an error message when I run Azure machine learning automated ML

Question

I follow the the instruction of AZURE Machine learning automated ML, My job always fails showing the error message:

I believed that the error may be related to data asset:

my data is located in Azure blob storage with name "workspaceartiftactstore",

This data is uploaded through container "Azureml"

I can see my file in "workspaceartiftactstore". Why the container name change completely? ( my first confusion). I chose "dat_final_05.10.2023.csv" for my data assets. I can see the data correctly from its preview.

However, I submit my job. It always get the above error message. It is very confusing for me at this time.

Answer 1

Hi Yu Chai.

It seems like there might be a small mix-up with the container references in your Azure Machine Learning job setup. Why not take a moment to double-check the container names listed in your Azure portal? It's quite easy to overlook details when dealing with similar names.

Would you mind verifying that the container azureml-blobstore-ed619ea0-85f4-45e8-88d9-672ac6a082dc is indeed present and correctly spelled in your Azure Blob Storage? Sometimes, the simplest solution is the correct one.

Also, consider taking a peek at your job configuration settings. There's a chance that a file or a parameter might be pointing to a different container than intended. It's a good practice to ensure that all references align with the container where your dat_final_05.10.2023.csv file resides.

Permissions can be quite the tricksters as well. It wouldn't hurt to revisit the IAM settings to confirm that your Azure ML service is still on friendly terms with the Blob Storage container.

And here's a thought – if you've registered a datastore in Azure ML with a unique name, it could be that the name being used in your job configurations needs a quick update to match. Datastore names can sometimes differ from the actual storage container names, so it's worth a check.

Lastly, if you're employing connection strings or keys, a routine verification that they're current could save the day. These credentials have a habit of changing when we least expect them to.

I hope this helps?

Answer 2

Hello, I still feel confused. I specifically create a Azure blob storage container in my Azure storage account, "yucai_ml".

In the data asset, I see the data is created.

However, the data format can NOT be recognized correctly in Azure machine learning. If the same data I put it in "azureml" container, Azure machine learning can recognize the data correctly.

---

e

The data header can NOT be shown correctly in data preview as you see "Target column" does not show correctly.

If I put the data in container "azureml", this is the correct format Azure machine learning recognized.

Answer 3

I feel confused mostly. I explicitly choose a data from Azure data storage, and in its preview, I can see this dataset. It means the container and data exists. Since data asset only create a reference, why does it show error message "Error: The container azureml-blobstore-ed619ea0-85f4-45e8-88d9-672ac6a082dc does not exist." If it is the permissions issue, how can I check it? But I can see the data in the preview. It indicates that the permission is allowed, am I correct?

This is what I see my access in IAM. It shows "Has full access to all resources in the subscription"

Answer 4

Hello I am trying to assign my storage account to my Azure machine learning.

Then I click "add role assignment"

I can not find machine learning in its list. It is very confusing.

Answer 5

For this part "Permissions can be quite the tricksters as well. It wouldn't hurt to revisit the IAM settings to confirm that your Azure ML service is still on friendly terms with the Blob Storage container.", could you provide more details how to check it?