File copy results in stream being larger than file length

[eben-roux]

I have encountered an odd scenario where copying a file from SharePoint to an Azure Blob Container results in the file stream returning a different number of bytes to the length of the file (see BlobCopyException below):

private async Task CopyAsync(IFile file, BlockBlobClient blockBlobClient, int chunkSize, CancellationToken cancellationToken)
{
    var buffer = new byte[chunkSize];
    var blockCopiedBytes = 0;
    var blockIds = new List<string>();
    var copiedLength = 0L;

    var fileStream = await file.GetContentAsync(true);

    using (var blockStream = new MemoryStream())
    {
        int bytesRead;

        while ((bytesRead = await fileStream.ReadAsync(buffer, 0, buffer.Length, cancellationToken)) != 0)
        {
            if (blockCopiedBytes + bytesRead > MaximumBlockSize)
            {
                var blockId = Convert.ToBase64String(Guid.NewGuid().ToByteArray());
                await blockBlobClient.StageBlockAsync(blockId, new MemoryStream(blockStream.GetBuffer(), 0, blockCopiedBytes), cancellationToken: cancellationToken);
                blockIds.Add(blockId);

                blockCopiedBytes = 0;
                blockStream.SetLength(0);
            }

            await blockStream.WriteAsync(buffer, 0, bytesRead, cancellationToken);

            copiedLength += bytesRead;
            blockCopiedBytes += bytesRead;
        }

        if (blockCopiedBytes > 0)
        {
            var blockId = Convert.ToBase64String(Guid.NewGuid().ToByteArray());
            await blockBlobClient.StageBlockAsync(blockId, new MemoryStream(blockStream.GetBuffer(), 0, blockCopiedBytes), cancellationToken: cancellationToken);
            blockIds.Add(blockId);
        }
    }

    if (copiedLength != file.Length)
    {
        throw new BlobCopyException(message);
    }

    await blockBlobClient.CommitBlockListAsync(blockIds, cancellationToken: cancellationToken);
}

I was wondering if someone has come across this before.

It is also not consistently the same file. The file size also doesn't seem to be a factor as some are only around 50Kb. This is only currently happening in a customer's environment. I'm using the same copy mechanism for internal archiving where we've copied thousands of files, some quite large, without any issues.

I've also added logging to check the file.LockedByUser.UserPrincipalName and file.CheckedOutByUser.UserPrincipalName but I suspect that it wouldn't be that.

I'm at a bit of a loss, so any assistance would be appreciated.