Cloudera Docs

Creating a Cloudera Machine Learning data connection to Impala

Learn how to connect natively to data stored in Impala when using Cloudera Data Visualization in Cloudera Machine Learning.

Before you start using data modeling and visualization functions, you must connect to your data. The following steps show you how to create a new Cloudera Machine Learning data connection in Cloudera Data Visualization to an Impala data warehouse.

When you create a connection, you automatically get privileges to create and manage the associated datasets. You can also build dashboards and visuals within these datasets.

  • For more information on the Manage data connections privilege, see RBAC permissions.
  • For instructions on how to define privileges for a specific role, see Setting role privileges.
  • For instructions on how to assign the administrator role to a user, see Promoting a user to administrator.

If you are using a Cloudera Private Cloud Base cluster running Impala with Kerberos for authentication, make sure that Kerberos credentials are configured in Cloudera Machine Learning before creating a Cloudera Machine Learning data connection to the Impala data warehouse. This ensures seamless integration and authentication between Cloudera Data Visualization and the Impala cluster. If you add Kerberos credentials after launching the Cloudera Data Visualization app, you need to restart the app for the changes to take effect.

For more information on using Kerberos for authentication in Cloudera Machine Learning, see Hadoop Authentication for AI Workspaces.

  1. On the main navigation bar, click DATA.
    The DATA interface opens, displaying the Datasets tab.
  2. On the side menu bar, click NEW CONNECTION.

    The Create New Data Connection modal window appears.

  3. Choose Impala from the Connection type drop-down list and assign a name to your connection.

    In this example, the Impala connection is made through Knox. Knox always uses TLS encryption and port 443 is the default HTTPS port.

  4. Enter the hostname or IP address of the running coordinator.
    You can retrieve this information from the JDBC URL of the Impala DW.
  5. Add 443 in the Port # field.
  6. Enter your workload username and password as credentials.
  7. Click the Advanced tab to configure additional details.
    1. For HTTP connection mode, locate the Impala Endpoint for the Data Hub.
    2. Copy and paste it into the HTTP Path field.
    3. Set any additional details as required.
  8. Check the Parameters and Data tabs for more configuration options.
  9. Once you finish configuring the settings, click TEST to check the connection.
  10. Click CONNECT to establish the connection.
You have successfully set up a connection to a running Impala DW.