How to automatically cancel the order if it is not paid within 30 minutes?

This article has been included in the Github warehouse, which contains core knowledge points such as computer foundation, Java core knowledge points, multithreading, JVM, common frameworks, distributed, microservices, design patterns, architecture, etc. Welcome to star~

Address: github.com/Tyson0314/J…

Table of contents

  • Understand the needs
  • Scenario 1: Database Polling
  • Solution 2: JDK's delay queue
  • Solution 3: Time Wheel Algorithm
  • Solution 4: redis cache
  • Scenario 5: Using message queues

Understand the needs

In development, we often encounter some requirements for delayed tasks.

For example

  • If the order is not paid within 30 minutes, it will be automatically canceled
  • Send a text message to the user 60 seconds after the order is generated

For the above tasks, we give a professional name to describe them, that is delayed tasks. Then there will be a question here, what is the difference between this delayed task and a scheduled task? There are a total of the following differences

Scheduled tasks have a clear trigger time, delayed tasks do not

Timing tasks have an execution cycle, while delayed tasks are executed within a period of time after an event is triggered, without an execution cycle

Timing tasks generally execute batch operations that are multiple tasks, while delayed tasks are generally a single task

Next, we take judging whether the order is timed out as an example to analyze the solution

Scenario 1: Database Polling

train of thought

This solution is usually used in small projects, that is, to scan the database regularly through a thread, judge whether there is an overtime order by the order time, and then perform operations such as update or delete

accomplish

Quartz can be used to achieve, a brief introduction

The maven project introduces a dependency as shown below

<dependency>
    <groupId>org.quartz-scheduler</groupId>
    <artifactId>quartz</artifactId>
    <version>2.2.2</version>
</dependency>
复制代码

Call the Demo class MyJob as shown below

package com.rjzheng.delay1;
​
import org.quartz.*;
import org.quartz.impl.StdSchedulerFactory;
​
public class MyJob implements Job {
​
    public void execute(JobExecutionContext context) throws JobExecutionException {
        System.out.println("要去数据库扫描啦。。。");
    }
​
    public static void main(String[] args) throws Exception {
        // 创建任务
        JobDetail jobDetail = JobBuilder.newJob(MyJob.class)
                .withIdentity("job1", "group1").build();
        // 创建触发器 每3秒钟执行一次
        Trigger trigger = TriggerBuilder
                .newTrigger()
                .withIdentity("trigger1", "group3")
                .withSchedule(
                        SimpleScheduleBuilder
                                .simpleSchedule()
                                .withIntervalInSeconds(3).
                                repeatForever())
                .build();
        Scheduler scheduler = new StdSchedulerFactory().getScheduler();
        // 将任务及其触发器放入调度器
        scheduler.scheduleJob(jobDetail, trigger);
        // 调度器开始调度任务
        scheduler.start();
    }
​
}
复制代码

Run the code, you can find that every 3 seconds, the output is as follows

要去数据库扫描啦。。。
复制代码

advantage

Simple and easy to implement, supports cluster operation

shortcoming

  • Consumes a lot of memory on the server
  • There is a delay, for example, if you scan every 3 minutes, the worst delay is 3 minutes
  • Suppose you have tens of millions of orders, scanning like this every few minutes, the database loss is huge

Solution 2: JDK's delay queue

train of thought

This solution is realized by using the DelayQueue that comes with JDK. This is an unbounded blocking queue. The queue can only get elements from it when the delay expires. Objects placed in the DelayQueue must implement the Delayed interface.

The DelayedQueue implementation workflow is shown in the figure below

其中 Poll():获取并移除队列的超时元素,没有则返回空

take():获取并移除队列的超时元素,如果没有则 wait 当前线程,直到有元素满足超时条件,返回结果。

实现

定义一个类 OrderDelay 实现 Delayed,代码如下

package com.rjzheng.delay2;
​
import java.util.concurrent.Delayed;
import java.util.concurrent.TimeUnit;
​
public class OrderDelay implements Delayed {
​
    private String orderId;
​
    private long timeout;
​
    OrderDelay(String orderId, long timeout) {
        this.orderId = orderId;
        this.timeout = timeout + System.nanoTime();
    }
​
    public int compareTo(Delayed other) {
        if (other == this) {
            return 0;
        }
        OrderDelay t = (OrderDelay) other;
        long d = (getDelay(TimeUnit.NANOSECONDS) - t.getDelay(TimeUnit.NANOSECONDS));
        return (d == 0) ? 0 : ((d < 0) ? -1 : 1);
    }
​
    // 返回距离你自定义的超时时间还有多少
    public long getDelay(TimeUnit unit) {
        return unit.convert(timeout - System.nanoTime(), TimeUnit.NANOSECONDS);
    }
​
    void print() {
        System.out.println(orderId + "编号的订单要删除啦。。。。");
    }
​
}
复制代码

