Azure MLS: NoApplicationPermissons_Warning when trying to run a .ipyjp notebook

Question

Errors/Warnings received when opening notebooks. Worked prior to 7/22.

NoApplicationPermissons_Warning: You do not have permission to run cells or access external application. Check with your admin to change permissions.

Error: User does not have access to this compute instance. Please check if this compute instance is assigned to you and you have access to the workspace. Additionally, verify that you are on the correct network to access this compute instance.

<compute name> · Kernel error

Error present for user with standard MLS RBACs (screenshot attached) and for second user with contributor/owner role over the subscription.

Answer 1

I am not expert in this subject but when I check old threads it is always an issue of permission changes or network issues.

So start by verifying if that the compute instance you are trying to use is assigned to you. You can check this in the Azure Machine Learning workspace under the "Compute" tab.

Given that you have the standard MLS RBACs and a second user has the contributor/owner role, verify if these roles have not been modified recently.

az role assignment list --assignee <your_user_id>

az role assignment create --assignee <your_user_id> --role "Contributor" --scope /subscriptions/<subscription_id>/resourceGroups/<resource_group_name>/providers/Microsoft.MachineLearningServices/workspaces/<workspace_name>

Sometimes, simply restarting the kernel or the compute instance can resolve transient issues. You can do this through the Azure Machine Learning UI or using the following command:

az ml compute restart --name <compute_instance_name> --resource-group <resource_group_name> --workspace-name <workspace_name>

Answer 2

Hello,

Thanks for reaching out to us, we did not hear from you. I hope you have solved your issue. In case you need more information, I am adding a checklist for permission issues with your Azure Machine Learning (AML) workspace and compute resources. Some general items you may want to check as below, let us know more details so that we can help more.

1. Check User Permissions

Ensure that the user experiencing the issue has the appropriate permissions to run notebooks and access the compute instance:

Workspace Permissions: Verify that the user has the required role in the AML workspace. Even if the user has the Contributor or Owner role at the subscription level, they still need specific permissions within the AML workspace. Check if they have at least the "Machine Learning Contributor" role in the workspace.

Compute Instance Permissions: Ensure that the compute instance is properly assigned to the user. You can check this by navigating to the Azure portal:

1. Go to the Azure Machine Learning workspace.
   1. Select the "Compute" section.
      1. Click on the specific compute instance.
         1. Verify if the compute instance has proper access roles set up for the users.
2. Verify Network Access

Check if there are any network restrictions that might prevent access to the compute instance:

VNet and Firewall Settings: If your AML workspace and compute instance are behind a Virtual Network (VNet) or have firewall rules, ensure that the user’s network is properly configured to access these resources.

Public Network Access: If your compute instance is configured to only allow public network access, make sure that your IP address is not blocked.

3. Check Application Permissions

The NoApplicationPermissons_Warning suggests there might be issues with application permissions:

Service Principal or Managed Identity: Ensure that the service principal or managed identity used by your notebook or AML workspace has the appropriate permissions to access and execute operations on the compute instance.

Role Assignments: Confirm that the service principal or managed identity has been assigned the correct role for the AML workspace and compute instance.

4. Kernel and Compute Configuration

The kernel error may indicate issues with the compute instance configuration or the notebook kernel itself:

Restart Compute Instance: Sometimes, simply restarting the compute instance can resolve temporary issues. Go to the Azure portal, find the compute instance, and restart it.

Kernel Version: Ensure that the notebook kernel is compatible with the compute instance. Sometimes updates or changes in the compute environment might require kernel adjustments.

5. Recent Changes and Updates

Since the issue began occurring after a specific date, consider any recent changes:

• AML Workspace Updates: Check if there were any updates or changes made to the AML workspace settings or compute instance configurations around that date.
• Notebook Configuration: Verify if any recent changes were made to the notebook configuration that might affect its execution.

Please have a try and I hope it helps.

Regards,

Yutong

-Please kindly accept the answer if you feel helpful to support the community, thanks a lot.