CountDownLatch使用错误+未最终断开连接导致线程池资源耗尽

发布于:2024-04-24 ⋅ 阅读:(30) ⋅ 点赞:(0)

错误描述:

        我设置了CountDownLatch对线程的协作做出了一些限制,但是我发现运行一段时间以后便发现定时任务不运行了。

具体代码:

public void sendToCertainWeb() throws IOException, InterruptedException {
        List<String> urlList = scheduleplanMapper.getRandomUrlList();
        Thread.sleep(6000);
        CountDownLatch countDownLatch = new CountDownLatch(20);
        for (String s : urlList) {
            transportThreadPool.execute(()->{
                try {
                    URL url = new URL(s);
                    // 打开连接
                    HttpURLConnection connection = (HttpURLConnection) url.openConnection();

                    // 设置请求方法为GET
                    connection.setRequestMethod("GET");
                    connection.setConnectTimeout(100000);
                    connection.setReadTimeout(100000);

                    // 添加自定义的请求头信息
                    String agent = scheduleplanMapper.getRandomAgent();
                    connection.addRequestProperty("User-Agent", agent);
                    connection.addRequestProperty("Accept-Language", "en-US,en;q=0.9");

                    // 获取服务器返回的状态码
                    int responseCode = connection.getResponseCode();

                    if (responseCode == HttpURLConnection.HTTP_OK) {
                        // 读取服务器返回的数据
                        BufferedReader reader = new BufferedReader(new InputStreamReader(connection.getInputStream()));

                        String line;
                        StringBuilder response = new StringBuilder();

                        while ((line = reader.readLine()) != null) {
                            response.append(line);
                        }
                        reader.close();
                        log.info("Right Code: " + responseCode);
                    } else {
                        log.error("Error Code: " + responseCode);
                    }

                    // 关闭连接
                    connection.disconnect();
                    countDownLatch.countDown();
                }catch (Exception e){
                    log.error(JSON.toJSONString(e));

                }

            });
        }
        countDownLatch.await();
    }

报错以后定时任务不运行了 

错误排查:

 打印线程日志发现定时任务的线程在第86行代码停着不动了。

正常的线程日志应该是这样的。

查看第86行代码,发现这里并没有唤醒主线程 ,导致线程一直处于运行状态,无法继续下一个任务。

        错误的原因是countDownLatch.countDown()并没有放在finally块里因此发生了错误并不会走这块代码,导致线程没有countDown

错误修改:

把countDownLatch.countDown();放在finally代码块里保证一定会进行countDown这个动作

正确代码:

    public void sendToCertainWeb() throws IOException, InterruptedException {
        List<String> urlList = scheduleplanMapper.getRandomUrlList();
        Thread.sleep(6000);
        CountDownLatch countDownLatch = new CountDownLatch(20);
        for (String s : urlList) {
            transportThreadPool.execute(()->{
                try {
                    URL url = new URL(s);
                    // 打开连接
                    HttpURLConnection connection = (HttpURLConnection) url.openConnection();

                    // 设置请求方法为GET
                    connection.setRequestMethod("GET");
                    connection.setConnectTimeout(100000);
                    connection.setReadTimeout(100000);

                    // 添加自定义的请求头信息
                    String agent = scheduleplanMapper.getRandomAgent();
                    connection.addRequestProperty("User-Agent", agent);
                    connection.addRequestProperty("Accept-Language", "en-US,en;q=0.9");

                    // 获取服务器返回的状态码
                    int responseCode = connection.getResponseCode();

                    if (responseCode == HttpURLConnection.HTTP_OK) {
                        // 读取服务器返回的数据
                        BufferedReader reader = new BufferedReader(new InputStreamReader(connection.getInputStream()));

                        String line;
                        StringBuilder response = new StringBuilder();

                        while ((line = reader.readLine()) != null) {
                            response.append(line);
                        }
                        reader.close();
                        log.info("Right Code: " + responseCode);
                    } else {
                        log.error("Error Code: " + responseCode);
                    }

                    // 关闭连接
                    connection.disconnect();

                }catch (Exception e){
                    log.error(JSON.toJSONString(e));
                }finally {
                    countDownLatch.countDown();
                }

            });
        }
        countDownLatch.await();
    }

错误总结:

         我们一般认为线程处于blocked状态的时候线程才是处于阻塞状态,但是这个状态只是对于计算机来说的。对于我们来说,只要业务不执行了,线程就是处于阻塞状态的,因此任何状态下的线程对于业务来说都是阻塞的。 我这个项目是爬虫项目,会去爬取别人网站的数据,有些网站识别爬虫之后不仅会拒绝你访问,还会通过一直不给响应使得你的服务器线程占满,进而导致你的爬虫服务器崩溃。

参考文章: 

未设置超时时间导致线程池资源耗尽,排查过程-CSDN博客