nicomak/spark repository overview

⁠To start the master:

docker run -d --name sparkmaster --net=host -e ROLE=master nicomak/spark:2.0.1

⁠To start slaves:

docker run -d --name sparkslave --net=host -e ROLE=slave -e MASTER_URL=spark://[master_host]:7077 nicomak/spark:2.0.1

⁠More info

• The --net=host network mode if prefered because Spark opens a lot of random ports.
• Any Spark environment variables can be defined in the container run command through -e options.

⁠Reference

More details about this image on my blog: http://blog.ditullio.fr/2016/11/09/docker-spark-analyse-cassandra-data/⁠