Scaling the HBase REST server in COD [Technical Preview]

Using Apache HBase REST API, you can scale up the HBase REST server for better connectivity to COD.

To use this feature in your CDP environment, you must have the COD_RESTWORKERS entitlement enabled. Contact Cloudera Support or your Cloudera account team if you do not have this entitlement.

Multiple Knox Gateway servers are required to scale up the HBase REST servers and support the increased load, so you must have multiple Gateway nodes. When you define multiple Gateway nodes, the Knox Gateway instances work in an HA mode and the load is balanced among the multiple Knox instances.

You need a minimum of two Gateway nodes to utilize this functionality. The required number of Gateway nodes can be specified using the --num-gateway-nodes option in the create-database command. Each Gateway node hosts an Apache Knox Gateway instance that provides access to the HBase REST server hosted on a REST worker node. The maximum number of Gateway nodes supported is 10.

You cannot modify the number of Gateway nodes after the database is created, so plan the number of requested Gateway nodes based on the estimated load.

You can specify the required number of REST worker nodes using the --num-restworker-nodes option in the create-database command. This optional parameter can only be defined when you specify the --num-gateway-nodes option. You can scale up the performance of the HBase REST servers by increasing the number of the REST worker nodes during the database creation using the cdp datahub scale-cluster command. The default number of REST worker nodes supported is 0.

  • You need to create a minimum of two Gateway nodes to enable this functionality.
  • You must have the ODAdmin rights to make changes to the COD database.
  • You must download and install the latest CDP CLI beta version. For more information, see Installing Beta CDP CLI.
  1. Launch the CDP CLI tool.
  2. Create the database with the desired number of Gateway nodes using the --num-gateway-nodes option.
    cdp opdb create-database --environment-name env_name --database-name database_name --num-gateway-nodes integer

    Optionally, you can specify the number of REST worker nodes using the --num-restworker-nodes option.

    cdp opdb create-database --environment-name env_name --database-name database_name --num-gateway-nodes integer --num-restworker-nodes integer

If you prefer to scale up or down the HBase REST server, use the following DataHub command.

cdp datahub scale-cluster --cluster-name cluster_name --instance-group-name instance_name --instance-group-desired-count integer

For example,

cdp datahub scale-cluster --cluster-name cod-test --instance-group-name restworker --instance-group-desired-count 4