[azure-storage-blob] BlobClient upload_blob does not use headers parameter for multi-block uploads.

[jboelter]

• Package Name: azure-storage-blob
• Package Version: 12.4.0
• Operating System: WSL / Ubuntu 20.04.1 LTS / 4.19.104-microsoft-standard / Windows 10.0.19041.450
• Python Version: 3.8.2

Describe the bug
BlobClient upload_blob() does not inject the headers parameter if the upload takes the multi-block upload code path (single block upload works).

To Reproduce

The provided headers will not be injected into the request if the object size exceeds blob_settings.max_single_put_size (default 64 * 1024 * 1024). Starting from a SAS URL.

        object = io.BytesIO(bytearray(1024*1024*100))
        blob = BlobClient.from_blob_url(url)
        blob.upload_blob(
           object,
           headers={'x-custom-header':'value'}, # BUG in the Azure SDK
        )

Expected behavior

The headers parameter to upload_blob should work for both single PUT uploads and multi-block uploads.

Fix appears to be just passing headers to upload_data_chunks 1 2 and upload_substream_blocks 1 2 in upload_block_blob. Not sure where else this may show up.

Additional context

We use a custom outgoing reverse proxy that expects certain headers.

Workaround the issue with a raw_request_hook.

        def hook(args):
            args.http_request.headers.update({'x-custom-header':'value'})

        blob = BlobClient.from_blob_url(url)
        blob.upload_blob(
            object,
            raw_request_hook=hook,
        )

[ghost]

Thanks for the feedback! We are routing this to the appropriate team for follow-up. cc @xgithubtriage.

[xiafu-msft]

Hi @jboelter

Thanks for reaching out and finding the workaround to unblock yourself.
It looks like we are passing header value when commit the block list,

azure-sdk-for-python/sdk/storage/azure-storage-blob/azure/storage/blob/_upload_helpers.py

Line 150 in 900aa7d

headers=headers,

Do you need that customized header on each put block request?

[ghost]

Hi, we're sending this friendly reminder because we haven't heard back from you in a while. We need more information about this issue to help address it. Please be sure to give us your input within the next 7 days. If we don't hear back from you within 14 days of this comment the issue will be automatically closed. Thank you!

[jboelter]

Hi @jboelter

Thanks for reaching out and finding the workaround to unblock yourself.
It looks like we are passing header value when commit the block list,

azure-sdk-for-python/sdk/storage/azure-storage-blob/azure/storage/blob/_upload_helpers.py

Line 150 in 900aa7d

headers=headers,

Do you need that customized header on each put block request?

The headers should be on every outgoing request. As noted in the bug, the headers are absent from upload_data_chunks and upload_substream_blocks. Not if this may show up in other call paths.

[xiafu-msft]

Hi @jboelter

It looks like we intended to only add the customized header in the exact "write" operation (change blob content) which is upload blob or commit block list operation. So it's not a bug....
While I think it doesn't hurt if we add it to all internal requests! We will post a pr to add customized header to all sub requests soon!

[tasherif-msft]

Hi @jboelter just wanted to let you know that we have a PR up for this issue. I'll let you know once its merged.