The hierarchy is defined through level in the path yarn.scheduler.capacity.. For instance: The a level (yarn.scheduler.capacity.root.a.queues) is a level below root (yarn.scheduler.capacity.root.queues) * Several queues can be defined by level. * Children do not inherit properties ===== Configuration Management ===== ==== capacity-scheduler.xml ==== The configuration follows the hadoop configuration principles and are stored in the file capacity-scheduler.xml Example with 3 levels (root, a and b) and 8 queues (a,b,c,a1,a2,b1,b2,b3) yarn.scheduler.capacity.root.queuesa,b,cThe queues at the this level (root is the root queue).yarn.scheduler.capacity.root.a.queuesa1,a2The queues at the this level (root is the root queue).yarn.scheduler.capacity.root.b.queuesb1,b2,b3The queues at the this level (root is the root queue). More Configuration ==== Modification ==== === via Cli === Yarn - Yarn CLI vi HADOOP_CONF_DIR/capacity-scheduler.xmlHADOOP_YARN_HOME/bin/yarn rmadmin -refreshQueues === via Rest === Yarn - ResourceManager Rest API https://hadoop.apache.org/docs/current/hadoop-yarn/hadoop-yarn-site/ResourceManagerRest.html#Scheduler_Configuration_Mutation_API ===== Documentation / Reference ===== * https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.6.0/bk_yarn-resource-management/content/enabling_capacity_scheduler.html * https://hadoop.apache.org/docs/current/hadoop-yarn/hadoop-yarn-site/CapacityScheduler.html