Managing Data Operating System
Also available as:
loading table of contents...

Enable System Service Mode On a Newly Installed Cluster

On a new HDP 3.0+ cluster that meets the capacity requirements for HBase to run in the system service mode, you must configure a separate queue for running ats-hbase.

Perform the following recommended steps to run ats-hbase on your cluster after installing HDP 3.0+.
  1. Configure a separate queue at the root level, say yarn system, for the ats-hbase service instead of using the default queue.
  2. Allocate required resources to the yarn-system queue for launching the ats-hbase service.
    The minimum amount of memory required for launching ats-hbase is 12 GB.
    To allocate resources, update the value of the yarn.scheduler.capacity.root.<queue_path>.capacity property in capacity-scheduler.xml. The following is an example:
  3. Set ACLs on the yarn-system queue such that the yarn-ats user can manage the ats-hbase service.
    To set ACLs on the queue, update values of the yarn.scheduler.capacity.root.<queue-path>.acl_submit_applications and yarn.scheduler.capacity.root.<queue-path>.acl_administer_queue in capacity-scheduler.xml.
  4. Disable preemption on the yarn-system queue.
    Set yarn.scheduler.capacity.root.<queue-path>.disable_preemption and yarn.scheduler.capacity.root.<queue-path>.intra-queue-preemption.disable_preemption in capacity-scheduler.xml to true.
  5. Assign the highest priority to the yarn-system queue to prevent inadvertent deletion or preemption.
    Set the value of yarn.scheduler.capacity.root.<queue-path>.priority in capacity-scheduler.xml.
  6. Restart YARN for all the configuration changes to take effect.
  7. Move ats-hbase from the default queue to the yarn-system queue.
    yarn application -changeQueue yarn-system -appId <app-id>

    Here, <app-id> is the ID of the ats-hbase service.