First chunk of StorageStreamDownloader.chunks() is always 32MB big

[mth-cbc]

• Package: azure-storage-blob
• Package Version: 12.6.0
• Operating System: Ubuntu 18.04
• Python Version: 3.8.0

Describe the bug
When using the StorageStreamDownloaders chunk interator to download/iterator offer a blob the first chunk is always 32*1024*1024 bytes big, regardless of the setting for max_chunk_get_size.

To Reproduce
Steps to reproduce the behavior:

1. instantiate a BlobClient for a file bigger than 32 MB
2. get a StorageStreamDownloader for the blob: stream = blob_client.download_blob()
3. iterate over the blob using the chunks() function and print the chunk size:

for chunk in stream.chunks():
    print(len(chunk))

Expected behavior
Each chunk, apart from the last one, should have the size given by the max_chunk_get_size parameter, so by default 4*1024*1024 bytes.

[xiafu-msft]

Hi @mth-cbc

Thanks so much for reporting this issue, it's an SDK bug.
Currently you can use this workaround by setting max_single_get_size=your expected chunk size, then the first chunk size should be changed.

Let me know if you have any concern! Sorry about the inconvenience.

[mth-cbc]

Hi @xiafu-msft

Thanks for your answer, but I observe the same behaviour when specifing an explicit chunk size via max_chunk_get_size.

[xiafu-msft]

Sorry I had a typo there! I mean max_single_get_size, updated the previous comment. set max_chunk_get_size and max_single_get_size to the same value will give you chunks with equal size when you call chunks()

[mth-cbc]

Thanks for the workaround, now it works as expected.

[tasherif-msft]

This pr #17559 will resolve this issue in the future :)
I will go ahead and close this issue!