What is the recommended approach to having cluster wide scheduling and job management? We want to have a single scheduler firing off jobs for any server in the cluster to pickup and run.
