Spark config

Use global Hive metastore

Repartition non-HDFS inputs in

Persist intermediate RDDs on

* “2” = replicate each partition on two cluster nodes

Override configuration

Define here Spark configuration keys that should be overridden for this job