Navigator to Atlas Sidecar migration from CDH to CDP

To migrate from Navigator to Apache Atlas, you can extract metadata from the source cluster and then transform and import the metadata on the destination cluster.

You can migrate your business metadata such as tags, custom properties (definitions and entity assignments), Managed metadata properties (definitions and entity assignments), original and updated entity names and descriptions.

You can also migrate technical metadata from Hive, Impala, Spark, and Referenced HDFS or S3.

Navigator audit data is not migrated to Atlas. You can, however, archive the Navigator audits on the original cluster, and Atlas will begin collecting audit data on the new cluster. See Transitioning Navigator audits.

Entity metadata from unreferenced S3 and HDFS entities, and metadata for Sqoop, Pig, MapReduce, Oozie, and YARN, Navigator policies, Navigator configuration settings are not transitioned to Atlas.

To migrate the technical and business Navigator metadata, you run a command-line tool on the destination= cluster, pointing it to the source cluster that extracts the metadata from Navigator and then run another tool to convert the extracted data for use in Atlas.

  1. If Navigator is configured with TLS/SSL, run the following commands on the destination cluster:
    export KEYSTORE_PATH=<keystore-path-on-source-cluster>; 
    export KEYSTORE_PASSWORD=<keystore-password-on-source-cluster>;
    export TRUSTSTORE_PATH=<truststore-path-on-source-cluster>; 
    export TRUSTSTORE_PASSWORD=<truststore-password-on-source-cluster>
  2. Extract ​​Cloudera Navigator data:
    1. Log in to the Cloudera Manager server host on the destination cluster.
    2. Ensure that the JAVA_HOME environment variable points to the JDK used by the host.
    3. Open the following file in a text editor:
    4. Run the the script using the following options:
      -n <URL of Navigator server on source cluster>:7187 
      -u <Navigator username> 
      -p <Navigator Password> 
      -c <cluster name> 
      -o <path to output file>  
  3. Transform the metadata:

    On the Cloudera Manager server, run the following command:
    -f <output file name used in>
    -o  ‘/tmp/nav/nav2atlas/[***clustername***]’ clusterName ‘[***clustername***]
  4. Start the Atlas service.

    Atlas will automatically begin to import the entities.

For more information, see Transitioning Navigator content to Atlas.