Is it possible to configure Ruby Spark to run against a cluster that uses yarn not meso? If so what is the syntax to configure Ruby Spark to do this?