What's New

Major features and updates for the Cloudera Machine Learning data service.

August 30, 2022

Release notes and fixed issues for version 2.0.32.

New Features / Improvements

  • Private Cluster on Azure - This option is temporarily disabled due to stability issues. (CDPAM-3279)
  • Backup / Restore Jobs - Timeouts can now be customized in both the Backup Workspace and Restore Workspace UI.

Fixed Issues

  • Workspace Installation (DSE-22545) - Increased the workspace installation timeout limit to 1 hour from 30 min
  • CML Workspace (DSE-22638) - Fixed an issue where CML workspace creation using a machine user caused an error.

July 21, 2022

Release notes and fixed issues for version 2.0.32.

New Features / Improvements

  • Garbage collection for deleted projects - This feature allows you to trigger cleanup of deleted projects. A separate feature allows older orphaned projects to be marked for cleanup. For more information, see Project Garbage Collection.
  • Disable Runtimes - It is now possible to disable and enable runtimes. For more information, see Disabling Runtimes.
  • Monitoring for Applications - This feature allows you to monitor the technical health of deployed Applications, including statistics and visualizations of CPU and memory usage. For more information, see Monitoring Applications.
  • Custom polling endpoints for applications - This feature allows the application creator to define what application endpoint servers poll to detect if the application is running, that avoids problems some applications have with polling the root endpoint. For more information, see Application polling endpoint.
  • PBJ Workbench Runtimes (Tech Preview) now work with Sessions, Experiments, Jobs and Applications - This feature enables the classic workbench UI backed by the open-source Jupyter protocol. This architectural change improves consistency, stability, and ease of customization while eliminating the dependency on proprietary CML code. For more information, see PBJ Workbench in Preview Features.
  • Kubernetes - Kubernetes 1.22 is now supported for both AWS and Azure.

Fixed Issues

  • Job quotas (DSE-12664) - Fixed an issue where subsequent jobs in a job pipeline can fail if quota is enabled.
  • HDFS via Python (DSE-19775) - Fixed an issue where accessing HDFS via Python libraries that connect natively to HDFS, such as Tensorflow or PyArrow, may fail due to an error that the libhdfs.so file cannot be found.

May 31, 2022

Release notes and fixed issues for version 2.0.30.

New Features / Improvements

  • ML Discovery and Exploration, SQL and Visualization (GA) - This feature enables Data Scientists to understand their data using a SQL editor and drag-and-drop Visual Dashboards within CML. Users can start with their pre-configured Data Connections and create Datasets that they can rely on for model development. For more information, see ML Discovery and Exploration.
  • CML Endpoint Stability - This feature adds the ability for CML admin to define the prefix for the URL for the CML Workspace. This enables a new CML Workspace to be created and leverage the endpoint of a previously deleted CML Workspace. This ensures that deployed models and applications deployed in the new Workspace will have the same endpoint as the same models and applications deployed in the old Workspace. For more information, see CML Static Subdomain in Provisioning ML Workspaces.
  • Add/Delete GPU Nodes - This feature enables MLAdmins to reconfigure CML Workspaces by adding or removing GPU Worker groups for existing deployments.