Cloudera Docs

Pre-migration steps

There are steps to take before you begin a Sidecar migration.

Planning

To plan your migration, you need to consider a variety of items.

  • When you provision the new CDP deployment, you should consider the type and capacity of the hosts you deploy in your clusters. Ensure that hardware is sufficient in speed and capacity to handle your current and planned workloads.

  • Depending on the software running on the source cluster, not all services deployed in CDH clusters have exact matches in CDP. Consider how your services will map to updated versions of those services or to new services designed to replace the legacy software. Configurations and performance tuning may need to be adjusted.

  • Plan for any updates that may be required in your application code. Newer versions may have updated or deprecated APIs and replacement services could require you to modify or re-architect your applications to use the correct APIs.

Replicate configurations

When you create your new CDP Private Cloud Base cluster(s) you may need to copy some of your custom/tuning configuration from the source cluster to the new cluster. These can include properties like thresholds, memory limits, ports numbers, directory locations, log file locations, etc. These configurations exist for services, clusters, and hosts. Note that many CDP services have been updated from earlier versions and there may not be a one-to-one correspondence of configuration parameters from the old service to the new service. You will need to ensure that you understand the new set of configuration options in CDP and that your configurations are valid and appropriate for your workloads and environment in which the new cluster is deployed..

For migrations from CDH, ​​Cloudera Manager can help you identify configuration properties that have been modified from their default settings so you can better assess what, if any, configuration settings should be explicitly brought over to the new CDP cluster:

  1. Open the Cloudera Manager Admin Console on the source deployment.
  2. Click Configuration > Non-default values.
    A list of host and service properties and their modified values displays.
  3. Open the Cloudera Manager Admin Console on the destination deployment.
  4. Manually set the desired configuration values on services and/or hosts in the new cluster. For more information, see Modifying Configuration Properties Using Cloudera Manager.
  5. For services that have been updated or replaced, you will need to evaluate any functional or performance tuning that may be required in the new deployment. The CDP documentation has a great deal of information about setting up Cloudera Runtime services. Cloudera Professional Services can also assist you with tuning the new services.
    If you are migrating the Hive service, see Hive Sidecar Migration from CDH to CDP for additional guidance on configuring Hive in CDP.

Set up Kerberos

See Configuring Authentication in Cloudera Manager

Important: If you enable Apache Ranger on the cluster, you must configure Kerberos to enforce strong authentication of users. Running Apache Ranger on a non-Kerberized CDP environment is not supported.