This allows sparkd and the api server to run simultaneously in the same docker, w/ independent data (on disk)