Never lose jobs with HA Jobtracker

MapR is the only Hadoop distribution with no single points of failure.

The MapR JobTracker HA improves recovery time objectives and provides for a self-healing cluster. Upon failure, the MapR JobTracker automatically restarts on another node in the cluster. TaskTrackers will automatically pause and then reconnect to the new JobTracker. Any currently running jobs or tasks continue without losing any progress or failing.

Unfortunately in other Hadoop distributions, the JobTracker runs only on a single node, and represents a single point of failure for the entire cluster. If the JobTracker fails, all running jobs fail and all progress is lost. In addition, administrators must first detect this situation, and then manually restart the JobTracker.