服务器之家:专注于服务器技术及软件下载分享
分类导航

PHP教程|ASP.NET教程|Java教程|ASP教程|编程技术|正则表达式|C/C++|IOS|C#|Swift|Android|VB|R语言|JavaScript|易语言|vb.net|

服务器之家 - 编程语言 - Java教程 - 5个并发处理技巧代码示例

5个并发处理技巧代码示例

2021-01-28 11:40Igor Sorokin Java教程

这篇文章主要介绍了5个并发处理技巧代码示例,具有一定参考价值,需要的朋友可以了解下。

【译者注】在本文中,作者总结出了5个关于处理并发性程序的技巧,并给出代码示例,让读者更好地理解和使用这5种方法。 以下为译文:

1.捕获InterruptedException错误

请检查下面的代码片段:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
public class Task implements Runnable {
    private final BlockingQueue queue = ...;
    @Override
     public void run() {
        while (!Thread.currentThread().isInterrupted()) {
            String result = getOrDefault(() -> queue.poll(1L, TimeUnit.MINUTES), "default");
            //do smth with the result
        }
    }
    T getOrDefault(Callable supplier, T defaultValue) {
        try {
            return supplier.call();
        }
        catch (Exception e) {
            logger.error("Got exception while retrieving value.", e);
            return defaultValue;
        }
    }
}

代码的问题是,在等待队列中的新元素时,是不可能终止线程的,因为中断的标志永远不会被恢复:

1.运行代码的线程被中断。
2.BlockingQueue # poll()方法抛出InterruptedException异常,并清除了中断的标志。
3.while中的循环条件 (!Thread.currentThread().isInterrupted())的判断是true,因为标记已被清除。

为了防止这种行为,当一个方法被显式抛出(通过声明抛出InterruptedException)或隐式抛出(通过声明/抛出一个原始异常)时,总是捕获InterruptedException异常,并恢复中断的标志。

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
T getOrDefault(Callable supplier, T defaultValue) {
    try {
        return supplier.call();
    }
    catch (InterruptedException e) {
        logger.error("Got interrupted while retrieving value.", e);
        Thread.currentThread().interrupt();
        return defaultValue;
    }
    catch (Exception e) {
        logger.error("Got exception while retrieving value.", e);
        return defaultValue;
    }
}

2.使用特定的执行程序来阻止操作

因为一个缓慢的操作而使整个服务器变得无响应,这通常不是开发人员想要的。不幸的是,对于RPC,响应时间通常是不可预测的。

假设服务器有100个工作线程,有一个端点,称为100 RPS。在内部,它发出一个RPC调用,通常需要10毫秒。在某个时间点,此RPC的响应时间变为2秒,在峰值期间服务器能够做的惟一的一件事就是等待这些调用,而其他端点则无法访问。

?
1
2
3
4
5
6
7
@GET
@Path("/genre/{name}")
@Produces(MediaType.APPLICATION_JSON)
public Response getGenre(@PathParam("name") String genreName) {
    Genre genre = potentiallyVerySlowSynchronousCall(genreName);
    return Response.ok(genre).build();
}

解决这个问题最简单的方法是提交代码,它将阻塞调用变成一个线程池:

?
1
2
3
4
5
6
7
8
9
10
11
@GET
@Path("/genre/{name}")
@Produces(MediaType.APPLICATION_JSON)
public void getGenre(@PathParam("name") String genreName, @Suspended AsyncResponse response) {
    response.setTimeout(1L, TimeUnit.SECONDS);
    executorService.submit(() -> {
        Genre genre = potentiallyVerySlowSynchronousCall(genreName);
        return response.resume(Response.ok(genre).build());
    }
    );
}

3.传MDC的值

MDC(Mapped Diagnostic Context)通常用于存储单个任务的特定值。例如,在web应用程序中,它可能为每个请求存储一个请求id和一个用户id,因此MDC查找与单个请求或整个用户活动相关的日志记录变得更加容易。

?
1
2017-08-27 14:38:30,893 INFO [server-thread-0] [requestId=060d8c7f, userId=2928ea66] c.g.s.web.Controller - Message.

可是如果代码的某些部分是在专用线程池中执行的,则线程(提交任务的线程)中MDC就不会被继续传值。在下面的示例中,第7行的日志中包含“requestId”,而第9行的日志则没有:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
@GET
@Path("/genre/{name}")
@Produces(MediaType.APPLICATION_JSON)
public void getGenre(@PathParam("name") String genreName, @Suspended AsyncResponse response) {
    try (MDC.MDCCloseable ignored = MDC.putCloseable("requestId", UUID.randomUUID().toString())) {
        String genreId = getGenreIdbyName(genreName);
        //Sync call
        logger.trace("Submitting task to find genre with id '{}'.", genreId);
        //'requestId' is logged
        executorService.submit(() -> {
            logger.trace("Starting task to find genre with id '{}'.", genreId);
            //'requestId' is not logged
            Response result = getGenre(genreId) //Async call
            .map(artist -> Response.ok(artist).build())
               .orElseGet(() -> Response.status(Response.Status.NOT_FOUND).build());
            response.resume(result);
        }
        );
    }
}

