How do I troubleshoot the Azure Machine Learning error: 'Datastore (workspaceblobstore) cannot be found'?

Question

I'm encountering the error message "Datastore (workspaceblobstore) cannot be found" in Azure Machine Learning.

Question: What are the possible causes of this error, and how can I resolve it?

Job status: Failed

Error message at job:

Datastore (workspaceblobstore) can not be found. If its workspaceblobstore, please check if key vault has been changed after datastore was created; If the Datastore has secret registered, please make sure that RBAC action "metadata/secrets/read" is permitted.

Background information

As a cloud engineer, I'm preparing our AML environment for data scientist by stepping trough the following tutorial:

https://github.com/Azure/azureml-examples
/tutorials/azureml-getting-started/azureml-getting-started-studio.ipynb

Step 5:

# submit the command job
ml_client.create_or_update(job)

This step creates a job which is failing with the above error.

Troubleshooting:

My observations during troubleshooting this issue.

We use user assigned managed identity for the compute instance and the clusters as well. The roles should be configured correctly. I used the following guides to configure:

User-assigned managed identity - Link

I've added the role to managed identity shown below, but did not solve the issue:

Data scientist - Link <-- Possibly not required

I've tried to start the job manually to repeat the failed one with "Edit and submit" button. The "Training script" section shows the following error:

Get default datastore error: Cannot upload file if default workspace storage is not Azure blob type and default workspace blob doesn't exist

Based on the above error message I was able to proceed:

After verifying and creating the workspaceblobstore and selecting it as default the original code was able to Queue a job and it was Completed successfully. <-- Issue was resolved for us at this point

Create datastores - Link