Cloudera Data Warehouse fails to start after NameNode migration

When you perform an HDFS NameNode migration in Cloudera Base on premises, you may encounter issues with your Cloudera Data Warehouse instance for which you are required to deactivate and reactivate the Cloudera Base on premises environment, and then recreate your Hive and Impala Virtual Warehouses. Learn how you can mitigate these issues without having to deactivate your environment.

Enabled quota management

If you have enabled quota management in Cloudera Data Warehouse, you may encounter issues with Hue Pod scheduling during creating, refreshing, rebuilding, or modifying the configuration of Hive or Impala Virtual Warehouses. Perform the following steps to address these issues:

Click on the Resource Templates menu on the left navigation pane.
Select the Default Resources under the HIVE and IMPALA tab and then click Actions > Make a Copy.
Provide a name of the resource template and add resources (increase CPUs and Memory) for any of the components, such as Query Coordinator or Impala Autoscaler, and then click Apply Changes.
tip
Select the Set As Default option to make this resource template the default template for all new Cloudera Data Warehouse entities; in this case all new Hive or Impala Virtual Warehouses.
Refresh the page on your browser to view the newly created resource template. You can now create your Virtual Warehousing using this newly created resource template.
Once the statefulset/deployment of the modified component is present, reset its resources back to its original values. By doing this, you are freeing up enough resources for the Hue front end to get scheduled.

Enabled Active-Active High Availability (HA)

If you have enabled Active-Active HA for your Impala Virtual Warehouse, then you may notice pod start failures in Catalogd pods. To address this issue, you must uncheck the Enable Impala Catalog Server HA option from the Sizing And Scaling tab of the Impala Virtual Warehouse details page.