Closed boynextdoor-cze closed 1 month ago
Thank you for your feedback. Tagging and routing to the team member best able to assist.
Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @Azure/azure-ml-sdk @azureml-github.
Hi @boynextdoor-cze - Thanks for opening an issue. We'll take a look at this asap!
assuming error in occurring while uploading the image file to storage account, can you please help with full exception trace? I think what you need a key less data store. https://github.com/Azure/azureml-examples/blob/main/sdk/python/resources/datastores/datastore.ipynb
Problem has been solved. I granted the app registration the reader & contributor access and then everything works well
I am creating a training environment in my Azure ML Workspace using docker, as the code below:
As Microsoft has disabled SAS access to storage account internally, when deploying the code above on Azure DevOps I got the following error:
azure.core.exceptions.HttpResponseError: Operation returned an invalid status 'Key based authentication is not permitted on this storage account.'
. I guess the reason is thatcreate_or_update
methods uses SAS URI for environment creation so that the problem seems to be inevitable.I've tried using
ManagedIdentityCredential()
but the error message becameManagedIdentityCredential.get_token failed: ManagedIdentityCredential authentication unavailable, no response from the IMDS endpoint.
Therefore, I wonder if there's any alternative to avoid using SAS, either via Python SDK or Azure CLI