springboot @EnableAsync 多线程

在处理大数据或实时数据时，如果在主线程频繁创建大量对象，这些对象使用完后成为游离对象，不会立即被GC。当创建速度大于销毁速度时，可能导致内存持续上涨，最后内存溢出。
可以开启多线程来处理，线程内的对象会在执行结束后尽快的销毁，均分内存累加的负担，保证内存占用的稳定性。

springboot的多线程使用

1. 配置@EnableAsync

package com.cdgs.data.config;
import java.util.concurrent.Executor;
import org.springframework.aop.interceptor.AsyncUncaughtExceptionHandler;
import org.springframework.aop.interceptor.SimpleAsyncUncaughtExceptionHandler;
import org.springframework.context.annotation.ComponentScan;
import org.springframework.context.annotation.Configuration;
import org.springframework.scheduling.annotation.AsyncConfigurer;
import org.springframework.scheduling.annotation.EnableAsync;
import org.springframework.scheduling.concurrent.ThreadPoolTaskExecutor;

@Configuration
@EnableAsync(proxyTargetClass=true)//利用@EnableAsync注解开启异步任务支持
@ComponentScan("com.cdgs.data.service") //必须加此注解扫描包
public class CustomMultiThreadingConfig implements AsyncConfigurer{

    @Override
    public Executor getAsyncExecutor() {
        ThreadPoolTaskExecutor taskExecutor = new ThreadPoolTaskExecutor();
        taskExecutor.setCorePoolSize(10);
        taskExecutor.setMaxPoolSize(20);
        taskExecutor.setQueueCapacity(500);
      //当提交的任务个数大于QueueCapacity，就需要设置该参数，但spring提供的都不太满足业务场景，可以自定义一个，也可以注意不要超过QueueCapacity即可
      //taskExecutor.setRejectedExecutionHandler(new ThreadPoolExecutor.CallerRunsPolicy());        
        taskExecutor.setWaitForTasksToCompleteOnShutdown(true);
        taskExecutor.setAwaitTerminationSeconds(10);
        taskExecutor.setThreadNamePrefix("ES-IMOPRT-");
        taskExecutor.initialize();  
        return taskExecutor;
    }
    @Override
    public AsyncUncaughtExceptionHandler getAsyncUncaughtExceptionHandler() {
        return new SimpleAsyncUncaughtExceptionHandler();
    }
}

2. 配置@Async，修饰类时表示类里所有方法都是多线程异步执行

    @Async
    public Future jdbcToElasticsearch(Pageable pageable) {
        //通过实现ApplicationContextAware得到applicationContext，进而获取spring管理的bean
        BaseinfoRepository repository = (BaseinfoRepository)SpringBeanUtil.getBean(BaseinfoRepository.class);
        List list = new ArrayList<>() ;

        //实际查询数据库等业务代码...

        elasticsearchOperation.bulkSave(list);
        return new AsyncResult<>(list.size());
    }

3. 调用，注意：调用方法不能和异步方法在同一类里

    //oralce批量导入es
    @GetMapping("/baseinfo")
    public String importBaseinfo(@PageableDefault(size=1000)Pageable pageable) {
        long count = autoImportService.count();
        int size = pageable.getPageSize();
        long loops = count%size>0?count/size+1:count/size;
        ArrayList> futureList = new ArrayList<>();
        for (int i = 0; i < loops; i++) {
            //异步执行任务，返回参数使用Future封装接收
            Future future = autoImportService.jdbcToElasticsearch(new PageRequest(pageable.getPageNumber() + i, size));
            futureList.add(future);
        }
        importCount = checkTaskDone(futureList);
        System.out.println(importCount);
    }

    public static long checkTaskDone(ArrayList> futureList) {
        //判断异步调用的方法是否全都执行完了
        while(true) {
            int doneSize = 0;
            for ( Future future  : futureList) {
                //该异步方法是否执行完成
                if(future.isDone()) {
                    doneSize++;
                }
            }
            //如果异步方法全部执行完，跳出循环
            if(doneSize == futureList.size()) {
                break;
            }
            try {
                Thread.sleep(2000);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }//每隔2秒判断一次
        }
        
        long importCount = 0;
        for ( Future future  : futureList) {
            try {
                importCount += future.get();
            } catch (InterruptedException | ExecutionException e) {
                e.printStackTrace();
            }
        }
        return importCount;
    }

调用该方法，观察内存能够稳定在一定范围

QQ图片20190304090025.png