An application master that runs the user's driver program and allocates executors.
Version of org.apache.spark.deploy.yarn.ClientBase tailored to YARN's stable API.
The entry point (starting in Client#main() and Client#run()) for launching Spark on YARN.
Client side methods to setup the Hadoop distributed cache
An application master that allocates executors on behalf of a driver that is running outside the cluster.
Contains util methods to interact with Hadoop from spark.