New Features in Cloudera Data Science Workbench 1.4.0

Major features and updates for Cloudera Data Science Workbench.

Models and Experiments
Cloudera Data Science Workbench 1.4 extends the machine learning platform experience from research to production. Now you can use Cloudera Data Science Workbench to build, train, and deploy models in a unified workflow.
  • Experiments - Train and compare versioned, reproducible models
  • Models - Deploy and manage models as REST APIs to serve predictions
External Authentication
LDAP/SAML users can now restrict access to Cloudera Data Science Workbench to specific LDAP/SAML groups. Additionally, you can now specify groups that should automatically be granted site administrator privileges when they log in to Cloudera Data Science Workbench. For details, see Configuring External Authentication with LDAP and SAML.

Cloudera Data Science Workbench now supports multiple identity provider signing certificates for SAML 2.0 authentication.

Cloudera Data Science Workbench now supports SAML 2.0 Errata 05 E43 for SAML 2.0 authentication.

Projects and Workbench
  • Site administrators can now disable individual built-in template projects by using a checkbox in the Project Templates table at Admin > Settings. Only enabled project templates will be displayed in the dropdown menu when creating a new project.

  • The default .gitignore file that is created with each new project has been updated to:

  • Added support for multiple Terminal windows within a single session.

  • Cloudera Data Science Workbench now supports DNS resolution of localhost to non-local IP address (not

  • Cloudera Data Science Workbench now appends the following default values to the NO_PROXY parameter if any of the following properties are configured: HTTP_PROXY, HTTPS_PROXY, or ALL_PROXY.
Installation Validation Checks
Improved validation checks run during the installation process. Cloudera Data Science Workbench now:
  • Verifies that the wildcard DNS subdomain has been configured.
  • Verifies that resolv.conf is not pointing to
  • Validates iptables chains to ensure there are no custom rules being set.
  • Throws a warning if you are using a self-signed TLS certificate, an expired certificate, or if the certificate is not valid for the wildcard domain used for Cloudera Data Science Workbench.
Command Line
Added a verbose option to the cdsw status command.
cdsw status [-v|--verbose]
Kubernetes has been upgraded to version 1.8.12.