Incorporated Patrick's feedback comment on #211 and made maven build/dep-resolution atleast a bit faster.

Allow spark on yarn to be run from HDFS. Allows the spark.jar, app.jar, and log4j.properties to be put into hdfs.

This uses a seperate Hadoop version for YARN artifact. This means when people link against
spark-yarn, things will resolve correctly.

Also rename parent project to spark-parent (otherwise it shows up as
"parent" in IntelliJ, which is very confusing).

See the discussion at https://github.com/mesos/spark/pull/355 for why
default profile activation is a problem.

by using -Dhadoop -Phadoop2.

This will make it easier to make the "run" script work with a Maven build