Before the Databricks Unit Catalog's release, we used init scripts stored in DBFS to generate the pip.conf file during cluster startup, allowing each cluster its unique auth token. But with init scripts no longer available in the Unit Catalog's shared mode, an alternative approach is required.
Suppose we're running following job/task in a Azure Databricks workspace:
jobId: "1111"
jobRunId: "2222"
taskRunId: "3333"
jobName: "ths job name"
taskName: "first-task"
databricksWorkspaceUrl: https://adb-4444444444.123.azuredatabricks.net/
Run below command in a Databricks job (task precisely):
Most of Databricks management can be done from the GUI or CLI, but for Azure Service Principal, we can only manage it by the SCIM API. There's an open PR for adding support of SCIM API in Databricks CLI, but the lastest update is back to the beginning of 2021.
This post is to add some tips that not covered by the official API docs.