The error occurs because the automation script explicitly sets WORKSPACE_GROUPS="<ALL>" and DATABASES="<ALL>", which the UCX installer interprets literally as a schema called "ALL"—instead of using the special meaning that the manual prompt does when you just press Enter or accept the default. When you install manually and accept the default <ALL>, the CLI substitutes this internally to mean "all schemas" and "all workspace groups," but the script input is treated as a specific schema name, which does not exist—hence the "SCHEMA_NOT_FOUND" error.
Correct Way to Use <ALL> in Automation
-
Do not set DATABASES="<ALL>" directly in your automation script.
-
Leave the environment variable unset, or omit the parameter entirely in your automation, so that the CLI falls back to its default behavior and interprets <ALL> correctly.
-
If your script must provide a value, try passing an empty string ("") or not setting the DATABASES and WORKSPACE_GROUPS environment variables at all.
Recommended Fix
-
Remove WORKSPACE_GROUPS and DATABASES variable assignments from your script when you want all databases/groups to be processed.
-
Alternatively, update your script to match exactly how the manual process works—let the installer prompt for these fields, or investigate if simply supplying an empty value ("") works.
Additional Notes
-
This behavior is a known issue in Databricks Labs UCX installations and is noted in troubleshooting forums.
-
Always test automation scripts with parameter variations to determine whether a default prompt or empty string substitutes for <ALL> correctly.
If you continue to encounter issues, consult the UCX troubleshooting guide or Databricks community discussions for more advanced script configuration examples
