@Async源码探究

1. @async源码探究

1.1. 上代码

@springbootapplication
@enableasync
public class springbootlearnapplication {

    public static void main(string[] args) {
        springapplication.run(springbootlearnapplication.class, args);
    }

}

@service
public class creatingthread08service {

    @async
    public void call(countdownlatch countdownlatch) {
        try {
            thread.sleep(1000);
        } catch (interruptedexception e) {
            e.printstacktrace();
        }
        system.out.println(thread.currentthread().getname() + " is running");
        countdownlatch.countdown();
        system.out.println(thread.currentthread().getname() + " is over");

    }
}

@runwith(springrunner.class)
@springboottest
public class springbootlearnapplicationtests {

    @autowired
    private creatingthread08service creatingthread08service;

    private int count = 4;

    private countdownlatch countdownlatch = new countdownlatch(count);

    @test
    public void contextloads() {

        stopwatch stopwatch = new stopwatch("async test");
        stopwatch.start();
        for (int i = 0; i < count; i++) {
            creatingthread08service.call(countdownlatch);
        }
        try {
            countdownlatch.await();
        } catch (interruptedexception e) {
            e.printstacktrace();
        }
        stopwatch.stop();
        system.out.println(stopwatch.prettyprint());
    }

}

结果

task-2 is running
task-2 is over
task-4 is running
task-4 is over
task-1 is running
task-1 is over
task-3 is running
task-3 is over
stopwatch 'async test': running time (millis) = 1018
-----------------------------------------
ms     %     task name
-----------------------------------------
01018  100%  

1.2. 提问

1.2.1. 加了该注解的方法，如果同时被调用n次，难道会创建n个线程？

• 通过debugger源码找到如下信息，为它默认线程池设置属性
  

• 我们把运行数量加大到count=20

• 结果

task-1 is running
task-1 is over
task-8 is running
task-8 is over
task-5 is running
task-5 is over
task-3 is running
task-3 is over
task-4 is running
task-4 is over
task-7 is running
task-7 is over
task-2 is running
task-2 is over
task-6 is running
task-6 is over
task-1 is running
task-1 is over
task-8 is running
task-3 is running
task-3 is over
task-5 is running
task-5 is over
task-8 is over
task-4 is running
task-6 is running
task-6 is over
task-7 is running
task-7 is over
task-2 is running
task-2 is over
task-4 is over
task-5 is running
task-1 is running
task-8 is running
task-8 is over
task-1 is over
task-3 is running
task-3 is over
task-5 is over
stopwatch 'async test': running time (millis) = 3021
-----------------------------------------
ms     %     task name
-----------------------------------------
03021  100%  

• 可以看出，如截图一致，它的运行核心执行线程数为8，且队列数很大，所以几乎不会再创建新的线程数，我特意sleep了1秒，20个任务，足够该线程池运行3遍，所以最大延迟3秒多

1.3. 那么核心线程数这个值能够变吗？

• 可以，通过application.properties中修改spring.task.execution.pool.core-size=20的值，比如我现在改成20，那么打印结果如下

task-1 is running
task-9 is running
task-12 is running
task-15 is running
task-17 is running
task-11 is running
task-11 is over
task-1 is over
task-10 is running
task-10 is over
task-8 is running
task-8 is over
task-6 is running
task-6 is over
task-4 is running
task-4 is over
task-7 is running
task-7 is over
task-5 is running
task-3 is running
task-3 is over
task-20 is running
task-20 is over
task-17 is over
task-19 is running
task-19 is over
task-2 is running
task-2 is over
task-18 is running
task-18 is over
task-15 is over
task-16 is running
task-16 is over
task-12 is over
task-14 is running
task-14 is over
task-13 is running
task-13 is over
task-9 is over
task-5 is over
stopwatch 'async test': running time (millis) = 1020
-----------------------------------------
ms     %     task name
-----------------------------------------
01020  100%  

• 所有线程只执行一遍，最大延迟就1秒多了

1.4. 总结

• 使用该注解，会创建默认核心线程为8的线程池，它的很多属性是可以通过配置设置的，如下

spring.task.execution.pool.core-size=10
spring.task.execution.thread-name-prefix=mytask-
spring.task.execution.pool.queue-capacity=10
spring.task.execution.pool.max-size=20
spring.task.execution.pool.keep-alive=60s
spring.task.execution.pool.allow-core-thread-timeout=true

• 根据业务需求设置合理的属性值，就相当于spring给你创建了线程池了