“Databricks 0 running” could mean a few things. Here’s what it might refer to and how to troubleshoot it:

1. No Active Tasks on a Cluster:

• Explanation:  You might see “0 running” in the Spark UI on a Databricks cluster. This means no active Spark tasks are executing at that moment.
• Troubleshooting:
  • Check the Cluster Status: Ensure your cluster is running and not in an error state.
  • Check Job Progress: If you have a job running, check its progress. It might be between stages or waiting for resources.
  • Review Spark UI: Look at the completed stages in the Spark UI to see if any tasks were executed previously.
  • Examine Logs: Check the cluster logs for any errors or warnings.

2. Running Command Stuck:

• Explanation: A command you’ve submitted might get stuck in the “Running Command…” state in the notebook.
• Troubleshooting:
  • Cluster Resources: If your command is resource-intensive, it might take longer or wait for resources. Consider increasing the cluster size.
  • Network Issues: Check if any network problems prevent the command from completing.
  • Simple Test: Try running a simple command (e.g., print(“Hello”)) to see if it executes successfully.
  • Logs: Review the cluster and driver logs for any errors.

3. Job Finished but Still Running:

• Explanation: Sometimes, even after all tasks are completed, a Spark job might show as running if the SparkContext is still active.
• Troubleshooting:
  • Driver Logs: Look at the driver’s standard error (stderr) logs to see what it’s doing.
  • File Commit: The driver might still commit the work if you write output files.
  • Manually Stop: You should manually stop the SparkContext or restart the cluster.

Additional Tips:

• Community Resources: Check Databricks forums or Stack Overflow for similar issues.
• Databricks Support: If the problem persists, contact Databricks support for assistance.