To build run
Add % provided to spark dependency before you submit it to the cluster. example: ("org.apache.spark" %% "spark-core" % "1.0.0" % "provided").
sbt assembly
Copy the file in target/...assembly..jar to the server. Da shit: scp -P 20022 dev/SparkPlayingField/target/scala-2.10/sparkplayingfield-assembly-0.1.jar hduser@131.220.9.194:/home/hduser/
Ssh into the server and rename the file to maderfaker.jar
To run use this:
bin/spark-submit --class org.menthal.NewAggregations --master yarn-cluster --executor-memory 500M --num-executors 3 ../maderfaker.jar yarn-cluster hdfs:///user/hduser/events.small