Apache HttpClient4.5（三）

转载自：
http://blog.csdn.net/z69183787/article/details/78039601
http://blog.csdn.net/z69183787/article/details/77964021

HttpClient中的RequestConfig在使用时：

RequestConfig requestConfig = RequestConfig.custom()
    .setConnectionRequestTimeout(1000)      
    .setSocketTimeout(1000)
    .setConnectTimeout(1000)
    .build();

有三个超时：connectionRequestTimeout、connectTimeout、socketTimeout，它们各自含义：
connectTimeOut：指建立连接的超时时间；
connectionRequestTimeOut：指从连接池获取到连接的超时时间，如果是非连接池的话，该参数暂时没有发现有什么用处。
socketTimeOut：指客户端和服务进行数据交互的时间，是指两者之间如果两个数据包之间的时间大于该时间则认为超时，而不是整个交互的整体时间，比如如果设置1秒超时，如果每隔0.8秒传输一次数据，传输10次，总共8秒，这样是不超时的。而如果任意两个数据包之间的时间超过了1秒，则超时。

更深理解可以参照:http://www.baeldung.com/httpclient-connection-management

---

RetryHandler重试方式ServiceUnavailableRetryStrategy并且可以自定义retry时间的间隔：

public static CloseableHttpClient getHttpClient() {
        ServiceUnavailableRetryStrategy serviceUnavailableRetryStrategy = new ServiceUnavailableRetryStrategy() {
            /**
             * retry逻辑
             */
            @Override
            public boolean retryRequest(HttpResponse response, int executionCount, HttpContext context) {
                if (executionCount <= 3)
                    return true;
                else
                    return false;
            }

            /**
             * retry间隔时间
             */
            @Override
            public long getRetryInterval() {
                return 2000;
            }
        };
        PoolingHttpClientConnectionManager cm = new PoolingHttpClientConnectionManager();
        cm.setMaxTotal(MAX_TOTAL);
        cm.setDefaultMaxPerRoute(MAX_PERROUTE);
        CloseableHttpClient httpClient = HttpClients.custom().setRetryHandler(new DefaultHttpRequestRetryHandler())
                .setConnectionManager(cm).build();
        return httpClient;
    }

除此之外，还有StandardHttpRequestRetryHandler、DefaultHttpRequestRetryHandler，或则自定义，实现HttpRequestRetryHandler即可。

---

两个主机建立连接的过程是很复杂的一个过程，涉及到多个数据包的交换，并且也很耗时间。Http连接需要的三次握手开销很大，这一开销对于比较小的http消息来说更大。但是如果直接使用已经建立好的http连接，这样花费就比较小，吞吐率更大。
httpclient的连接池中只有一个httpclient的实例，可以看看CloseableHttpClient和PoolingHttpClientConnectionManager的源码,会发现httpclient实例通过execute执行get或post获取连接的时候，会通过实例关联的connectionManager的connect()来建立连接，这里的connectionManager是一个连接池的实现, 因此connect方法调用的是PoolingHttpClientConnectionManager重写的实现，这就是一个连接池建立连接的实现，具体看源码。
简单示例：

public class HttpClientUtil {
    static final int timeOut = 10 * 1000;
    private static CloseableHttpClient httpClient = null;
    private final static Object syncLock = new Object();

    private static void config(HttpRequestBase httpRequestBase) {
        // 设置Header等
        // httpRequestBase.setHeader("User-Agent", "Mozilla/5.0");
        // httpRequestBase
        // .setHeader("Accept",
        // "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8");
        // httpRequestBase.setHeader("Accept-Language",
        // "zh-CN,zh;q=0.8,en-US;q=0.5,en;q=0.3");// "en-US,en;q=0.5");
        // httpRequestBase.setHeader("Accept-Charset",
        // "ISO-8859-1,utf-8,gbk,gb2312;q=0.7,*;q=0.7");

        // 配置请求的超时设置
        RequestConfig requestConfig = RequestConfig.custom()
                .setConnectionRequestTimeout(timeOut)
                .setConnectTimeout(timeOut).setSocketTimeout(timeOut).build();
        httpRequestBase.setConfig(requestConfig);
    }

    public static CloseableHttpClient getHttpClient(String url) {
        String hostname = url.split("/")[2];
        int port = 80;
        if (hostname.contains(":")) {
            String[] arr = hostname.split(":");
            hostname = arr[0];
            port = Integer.parseInt(arr[1]);
        }
        if (httpClient == null) {
            synchronized (syncLock) {
                if (httpClient == null) {
                    httpClient = createHttpClient(200, 40, 100, hostname, port);
                }
            }
        }
        return httpClient;
    }

    public static CloseableHttpClient createHttpClient(int maxTotal,
            int maxPerRoute, int maxRoute, String hostname, int port) {
        ConnectionSocketFactory plainsf = PlainConnectionSocketFactory
                .getSocketFactory();
        LayeredConnectionSocketFactory sslsf = SSLConnectionSocketFactory
                .getSocketFactory();
        Registry registry = RegistryBuilder
                . create().register("http", plainsf)
                .register("https", sslsf).build();
        PoolingHttpClientConnectionManager cm = new PoolingHttpClientConnectionManager(
                registry);
        // 将最大连接数增加
        cm.setMaxTotal(maxTotal);
        // 将每个路由基础的连接增加
        cm.setDefaultMaxPerRoute(maxPerRoute);
        HttpHost httpHost = new HttpHost(hostname, port);
        // 将目标主机的最大连接数增加
        cm.setMaxPerRoute(new HttpRoute(httpHost), maxRoute);

