Cloudera Docs

July 26, 2024

This release of the Cloudera Data Warehouse (CDW) service on CDP Public Cloud has the following behavior changes:

Summary: Changes to the Azure Kubernetes Service (AKS) managed identities

Before this release: Hive inherited the Storage Blob Data Owner role from the Data Lake.

After this release: You must explicitly grant the Storage Blob Data Owner role to the AKS identity on your Data Lake storage account. This permission is required for activating an environment in CDW.

Summary: Overlay Networking replaces kubenet in Azure environments

Before this release: To avoid IP address exhaustion, you enable the kubenet networking feature when you activate an Azure environment to use with CDW.

After this release: To avoid IP address exhaustion, by default, CDW uses CNI overlay networking and replaces kubenet networking. You use the CDP CLI to set CIDR (Classless Inter-Domain Routing) ranges. For more information, see Overlay networking.

Summary: Selecting a resource template replaces configuring Database Catalog T-shirt sizes

Before this release: You could configure the following sizes of the Java Heap for your Database Catalog workload:
  • Small (default), 8Gb
  • Medium, 16Gb
  • Large, 24Gb
After this release: You can configure a CDW resource template, which fulfills all requirements for a Database Catalog for the type of Virtual Warehouse you have, for example a Hive or Impala Virtual Warehouse:
  • Default resources
  • Medium resources
  • Large resources
For more information, see Creating a Database Catalog.

Summary: Impala Autoscaling Dashboard replaces the older Impala Autoscaler Web UI

Before this release: The Impala Autoscaler Web UI provided the tools to monitor how Impala Virtual Warehouses autoscaled.

After this release: You can monitor Impala autoscaling in a warehouse that uses workload-aware autoscaling or the regular autoscaling using the new Impala Autoscaling Dashboard.

Summary: Apache Iceberg component version

Before this release: The supported Iceberg version was 1.3.0.

After this release: The supported Iceberg version is 1.4.3.

Summary: CDW_VERSIONED_DEPLOY entitlement

Before this release: To upgrade Azure Kubernetes Service for CDW to certain workload versions, you needed the CDW_VERSIONED_DEPLOY entitlement.

After this release: The status of Virtual Warehouse and Database Catalog workload version selections is now General availability (GA). The entitlement is no longer needed. For more information, see What's New in this release.

Summary: Uploading files to object store using Hue

Before this release: The Hue server performed the file upload operations. The maximum file size you could upload to S3 buckets or ADLS Gen2 containers was 2 GB. When multiple users performed major upload operations, it impacted the performance of Hue since all the backend calls were synchronous. The Hue server became slow as the upload blocked Hue services. You may have worked around it by disabling the upload from the main Hue server and having another setup with upload enabled and redirecting the end users wanting to upload files to the second instance.

After this release: The files are uploaded using an asynchronous task queue or job queue in Hue. This improves performance and allows you to upload files as large as 5 GB.

Summary: Change in the option name to enable Hue at the environment level

Before this release: The option to enable the shared Hue service (Hue at environment level) from the Environment Details page was called Enable Query Editor.

After this release: The option has been renamed to Enable Shared Hue Service.

Summary: Upgrading and rebuilding the shared Hue service

Before this release: In CDW version 1.8.1, you could not upgrade the Hue instances that are deployed at the environment level.

After this release: You can upgrade and rebuild the shared Hue service.