The Executing protocol currently holds the consumer with an address and thread Id. However when submitting a new task to the node it doesn't verify that the thread that runs the task is the one desired. This can cause the coordinator to have problems when re-registering the node as it may already be registered.