Spark cluster and Livy sessions

[ghukill]

Moving to Spark cluster on Combine server, as opposed to "local" instance via Livy, prompts a few todos, and provides a good opportunity to revisit how users manage Livy sessions

• provide links to Spark cluster GUI @ :8080
• allows users to restart Spark cluster via supervisor processes
  • all would be restarting spark:, or driver at spark_driver and worker at spark_worker
• rename "Livy/Spark Session" link to "Spark Cluster" or something more general?

As for Livy sessions, users should have the ability to start/stop/restart them if desired (e.g. code changes), but shouldn't have to concern themselves with them for running Jobs.

• consider setting in localsettings, AUTOSTART_LIVY_SESSION that would automatically fire a Livy session if not active
• refactor some of the code for confirming active livy sessions
  • specifically when CombineJob fires a Job, if a session is not active, it just sets as False and fails

[ghukill]

Spark cluster mostly in place, confirmed new build on dev branch with Ubuntu 18.04. Impetus for this work was better handling Livy, looking into that now...

[ghukill]

Closing, in favor of more specific issues.