为什么 Thread.activeCount() 计数的线程数比我的代码创建的线程数多？答案

【问题标题】：Why does Thread.activeCount() count more threads than my code creates?为什么 Thread.activeCount() 计数的线程数比我的代码创建的线程数多？
【发布时间】：2021-10-23 11:26:09
【问题描述】：

我是 Java 开发的初学者，我尝试在下面的代码中使用 CountDownLatch 编写多线程进程。但是 ExecutorService 不能正常工作。

ExecutorService 线程编号不像代码中定义的那样工作。

import java.net.URI;
import java.net.http.HttpClient;
import java.net.http.HttpRequest;
import java.net.http.HttpResponse;
import java.net.http.HttpResponse.BodyHandlers;
import java.util.concurrent.CountDownLatch;
import java.util.concurrent.ExecutorService;
import java.util.concurrent.Executors;

public class HTTPrequester {
    
    private ExecutorService exs;
    private CountDownLatch latch;
    private int thread_num = 50;
    private String[] url_list;
    
    public static void main(String[] args) {
        HTTPrequester req = new HTTPrequester();
        req.action();
    }
    
    
    public void action() {
        url_list = new String[]{"https://google.com/","https://microsoft.com/","https://yahoo.com/","https://stackoverflow.com/","https://youtube.com/","https://google.com/","https://microsoft.com/","https://yahoo.com/","https://stackoverflow.com/","https://youtube.com/",
                "https://google.com/","https://microsoft.com/","https://yahoo.com/","https://stackoverflow.com/","https://youtube.com/","https://google.com/","https://microsoft.com/","https://yahoo.com/","https://stackoverflow.com/","https://youtube.com/"};
        exs = Executors.newFixedThreadPool(thread_num);
        latch = new CountDownLatch(url_list.length);
        
        for(String url: url_list) {
            exs.submit(new Runner(latch,url));
        }
        
    }
    public class Runner implements Runnable {
        
        private CountDownLatch latch;
        private String url;
        public Runner(CountDownLatch latch, String url) {
            this.latch = latch;
            this.url = url;
        }

        @Override
        public void run() {
            try {
                HttpClient client = HttpClient.newHttpClient();
                HttpRequest request = HttpRequest.newBuilder().uri(URI.create(url)).build();
                HttpResponse<String> response = client.send(request, BodyHandlers.ofString());
                System.out.println("Thread Count: "+Thread.activeCount()+" - Latch Count: "+latch.getCount());
            }catch(Exception e) {}
            latch.countDown();
        }
        
    }
    
}

控制台日志；

Thread Count: 81 - Latch Count: 20
Thread Count: 81 - Latch Count: 20
Thread Count: 81 - Latch Count: 20
Thread Count: 81 - Latch Count: 17
Thread Count: 81 - Latch Count: 16
Thread Count: 81 - Latch Count: 15
Thread Count: 81 - Latch Count: 14
Thread Count: 81 - Latch Count: 13
Thread Count: 81 - Latch Count: 12
Thread Count: 81 - Latch Count: 11
Thread Count: 81 - Latch Count: 10
Thread Count: 81 - Latch Count: 9
Thread Count: 81 - Latch Count: 8
Thread Count: 81 - Latch Count: 7
Thread Count: 81 - Latch Count: 6
Thread Count: 81 - Latch Count: 5
Thread Count: 81 - Latch Count: 4
Thread Count: 81 - Latch Count: 3
Thread Count: 81 - Latch Count: 2
Thread Count: 81 - Latch Count: 1

我将线程号定义为 50，但是当它工作时Thread.activeCount() 函数返回 81。为什么？

【问题讨论】：

查看所有线程的名称。 HttpClient 创建了许多自己的线程。
你不是唯一一个创建线程的人！ Java 本身和许多框架、库等也创建线程。
嗯，那正常吗？我该怎么办？
为什么你认为有什么事情要做，或者你需要做什么？
我正在更改您问题的标题，以反映正文的实际要求。

标签： java multithreading concurrency

【解决方案1】：

您正在使用 Thread.activeCount()，它返回（估计）调用者线程的 ThreadGroup 中的线程数。

这与查询有多少线程正在使用您的固定大小的线程池不同（您已经知道答案，因为您已经创建了它）。

正如其他人所说，您不是唯一一个创建线程的人，具体来说，您不是唯一一个在（我猜）主线程的 ThreadGroup 中创建线程的人。

您创建的执行程序将创建属于您实际实例化它的线程的 ThreadGroup 的线程，因此您将始终拥有比线程池大小更高的计数。

如果您想查看（出于教育和调试目的）jvm 中所有线程的完整图片以及它们正在做什么（即它们的堆栈跟踪），请尝试使用 jstack（它应该捆绑在您选择的 JDK 中） .

https://docs.oracle.com/en/java/javase/11/tools/jstack.html

【讨论】：