YARN is a distributed resource management system.
YARN is the core component of the Hadoop system. Its main functions include resource management in the Hadoop cluster, scheduling, running and monitoring of jobs.
ResourceManager
Responsible for cluster resource management and scheduling, allocating resources for various types of jobs running on YARN.
The non-HA cluster is deployed on the Master node of EMR, and the HA cluster is deployed on multiple Master nodes of EMR to ensure high availability.
NodeManager
Responsible for node resource management, monitoring and job running.
Deployed on the Core or Task node of EMR.