I have a Spark app which runs well on local master. I'm now ready to put it on a cluster. What needs to be installed on the master? What needs to be installed on the workers?
If the cluster already has Hadoop or YARN or Cloudera, does it still need an install of Spark?
