What is the preferred way to execute mass indexing operation in multi-node environment, among the below options,
* invoking only from one of the nodes
* invoking from all the nodes concurrently
