3.5. Upgrade the YARN ResourceManager process or HA process pair

If High Availability is not enabled for the YARN ResourceManager, shut down the ResourceManager process, switch to the newer version, and start the ResourceManager process:

su - yarn -c "/usr/hdp/current/hadoop-yarn-resourcemanager/sbin/yarn-daemon.sh stop resourcemanager"

hdp-select set hadoop-yarn-resourcemanager 2.2.6.0-2800

su - yarn -c "/usr/hdp/current/hadoop-yarn-resourcemanager/sbin/yarn-daemon.sh start resourcemanager"

If High Availability is enabled for the YARN ResourceManager, you’ll need to upgrade both ResourceManager processes, starting with the standby process. This is similar to the process you used for the NameNode pair:

  1. Upgrade the standby process

  2. Failover from active to standby

  3. Upgrade the new standby process (formerly the active process)

In the following instructions, RM1 refers to the currently active ResourceManager. RM2 refers to the current standby ResourceManager.

  1. Shut down the standby Resource Manager (RM2), switch to the newer version, and start RM2:

    su - yarn -c "/usr/hdp/current/hadoop-yarn-resourcemanager/sbin/yarn-daemon.sh stop resourcemanager"

    hdp-select set hadoop-yarn-resourcemanager 2.2.6.0-2800

    su - yarn -c "/usr/hdp/current/hadoop-yarn-resourcemanager/sbin/yarn-daemon.sh start resourcemanager"

    RM2 will start, and will become the standby ResourceManager.

  2. To switch the standby ResourceManager, stop the active ResourceManager:

    su - yarn -c "/usr/hdp/current/hadoop-yarn-resourcemanager/sbin/yarn-daemon.sh stop resourcemanager"

  3. Shut down the new standby ResourceManager (RM1), switch to the new version, and start RM1:

    su - yarn -c "/usr/hdp/current/hadoop-yarn-resourcemanager/sbin/yarn-daemon.sh stop resourcemanager"

    hdp-select set hadoop-yarn-resourcemanager 2.2.6.0-2800

    su - yarn -c "/usr/hdp/current/hadoop-yarn-resourcemanager/sbin/yarn-daemon.sh start resourcemanager"

    RM1 will start and become the standby ResourceManager. For version info, see http://<node-name>:8088/cluster/cluster.


loading table of contents...