Question

I am trying to download a file from Azure blob, but in 0.1% cases it just hangs at the blockBlob.DownloadToStream(...) for more than 10 mins. Is there a way I can timeout this function in 2 mins?

k!

Answer 1

您可以配置整个DownloadToStream操作的超时。例如，您可以将MaximumExecutionTime设置为120秒（这适用于整个DownloadToStream操作），ServerTimeout设置为30秒（这适用于每个REST调用），并指定RetryPolicy（{ {3}}或ExponentialRetry，这适用于每个REST调用），以便在请求超时或失败时启用重试。通过这样做，每个单独的LinearRetry REST调用将在服务器端限制为30秒（如果遇到服务器超时问题，重试策略将生效），并且整个DownloadToStream操作将在客户端限制为2分钟。

请参阅下面的代码示例。注意：Get Blob对象只需构造一次，然后您可以将它用于下面所有与blob相关的操作。

        var maxRetryCount = 3;
        CloudBlockBlob blockBlob = container.GetBlockBlobReference(blobName);
        var blobRequestOptions = new BlobRequestOptions
        {
            ServerTimeout = TimeSpan.FromSeconds(30),
            MaximumExecutionTime = TimeSpan.FromSeconds(120),
            RetryPolicy = new ExponentialRetry(TimeSpan.FromSeconds(3), maxRetryCount),
        };

        using (var memoryStream = new MemoryStream())
        {
            blockBlob.DownloadToStream(memoryStream, null, blobRequestOptions);
        }

Answer 2

谢谢！我还得到了以下代码，以便在超时后重试下载blob。

 CloudBlockBlob blockBlob = container.GetBlockBlobReference(blobName);
 var cts = new CancellationTokenSource((int)TimeSpan.FromSeconds(30 * (retryCount + 1)).TotalMilliseconds);

 using (var memoryStream = new System.IO.MemoryStream())
 {
    Task task = blockBlob.DownloadToStreamAsync(memoryStream, cts.Token);
    await task.ConfigureAwait(false);

    if (!task.IsCanceled && !task.IsFaulted)
    {
        return Encoding.UTF8.GetString(memoryStream.ToArray());                            
    }
    else if (task.IsCanceled)
    {
        retryCount++;
        Console.WriteLine("Task cancelled happened for blob: {0}. Increasing timeout to: {1} sec", blobName, retryCount * 30);
    }
}

Azure blob DownloadToStream takes too long

2 个答案: