我有一个帮助程序例程，试图从S3中进行线程下载。很多时候(大约占请求的1％)，我会收到一条有关NoHttpResponseException的日志消息，过一会儿，当从SocketTimeoutException中读取时会导致S3ObjectInputStream。
我是在做错什么，还是仅仅是我的路由器/互联网？或者这是S3所期望的？我没有注意到其他地方的问题。
public void
fastRead(final String key, Path path) throws StorageException
{
final int pieceSize = 1<<20;
final int threadCount = 8;
try (FileChannel channel = (FileChannel) Files.newByteChannel( path, WRITE, CREATE, TRUNCATE_EXISTING ))
{
final long size = s3.getObjectMetadata(bucket, key).getContentLength();
final long pieceCount = (size - 1) / pieceSize + 1;
ThreadPool pool = new ThreadPool (threadCount);
final AtomicInteger progress = new AtomicInteger();
for(int i = 0; i < size; i += pieceSize)
{
final int start = i;
final long end = Math.min(i + pieceSize, size);
pool.submit(() ->
{
boolean retry;
do
{
retry = false;
try
{
GetObjectRequest request = new GetObjectRequest(bucket, key);
request.setRange(start, end - 1);
S3Object piece = s3.getObject(request);
ByteBuffer buffer = ByteBuffer.allocate ((int)(end - start));
try(InputStream stream = piece.getObjectContent())
{
IOUtils.readFully( stream, buffer.array() );
}
channel.write( buffer, start );
double percent = (double) progress.incrementAndGet() / pieceCount * 100.0;
System.err.printf("%.1f%%\n", percent);
}
catch(java.net.SocketTimeoutException | java.net.SocketException e)
{
System.err.println("Read timed out. Retrying...");
retry = true;
}
}
while (retry);
});
}
pool.<IOException>await();
}
catch(AmazonClientException | IOException | InterruptedException e)
{
throw new StorageException (e);
}
}
2014-05-28 08:49:58 INFO com.amazonaws.http.AmazonHttpClient executeHelper Unable to execute HTTP request: The target server failed to respond
org.apache.http.NoHttpResponseException: The target server failed to respond
at org.apache.http.impl.conn.DefaultHttpResponseParser.parseHead(DefaultHttpResponseParser.java:95)
at org.apache.http.impl.conn.DefaultHttpResponseParser.parseHead(DefaultHttpResponseParser.java:62)
at org.apache.http.impl.io.AbstractMessageParser.parse(AbstractMessageParser.java:254)
at org.apache.http.impl.AbstractHttpClientConnection.receiveResponseHeader(AbstractHttpClientConnection.java:289)
at org.apache.http.impl.conn.DefaultClientConnection.receiveResponseHeader(DefaultClientConnection.java:252)
at org.apache.http.impl.conn.ManagedClientConnectionImpl.receiveResponseHeader(ManagedClientConnectionImpl.java:191)
at org.apache.http.protocol.HttpRequestExecutor.doReceiveResponse(HttpRequestExecutor.java:300)
at com.amazonaws.http.protocol.SdkHttpRequestExecutor.doReceiveResponse(SdkHttpRequestExecutor.java:66)
at org.apache.http.protocol.HttpRequestExecutor.execute(HttpRequestExecutor.java:127)
at org.apache.http.impl.client.DefaultRequestDirector.tryExecute(DefaultRequestDirector.java:713)
at org.apache.http.impl.client.DefaultRequestDirector.execute(DefaultRequestDirector.java:518)
at org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:906)
at org.apache.http.impl.client.AbstractHttpClient.execute(AbstractHttpClient.java:805)
at com.amazonaws.http.AmazonHttpClient.executeHelper(AmazonHttpClient.java:385)