        // 请求重试处理
        HttpRequestRetryHandler httpRequestRetryHandler = new HttpRequestRetryHandler() {
            public boolean retryRequest(IOException exception,
                    int executionCount, HttpContext context) {
                if (executionCount >= 5) {// 如果已经重试了5次，就放弃
                    return false;
                }
                if (exception instanceof NoHttpResponseException) {// 如果服务器丢掉了连接，那么就重试
                    return true;
                }
                if (exception instanceof SSLHandshakeException) {// 不要重试SSL握手异常
                    return false;
                }
                if (exception instanceof InterruptedIOException) {// 超时
                    return false;
                }
                if (exception instanceof UnknownHostException) {// 目标服务器不可达
                    return false;
                }
                if (exception instanceof ConnectTimeoutException) {// 连接被拒绝
                    return false;
                }
                if (exception instanceof SSLException) {// SSL握手异常
                    return false;
                }

                HttpClientContext clientContext = HttpClientContext
                        .adapt(context);
                HttpRequest request = clientContext.getRequest();
                // 如果请求是幂等的，就再次尝试
                if (!(request instanceof HttpEntityEnclosingRequest)) {
                    return true;
                }
                return false;
            }
        };

        CloseableHttpClient httpClient = HttpClients.custom()
                .setConnectionManager(cm)
                .setRetryHandler(httpRequestRetryHandler).build();

        return httpClient;
    }

    private static void setPostParams(HttpPost httpost,
            Map params) {
        List nvps = new ArrayList();
        Set keySet = params.keySet();
        for (String key : keySet) {
            nvps.add(new BasicNameValuePair(key, params.get(key).toString()));
        }
        try {
            httpost.setEntity(new UrlEncodedFormEntity(nvps, "UTF-8"));
        } catch (UnsupportedEncodingException e) {
            e.printStackTrace();
        }
    }

    public static String post(String url, Map params) throws IOException {
        HttpPost httppost = new HttpPost(url);
        config(httppost);
        setPostParams(httppost, params);
        CloseableHttpResponse response = null;
        try {
            response = getHttpClient(url).execute(httppost,
                    HttpClientContext.create());
            HttpEntity entity = response.getEntity();
            String result = EntityUtils.toString(entity, "utf-8");
            EntityUtils.consume(entity);
            return result;
        } catch (Exception e) {
//          e.printStackTrace();
            throw e;
        } finally {
            try {
                if (response != null)
                    response.close();
            } catch (IOException e) {
                e.printStackTrace();
            }
        }
    }

    public static String get(String url) {
        HttpGet httpget = new HttpGet(url);
        config(httpget);
        CloseableHttpResponse response = null;
        try {
            response = getHttpClient(url).execute(httpget,
                    HttpClientContext.create());
            HttpEntity entity = response.getEntity();
            String result = EntityUtils.toString(entity, "utf-8");
            EntityUtils.consume(entity);
            return result;
        } catch (IOException e) {
            e.printStackTrace();
        } finally {
            try {
                if (response != null)
                    response.close();
            } catch (IOException e) {
                e.printStackTrace();
            }
        }
        return null;
    }

    public static void main(String[] args) {
        // URL列表数组
        String[] urisToGet = {
                "http://blog.csdn.net/catoop/article/details/38849497",
                "http://blog.csdn.net/catoop/article/details/38849497" };

        long start = System.currentTimeMillis();
        try {
            int pagecount = urisToGet.length;
            ExecutorService executors = Executors.newFixedThreadPool(pagecount);
            CountDownLatch countDownLatch = new CountDownLatch(pagecount);
            for (int i = 0; i < pagecount; i++) {
                HttpGet httpget = new HttpGet(urisToGet[i]);
                config(httpget);
                // 启动线程抓取
                executors
                        .execute(new GetRunnable(urisToGet[i], countDownLatch));
            }
            countDownLatch.await();
            executors.shutdown();
        } catch (InterruptedException e) {
            e.printStackTrace();
        } finally {
            System.out.println("线程" + Thread.currentThread().getName() + ","
                    + System.currentTimeMillis() + ", 所有线程已完成，开始进入下一步！");
        }

        long end = System.currentTimeMillis();
        System.out.println("consume -> " + (end - start));
    }

    static class GetRunnable implements Runnable {
        private CountDownLatch countDownLatch;
        private String url;

        public GetRunnable(String url, CountDownLatch countDownLatch) {
            this.url = url;
            this.countDownLatch = countDownLatch;
        }

        @Override
        public void run() {
            try {
                System.out.println(HttpClientUtil.get(url));
            } finally {
                countDownLatch.countDown();
            }
        }
    }
}