Cloudera Docs

Cloudera AI Studios Known Issues and Limitations

This topic describes known issues and workarounds in AI Studios.

Agent Studios

Tool Execution Failure in Azure Workbenches (DSE-50025)

Agent Studio workflow tools fail to execute in Cloudera AI Azure Workbenches, displaying the bwrap: Failed to make / slave: Permission denied error message. This occurs because Agent Studio requires Linux namespace isolation and elevated privileges for secure sandboxing, which are restricted in Azure Workbench environments.

Workaround: To enable tool execution in these environments, you can enable the Insecure Tool Execution mode. This allows tools to run without sandbox isolation.

To enable Insecure Tool Execution Mode:
  1. Open your project in the Cloudera AI Workbench.
  2. Navigate to Project Settings > Advanced > Environment Variables
  3. Add a new environment variable:

    1. Name: ALLOW_AGENT_STUDIO_INSECURE_TOOL_EXECUTION

    2. Value: true

  4. Click Submit and restart the Agent Studio application for the changes to take effect.

Remove or set ALLOW_AGENT_STUDIO_INSECURE_TOOL_EXECUTION to false, and restart Agent Studio application, to disable Insecure Tool Execution Mode.