Yarn Node Labels

Yarn Node Labels + Capacity-Scheduler

在yarn-site.xml中开启capacity-schedule

yarn-site.xml

<property>
    <name>yarn.resourcemanager.scheduler.class</name>
    <value>org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler</value>
  </property>

 

配置capacity-scheduler.xml文件

<configuration>
  <property>
    <name>yarn.scheduler.capacity.root.queues</name>
    <value>default,area0,area1,area2</value>
    <description>
      The queues at the this level (root is the root queue).
    </description>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.default.capacity</name>
    <value>25</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area0.capacity</name>
    <value>25</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area1.capacity</name>
    <value>25</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area2.capacity</name>
    <value>25</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.default.maximum-capacity</name>
    <value>100</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area0.maximum-capacity</name>
    <value>100</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area1.maximum-capacity</name>
    <value>100</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area2.maximum-capacity</name>
    <value>100</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.accessible-node-labels</name>
    <value>*</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area0.accessible-node-labels</name>
    <value>area0</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area1.accessible-node-labels</name>
    <value>area1</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area2.accessible-node-labels</name>
    <value>area2</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.accessible-node-labels.area0.capacity</name>
    <value>33</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.accessible-node-labels.area1.capacity</name>
    <value>33</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.accessible-node-labels.area2.capacity</name>
    <value>34</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area0.accessible-node-labels.area0.capacity</name>
    <value>100</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area1.accessible-node-labels.area1.capacity</name>
    <value>100</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area2.accessible-node-labels.area2.capacity</name>
    <value>100</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.default-node-label-expression</name>
    <value> ,area0,area1,area2</value>
  </property>
　<!--value为空，则root.default队列只能在没有label的node上执行，而我环境环境中不存在没有label的node，所以提交到root.default的任务卡死-->
  <property>
    <name>yarn.scheduler.capacity.root.default.default-node-label-expression</name>
    <value> </value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area0.default-node-label-expression</name>
    <value>area0</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area1.default-node-label-expression</name>
    <value>area1</value>
  </property>

  <property>
    <name>yarn.scheduler.capacity.root.area2.default-node-label-expression</name>
    <value>area2</value>
  </property>
</configuration>

添加Yarn Node Labels

#添加标签

yarn rmadmin -addToClusterNodeLabels area0,area1,area2
#节点添加标签
yarn rmadmin -replaceLabelsOnNode Host1:45454,area0
yarn rmadmin -replaceLabelsOnNode Host2:45454,area1
yarn rmadmin -replaceLabelsOnNode Host3:45454,area2
#查看标签
yarn node -status Host1:45454
#也可以通过Yarn管理页面查看Node Label

配置Yarn Node Labels存储目录，这样不用每次重启服务后重新配置Yarn Node Labels。

在yarn-site.xml中添加下列信息

 

<property>
<name>yarn.node-labels.enabled</name>
<value>true</value>
</property>
<property>
<name>yarn.nodemanager.address</name>
<value>0.0.0.0:45454</value>
</property>

<property>
<name>yarn.node-labels.manager-class</name>
<value>org.apache.hadoop.yarn.server.resourcemanager.nodelabels.RMNodeLabelsManager</value>
</property>
<property>
<name>yarn.node-labels.fs-store.root-dir</name>
<value>hdfs://Host0:8020/yarn/node-labels</value>
<description>标签数据在HDFS上的存储位置</description>
</property>