dylanwilder

This works well if configuration is uniform per node. However, the per resource determination is more flexible and could conceivably be used for more open ended cases.

Alternatively, cassandra supports running across multiple disks. Can this be supported as well?

@mesosphere-ci: retest this please

@triclambert is there a secret to the mesosphere-ci thing?

@mesosphere-ci: retest this please

retest this please

@mesosphere-ci: retest this please

@mesosphere-ci: retest this please

I can send a PR for this if no one objects.

+1