原文地址:http://www.baeldung.com/spring-async

1. Overview

In this article we’ll explore the asynchronous execution support in Spring – and the @Async annotation.

Simply put – annotating a method of a bean with @Async will make it execute in a separate thread i.e. the caller will not wait for the completion of the called method.

2. Enable Async Support

Let’s start by enabling asynchronous processing with Java configuration – by simply adding the @EnableAsync to a configuration class:

1
2
3
@Configuration
@EnableAsync
public class SpringAsyncConfig { ... }

The enable annotation is enough, but as you’d expect, there are also a few simple options for configuration as well:

  • annotation – by default, @EnableAsync detects Spring’s @Async annotation and the EJB 3.1 javax.ejb.Asynchronous; this option can be used to detect other, user defined annotation types as well
  • mode – indicates the type of advice that should be used – JDK proxy-based or AspectJ weaving
  • proxyTargetClass – indicates the type of proxy that should be used – CGLIB or JDK; this attribute has effect only if the mode is set to AdviceMode.PROXY
  • order – sets the order in which AsyncAnnotationBeanPostProcessor should be applied; by default it runs last, just so that it can take into account all existing proxies

Asynchronous processing can also be enabled using XML configuration – by using the task namespace:

1
2
<task:executor id="myexecutor" pool-size="5"  />
<task:annotation-driven executor="myexecutor"/>

3. The @Async Annotation

First – let’s go over the rules – @Async has two limitations:

  • it must be applied to public methods only
  • self invocation – calling the async method from within the same class – won’t work

The reasons are simple – the method needs to be public so that it can be proxied. And self-invocation doesn’t work because it bypasses the proxy and calls the underlying method directly.

3.1. Methods with void Return Type

Following is the simple way to configure a method with void return type to run asynchronously:

1
2
3
4
5
@Async
public void asyncMethodWithVoidReturnType() {
    System.out.println("Execute method asynchronously. "
      + Thread.currentThread().getName());
}

3.2. Methods With Return Type

@Async can also be applied to a method with return type – by wrapping the actual return in a Future:

1
2
3
4
5
6
7
8
9
10
11
12
13
@Async
public Future<String> asyncMethodWithReturnType() {
    System.out.println("Execute method asynchronously - "
      + Thread.currentThread().getName());
    try {
        Thread.sleep(5000);
        return new AsyncResult<String>("hello world !!!!");
    } catch (InterruptedException e) {
        //
    }
 
    return null;
}

Spring also provides a AsyncResult class which implements Future. This can be used to track the result of asynchronous method execution.

Now, let’s invoke the above method and retrieve the result of the asynchronous process using the Future object.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
public void testAsyncAnnotationForMethodsWithReturnType()
    throws InterruptedException, ExecutionException {
    System.out.println("Invoking an asynchronous method. "
      + Thread.currentThread().getName());
    Future<String> future = asyncAnnotationExample.asyncMethodWithReturnType();
 
    while (true) {
        if (future.isDone()) {
            System.out.println("Result from asynchronous process - " + future.get());
            break;
        }
        System.out.println("Continue doing something else. ");
        Thread.sleep(1000);
    }
}

4. The Executor

By default Spring uses a SimpleAsyncTaskExecutor to actually run these methods asynchronously. The defaults can be overridden at two levels – at the application level or at the individual method level.

4.1. Override the Executor at the Method Level

The required executor needs to be declared in a configuration class:

1
2
3
4
5
6
7
8
9
@Configuration
@EnableAsync
public class SpringAsyncConfig {
     
    @Bean(name = "threadPoolTaskExecutor")
    public Executor threadPoolTaskExecutor() {
        return new ThreadPoolTaskExecutor();
    }
}

Then the executor name should be provided as an attribute in @Async:

1
2
3
4
5
@Async("threadPoolTaskExecutor")
public void asyncMethodWithConfiguredExecutor() {
    System.out.println("Execute method with configured executor - "
      + Thread.currentThread().getName());
}

4.2. Override the Executor at the Application Level

The configuration class should implement the AsyncConfigurer interface – which will mean that it has the implement the getAsyncExecutor() method. It’s here that we will return the executor for the entire application – this now becomes the default executor to run methods annotated with @Async:

1
2
3
4
5
6
7
8
9
10
@Configuration
@EnableAsync
public class SpringAsyncConfig implements AsyncConfigurer {
     
    @Override
    public Executor getAsyncExecutor() {
        return new ThreadPoolTaskExecutor();
    }
     
}

5. Exception Handling

When a method return type is a Future, exception handling is easy – Future.get()method will throw the exception.

But, if the return type is voidexceptions will not be propagated to the calling thread. Hence we need to add extra configurations to handle exceptions.

We’ll create a custom async exception handler by implementingAsyncUncaughtExceptionHandler interface. The handleUncaughtException()method is invoked when there are any uncaught asynchronous exceptions:

1
2
3
4
5
6
7
8
9
10
11
12
public class CustomAsyncExceptionHandler  implements AsyncUncaughtExceptionHandler {
 
    @Override
    public void handleUncaughtException(Throwable throwable, Method method, Object... obj) {
        System.out.println("Exception message - " + throwable.getMessage());
        System.out.println("Method name - " + method.getName());
        for (Object param : obj) {
            System.out.println("Parameter value - " + param);
        }
    }
     
}

In the previous section we looked at the AsyncConfigurer interface implemented by the configuration class. As part of that, we also need to override thegetAsyncUncaughtExceptionHandler() method to return our custom asynchronous exception handler:

1
2
3
4
@Override
public AsyncUncaughtExceptionHandler getAsyncUncaughtExceptionHandler() {
    return new CustomAsyncExceptionHandler();
}

6. Conclusion

In this tutorial we looked at running asynchronous code with Spring. We started with the very basic configuration and annotation to make it work but also looked at more advanced configs such as providing our own executor, or exception handling strategies.