这可以通过MDC#getCopyOfContextMap()方法来解决:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
...
public void getGenre(@PathParam("name") String genreName, @Suspended AsyncResponse response) {
 try (MDC.MDCCloseable ignored = MDC.putCloseable("requestId", UUID.randomUUID().toString())) {
 ...
 logger.trace("Submitting task to find genre with id '{}'.", genreId); //'requestId' is logged
 withCopyingMdc(executorService, () -> {
  logger.trace("Starting task to find genre with id '{}'.", genreId); //'requestId' is logged
  ...
 });
 }
}
private void withCopyingMdc(ExecutorService executorService, Runnable function) {
 Map

4.更改线程名称

为了简化日志读取和线程转储,可以自定义线程的名称。这可以通过创建ExecutorService时用一个ThreadFactory来完成。在流行的实用程序库中有许多ThreadFactory接口的实现:

?
1
2
3
com.google.common.util.concurrent.ThreadFactoryBuilde+r in Guava.
org.springframework.scheduling.concurrent.CustomizableThreadFactory in Spring.
org.apache.commons.lang3.concurrent.BasicThreadFactory in Apache Commons Lang 3.
?
1
2
3
4
ThreadFactory threadFactory = new BasicThreadFactory.Builder()
 .namingPattern("computation-thread-%d")
 .build();
ExecutorService executorService = Executors.newFixedThreadPool(numberOfThreads, threadFactory);

尽管ForkJoinPool不使用ThreadFactory接口,但也支持对线程的重命名:

?
1
2
3
4
5
6
ForkJoinPool.ForkJoinWorkerThreadFactory forkJoinThreadFactory = pool -> {
 ForkJoinWorkerThread thread = ForkJoinPool.defaultForkJoinWorkerThreadFactory.newThread(pool);
 thread.setName("computation-thread-" + thread.getPoolIndex());
 return thread;
};
ForkJoinPool forkJoinPool = new ForkJoinPool(numberOfThreads, forkJoinThreadFactory, null, false);

将线程转储与默认命名进行比较:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
"pool-1-thread-3" #14 prio=5 os_prio=31 tid=0x00007fc06b19f000 nid=0x5703 runnable [0x0000700001ff9000]
 java.lang.Thread.State: RUNNABLE
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.TaskHandler.compute(TaskHandler.java:16)
...
"pool-2-thread-3" #15 prio=5 os_prio=31 tid=0x00007fc06aa10800 nid=0x5903 runnable [0x00007000020fc000]
 java.lang.Thread.State: RUNNABLE
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.HealthCheckCallback.recordFailure(HealthChecker.java:21)
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.HealthChecker.check(HealthChecker.java:9)
...
"pool-1-thread-2" #12 prio=5 os_prio=31 tid=0x00007fc06aa10000 nid=0x5303 runnable [0x0000700001df3000]
 java.lang.Thread.State: RUNNABLE
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.TaskHandler.compute(TaskHandler.java:16)
 ...

与自定义命名进行比较:

?
1
2
3
4
5
6
7
8
9
10
11
12
13
"task-handler-thread-1" #14 prio=5 os_prio=31 tid=0x00007fb49c9df000 nid=0x5703 runnable [0x000070000334a000]
 java.lang.Thread.State: RUNNABLE
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.TaskHandler.compute(TaskHandler.java:16)
...
"authentication-service-ping-thread-0" #15 prio=5 os_prio=31 tid=0x00007fb49c9de000 nid=0x5903 runnable [0x0000700003247000]
 java.lang.Thread.State: RUNNABLE
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.HealthCheckCallback.recordFailure(HealthChecker.java:21)
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.HealthChecker.check(HealthChecker.java:9)
...
"task-handler-thread-0" #12 prio=5 os_prio=31 tid=0x00007fb49b9b5000 nid=0x5303 runnable [0x0000700003144000]
 java.lang.Thread.State: RUNNABLE
at com.github.sorokinigor.article.tipsaboutconcurrency.setthreadsname.TaskHandler.compute(TaskHandler.java:16)
 ...

想象一下,可能会不止3个线程。

5.使用LongAdder计数器

在高竞争的情况下,会采用java.util.concurrent.atomic.LongAdder进行计数,而不会采用AtomicLong/AtomicInteger。

LongAdder可以跨越多个单元间仍保持值不变,但是如果需要的话,也可以增加它们的值,但与父类AtomicXX比较,这会导致更高的吞吐量,也会增加内存消耗。

?
1
2
3
4
LongAdder counter = new LongAdder();
counter.increment();
...
long currentValue = counter.sum();

总结

以上就是本文关于5个并发处理技巧代码示例的全部内容,希望对大家有所帮助。有什么问题可以随时留言,小编会及时回复大家的。感谢朋友们对本站的支持!

原文链接:http://geek.csdn.net/news/detail/235232

延伸 · 阅读

精彩推荐