AzureMLCompute Job Failed

Question

I am having this error when I was following the notebook here:

https://learn.microsoft.com/en-us/learn/modules/use-automated-machine-learning/

AzureMLCompute job failed. BFSMountError: unable to mount blob fuse file system Info: Could not mount Azure Blob Container azureml-blobstore-29.......xx12 at workspaceblobstore: Unauthorized. Cannot access the storage account with the given account key. Please verify that the account key is valid. Info: Job environment preparation failed on 10.0.0.4.

{
"message": "AzureMLCompute job failed.\nBFSMountError: unable to mount blob fuse file system\n\tInfo: Could not mount Azure Blob Container azureml-blobstore-29.......xx12 at workspaceblobstore: Unauthorized. Cannot access the storage account with the given account key. Please verify that the account key is valid.\n\tInfo: Job environment preparation failed on 10.0.0.4."
}

It says 'cannot access the storage account' however I am able to run a Jupiter notebook within this compute and access the storage account without any problem. I guess 'account key' is referring to the storage account key. If so, how can I provide it during deployment?

Note:

1. I am having a similar error when I try to deploy a model to ACI.
2. The AML workspace and so the storages are created in a VNET

Answer 1

Hi.

Thanks for reaching out to us and sorry for your experience. For this error, the root cause might be the Updated Storage keys were not synced correctly, because of which the ML service was still using the old storage keys.

I will suggest you to do:

Run "az ml workspace sync-keys -w myworkspace -g myresourcegroup" to sync up the key again

Please have a try to see if this can be solved.

Regards,
Yutong

Answer 2

It's not supported using the Azure Machine Learning designer or automated machine learning (from the studio) with resources inside a virtual network.” / https://learn.microsoft.com/en-us/azure/machine-learning/how-to-enable-virtual-network

In near future we will support private link for ML in general in near future.