The NameNode server

This component has the responsibility of scheduling and tracking jobs spread across the cluster. Until Hadoop 1, it was a single point of failure, but with Hadoop 2, this responsibility was delegated to YARN, which creates multiple job trackers across the cluster and ensures failover.