Merge pull request #341 from ash211/patch-5

Clarify spark.cores.max in docs

It controls the count of cores across the cluster, not on a per-machine basis.