Managing Engines

This topic describes how to manage engines and configure engine environments to meet your project requirements.

Site administrators and project administrators are responsible for making sure that all projects on the deployment have access to the engines they need. Site admins can create engine profiles, determine the default engine version to be used across the deployment, and white-list any custom engines that teams require. As a site administrator, you can also customize engine environments by setting global environmental variables and configuring any files/folders that need to be mounted into project environments on run time.

By default, Cloudera Machine Learning ships a base engine image that includes kernels for Python, R, and Scala, along with some additional libraries ( see Configuring Cloudera Machine Learning Engines for more information) that can be used to run common data analytics operations. Occasionally, new engine versions are released and shipped with Cloudera Machine Learning releases.

Engine images are available in the Site Administrator panel at Admin > Engines, under the Engine Images section. As a site administrator, you can select which engine version is used by default for new projects. Furthermore, project administrators can explicitly select which engine image should be used as the default image for a project. To do so, go to the project's Overview page and click Settings on the left navigation bar.

If a user publishes a new custom Docker image, site administrators are responsible for white-listing such images for use across the deployment. For more information on creating and managing custom Docker images, see Configuring the Engine Environment.