@Async in Spring--转的更多相关文章

  1. SpringBoot系列:Spring Boot异步调用@Async

    在实际开发中,有时候为了及时处理请求和进行响应,我们可能会多任务同时执行,或者先处理主任务,也就是异步调用,异步调用的实现有很多,例如多线程.定时任务.消息队列等, 这一章节,我们就来讲讲@Async ...

  2. spring boot使用@Async异步注解

    1.java的大部分接口的方法都是串行执行的,但是有些业务场景是不需要同步返回结果的,可以把结果直接返回,具体业务异步执行,也有些业务接口是需要并行获取数据,最后把数据聚合在统一返回给前端. 通常我们 ...

  3. 161021、spring异步调用,完美解决!

    前言 项目中,用户抢单,下单需要向对方推送消息,但是加上推送就会造成抢单和下单性能降低,反应变慢,因为抢单下单动作跟推送部分是同步的,现在想改成异步推送. 在Java应用中,绝大多数情况下都是通过同步 ...

  4. 【Spring】6、注解大全

    一.@interface Java用  @interface Annotation{ } 定义一个注解 @Annotation,一个注解是一个类. 二.@Override,@Deprecated,@S ...

  5. spring 类注入失败,解决之道

    1.今天偶尔发现的问题,如果你在一个类上面用了注解@Async,spring的异步注解之后,发现如果别的类用@Autowired导入这个类时会失败! 解决办法:用了@Async无非是想方便的用异步操作 ...

  6. Java方式配置Spring MVC

    概述 使用Java方式配置Spring MVC,以及回顾一下Spring MVC的各种用法. Spring MVC简述 关于Spring MVC的介绍网上有很多,这里就不再赘述了,只是要说一下,Spr ...

  7. 注解@Async解决异步调用问题

    序言:Spring中@Async 根据Spring的文档说明,默认采用的是单线程的模式的.所以在Java应用中,绝大多数情况下都是通过同步的方式来实现交互处理的. 那么当多个任务的执行势必会相互影响. ...

  8. springmvc实现long-pulling技术

    背景介绍: 项目中有一个通讯模块,本来是用websocket全双工技术实现的,但IE10下面不支持websocket,而国内的360.2345浏 览器封装的所有是IE10下面的内核,考虑到站点在国内的 ...

  9. servlet3异步原理与实践

    一.什么是Servlet servlet 是基于 Java 的 Web 组件,由容器进行管理,来生成动态内容.像其他基于 Java 的组件技术一样,servlet 也是基于平台无关的 Java 类格式 ...

  10. springboot-async

    在项目中,当访问其他人的接口较慢或者做耗时任务时,不想程序一直卡在耗时任务上,想程序能够并行执行, 我们可以使用多线程来并行的处理任务,也可以使用spring提供的异步处理方式@Async. Spri ...

随机推荐

  1. java-如何用eclipse打包jar

    Eclipse通过导出的方式(右键单击项目,之后选择Export)打包java类文件生成jar包. 方法一:(在项目工程没有引用外部jar包时,直接导出) 选中工程---->右键,Export. ...

  2. POJ2369 Permutations(置换的周期)

    链接:http://poj.org/problem?id=2369 Permutations Time Limit: 1000MS   Memory Limit: 65536K Total Submi ...

  3. 使用Dotfuscator 进行.Net代码混淆 代码加密的方法

    混淆代码能在一定程度上放置代码被盗用,保护我们的知识产权 1.打开vs2012,选择工具-〉Dotfuscator Software Services 2.选择你需要混淆的DLL 文件,可以多选择 3 ...

  4. An internal error occurred during: "Launching PmallSearch on Tomcat 7.x". java.lang.NullPointerException

    如果出现了上述的错误按照如下的3个步骤解决:1.首先关闭MyEclipse工作空间.2.然后删除工作空间下的“/.metadata/.plugins/org.eclipse.core.runtime/ ...

  5. Fedora17安装MySQL及配置

    一.安装 以下操作均是在root用户下进行操作.首先,如果以前有安装过,这次卸载后安装的话,最好是先将/var/lib/mysql/里面的内容全部删掉(如果不重要的话),我就遇到过没有删除这些东西,结 ...

  6. 一个微软的DDD架构图

  7. Microsoft开源跨平台的序列化库——Bond

    上个月Microsoft开源了Bond,一个跨平台的模式化数据处理框架.Bond支持跨语言的序列化/反序列化,支持强大的泛型机制能够对数据进行有效地处理.该框架在Microsoft公司内部的高扩展服务 ...

  8. UI控件(UIScrollView)

    @implementation ViewController - (void)viewDidLoad { [super viewDidLoad]; //创建一个scrollview UIScrollV ...

  9. Unity3D游戏开发初探—2.初步了解3D模型基础

    一.什么是3D模型? 1.1 3D模型概述 简而言之,3D模型就是三维的.立体的模型,D是英文Dimensions的缩写. 3D模型也可以说是用3Ds MAX建造的立体模型,包括各种建筑.人物.植被. ...

  10. 高灵活度,高适用性,高性能,轻量级的 ORM 实现

    ORM(Object-Relational Mapping 对象关系映射),它的作用是在关系型数据库和业务实体对象之间作一个映射,目的是提供易于理解的模型化数据的方法. ORM虽然有诸多好处,但是在实 ...