at com.amazonaws.http.AmazonHttpClient.execute(AmazonHttpClient.java:233)
at com.amazonaws.services.s3.AmazonS3Client.invoke(AmazonS3Client.java:3569)
at com.amazonaws.services.s3.AmazonS3Client.getObject(AmazonS3Client.java:1130)
at com.syncwords.files.S3Storage.lambda$fastRead$0(S3Storage.java:123)
at com.syncwords.files.S3Storage$$Lambda$3/1397088232.run(Unknown Source)
at net.almson.util.ThreadPool.lambda$submit$8(ThreadPool.java:61)
at net.almson.util.ThreadPool$$Lambda$4/1980698753.call(Unknown Source)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:744)
最佳答案
更新:为响应我在GitHub上创建的问题，对AWS开发工具包进行了更新。我不确定情况如何变化。该答案的第二部分(批评getObject)可能(希望如此)是错误的。
S3被设计为失败，并且它经常失败。
幸运的是，适用于Java的AWS开发工具包具有用于重试请求的内置工具。不幸的是，在下载S3对象时并没有涵盖SocketExceptions的情况(它们在上载和执行其他操作时可以完成的工作)。因此，类似于问题中的代码是必要的(请参见下文)。
当该机制按预期工作时，您仍然会在日志中看到消息。您可以选择通过从INFO过滤com.amazonaws.http.AmazonHttpClient日志事件来隐藏它们。 (AWS SDK使用Apache Commons Logging。)
根据您的网络连接和Amazon服务器的运行状况，重试机制可能会失败。如 lvlv 所指出的，配置相关参数的方法是通过ClientConfiguration。我建议更改的参数是重试次数，默认情况下为3。您可以尝试的其他方法是增加或减少连接和套接字超时(默认值为50s，不仅足够长，而且考虑到无论如何无论如何都会超时)，可能太长了，并使用TCP KeepAlive(默认值)离开)。
ClientConfiguration cc = new ClientConfiguration()
.withMaxErrorRetry (10)
.withConnectionTimeout (10_000)
.withSocketTimeout (10_000)
.withTcpKeepAlive (true);
AmazonS3 s3Client = new AmazonS3Client (credentials, cc);
甚至可以通过设置
RetryPolicy(同样在ClientConfiguration中)来覆盖重试机制。它最有趣的元素是RetryCondition，默认情况下:请参见SDKDefaultRetryCondition javadoc和source
SDK中隐藏在其他地方的半屁股重试工具
内置机制(在整个AWS开发工具包中使用的)无法处理的是读取S3对象数据。
如果您调用
AmazonS3.getObject (GetObjectRequest getObjectRequest, File destinationFile)，则AmazonS3Client使用其自己的重试机制。该机制位于ServiceUtils.retryableDownloadS3ObjectToFile(source)内部，该机制使用次优的硬连线重试行为(它将仅重试一次，而不会在SocketException上重试!)。 ServiceUtils中的所有代码似乎设计欠佳(issue)。我使用类似于以下代码:
public void
read(String key, Path path) throws StorageException
{
GetObjectRequest request = new GetObjectRequest (bucket, key);
for (int retries = 5; retries > 0; retries--)
try (S3Object s3Object = s3.getObject (request))
{
if (s3Object == null)
return; // occurs if we set GetObjectRequest constraints that aren't satisfied
try (OutputStream outputStream = Files.newOutputStream (path, WRITE, CREATE, TRUNCATE_EXISTING))
{
byte[] buffer = new byte [16_384];
int bytesRead;
while ((bytesRead = s3Object.getObjectContent().read (buffer)) > -1) {
outputStream.write (buffer, 0, bytesRead);
}
}
catch (SocketException | SocketTimeoutException e)
{
// We retry exceptions that happen during the actual download
// Errors that happen earlier are retried by AmazonHttpClient
try { Thread.sleep (1000); } catch (InterruptedException i) { throw new StorageException (i); }
log.log (Level.INFO, "Retrying...", e);
continue;
}
catch (IOException e)
{
// There must have been a filesystem problem
// We call `abort` to save bandwidth
s3Object.getObjectContent().abort();
throw new StorageException (e);
}
return; // Success
}
catch (AmazonClientException | IOException e)
{
// Either we couldn't connect to S3
// or AmazonHttpClient ran out of retries
// or s3Object.close() threw an exception
throw new StorageException (e);
}
throw new StorageException ("Ran out of retries.");
}