I have an ML dataset that consists of 2.7M images, ~1TB. For ML experiments, I uploaded these as blobs in an Azure storage account under "hot" access (which I thought would be appropriate for ML training). While not actively training, I would like this dataset to be in "Archive" mode to avoid accumulating "leaky faucet" charges.

The document on blob storage access tiers describes how it can be toggled on a per-blob basis, but AFAICT that is per-file (not even per-directory, which would be fine).

Is there a way to toggle the access tier of an entire storage account?

Absent that, what might be a good strategy? Perhaps creating a default-archive-access data_archive storage account and do a cloud-based copy to a temporary data_for_training_purposes hot-access storage account that I create and delete as needed?

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.