What's New

Major features and updates plus fixed issues for Cloudera Data Science Workbench 1.10.4.

New features and improvements

This release introduced the following improvements.

  • CDSW now supports RHEL 8.7.
  • CDSW includes ML Runtimes 2023.05. (Note: PBJ Runtimes are not supported with CDSW.)
  • Support backup of CDSW metadata and images
  • Exposed CML LDAP Client Configs to customers for more control.

Fixed issues

This release fixed the following issues.

  • Python 2 dependent script control.sh is modified to support Python 3.

    Cloudera Bug: DSE-26056

  • Feeding a large JSON data to a model no longer results in crashing the model-proxy pod.

    Cloudera Bug: DSE-21556

  • Fixed Projects list filter issue. For non-admin users, the creator filter and search filter now display the correct result.

    Cloudera Bug: DSE-25817

  • Allow cloning of files larger than pod's RAM size of 4GB.

    Cloudera Bug: DSE-24708

  • The ability to install both git and git-lfs is now available during project creation.

    Cloudera Bug: DSE-26995

  • cdswctl now accepts self-signed certificates.

    Cloudera Bug: DSE-15102

  • Resolve incompatibility of TLDR Python packages with user-installed newer Python packages.

    Cloudera Bug: DSE-21786

  • Asynchronously launch model builds and experiment builds. The client or web UI will poll for the status of the build. This avoids network transport layer timeouts for the clients.

    Cloudera Bug: DSE-22357

  • Job runs now validate Runtimes properly.

    Cloudera Bug: DSE-24678

  • Engine version 15 has been replaced with Engine version 16 in CDSW 1.10.4.

    Cloudera Bug: DSE-26217

  • Added User Environment variables to job runs in APIv2.

    Cloudera Bug: DSE-20727

For a list of supported CDSW versions, see Support Lifecycle Policy