problem is ClusterChunksResizePolicy referencing the target collection with a UUID that does not match the content of config.collections once renameCollection completes.