Running Spark jobs via Marathon

[B. Svenjord]

I've got a Marathon job running for the following:

./bin/spark-class org.apache.spark.deploy.mesos.MesosClusterDispatcher

However, following that I'd like to now be able to run individual Spark jobs as separate Marathon jobs with the command:

./bin/spark-submit ....

My question is:

how can I call spark-submit from a Mesos executor of the binaries are not installed on it? 

(Note: I'm aware that http://spark.apache.org/docs/latest/running-on-mesos.html#connecting-spark-to-mesos also recommends installing Spark on all the Mesos slaves, but is that the only option?)

Any guidance is much appreciated.

[Durga swaroop Perla]

If you're looking to have Mesos manage your spark cluster, then that is the only way. You need to have spark binaries on all the mesos agents but If you just want a simple standalone cluster then you don't need to do that.