运行的测试 Demo 为,我们设定延迟时间为 3 秒

package com.rjzheng.delay2;
​
import java.util.ArrayList;
import java.util.List;
import java.util.concurrent.DelayQueue;
import java.util.concurrent.TimeUnit;
​
public class DelayQueueDemo {
​
    public static void main(String[] args) {
        List<String> list = new ArrayList<String>();
        list.add("00000001");
        list.add("00000002");
        list.add("00000003");
        list.add("00000004");
        list.add("00000005");
​
        DelayQueue<OrderDelay> queue = newDelayQueue < OrderDelay > ();
        long start = System.currentTimeMillis();
        for (int i = 0; i < 5; i++) {
            //延迟三秒取出
            queue.put(new OrderDelay(list.get(i), TimeUnit.NANOSECONDS.convert(3, TimeUnit.SECONDS)));
            try {
                queue.take().print();
                System.out.println("After " + (System.currentTimeMillis() - start) + " MilliSeconds");
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
        }
    }
​
}
复制代码

输出如下

00000001编号的订单要删除啦。。。。
After 3003 MilliSeconds
00000002编号的订单要删除啦。。。。
After 6006 MilliSeconds
00000003编号的订单要删除啦。。。。
After 9006 MilliSeconds
00000004编号的订单要删除啦。。。。
After 12008 MilliSeconds
00000005编号的订单要删除啦。。。。
After 15009 MilliSeconds
复制代码

可以看到都是延迟 3 秒,订单被删除

优点

效率高,任务触发时间延迟低。

缺点

  • 服务器重启后,数据全部消失,怕宕机
  • 集群扩展相当麻烦
  • 因为内存条件限制的原因,比如下单未付款的订单数太多,那么很容易就出现 OOM 异常
  • 代码复杂度较高

方案 3:时间轮算法

思路

先上一张时间轮的图(这图到处都是啦)

时间轮算法可以类比于时钟,如上图箭头(指针)按某一个方向按固定频率轮动,每一次跳动称为一个 tick。这样可以看出定时轮由个 3 个重要的属性参数,ticksPerWheel(一轮的 tick 数),tickDuration(一个 tick 的持续时间)以及 timeUnit(时间单位),例如当 ticksPerWheel=60,tickDuration=1,timeUnit=秒,这就和现实中的始终的秒针走动完全类似了。

如果当前指针指在 1 上面,我有一个任务需要 4 秒以后执行,那么这个执行的线程回调或者消息将会被放在 5 上。那如果需要在 20 秒之后执行怎么办,由于这个环形结构槽数只到 8,如果要 20 秒,指针需要多转 2 圈。位置是在 2 圈之后的 5 上面(20 % 8 + 1)

实现

我们用 Netty 的 HashedWheelTimer 来实现

给 Pom 加上下面的依赖

<dependency>
    <groupId>io.netty</groupId>
    <artifactId>netty-all</artifactId>
    <version>4.1.24.Final</version>
</dependency>
复制代码

测试代码 HashedWheelTimerTest 如下所示

package com.rjzheng.delay3;
​
import io.netty.util.HashedWheelTimer;
import io.netty.util.Timeout;
import io.netty.util.Timer;
import io.netty.util.TimerTask;
​
import java.util.concurrent.TimeUnit;
​
public class HashedWheelTimerTest {
​
    static class MyTimerTask implements TimerTask {
​
        boolean flag;
​
        public MyTimerTask(boolean flag) {
            this.flag = flag;
        }
​
        public void run(Timeout timeout) throws Exception {
            System.out.println("要去数据库删除订单了。。。。");
            this.flag = false;
        }
    }
​
    public static void main(String[] argv) {
        MyTimerTask timerTask = new MyTimerTask(true);
        Timer timer = new HashedWheelTimer();
        timer.newTimeout(timerTask, 5, TimeUnit.SECONDS);
        int i = 1;
        while (timerTask.flag) {
            try {
                Thread.sleep(1000);
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
            System.out.println(i + "秒过去了");
            i++;
        }
    }
​
}
复制代码

输出如下

1秒过去了
2秒过去了
3秒过去了
4秒过去了
5秒过去了
要去数据库删除订单了。。。。
6秒过去了
复制代码

优点

效率高,任务触发时间延迟时间比 delayQueue 低,代码复杂度比 delayQueue 低。

缺点

  • 服务器重启后,数据全部消失,怕宕机
  • 集群扩展相当麻烦
  • 因为内存条件限制的原因,比如下单未付款的订单数太多,那么很容易就出现 OOM 异常

方案 4:redis 缓存

思路一

利用 redis 的 zset,zset 是一个有序集合,每一个元素(member)都关联了一个 score,通过 score 排序来取集合中的值

添加元素:ZADD key score member [score member …]

按顺序查询元素:ZRANGE key start stop [WITHSCORES]

查询元素 score:ZSCORE key member

移除元素:ZREM key member [member …]

测试如下

添加单个元素
redis> ZADD page_rank 10 google.com
(integer) 1
​
添加多个元素
redis> ZADD page_rank 9 baidu.com 8 bing.com
(integer) 2
​
redis> ZRANGE page_rank 0 -1 WITHSCORES
1) "bing.com"
2) "8"
3) "baidu.com"
4) "9"
5) "google.com"
6) "10"
​
查询元素的score值
redis> ZSCORE page_rank bing.com
"8"
​
移除单个元素
redis> ZREM page_rank google.com
(integer) 1
​
redis> ZRANGE page_rank 0 -1 WITHSCORES
1) "bing.com"
2) "8"
3) "baidu.com"
4) "9"
复制代码

那么如何实现呢?我们将订单超时时间戳与订单号分别设置为 score 和 member,系统扫描第一个元素判断是否超时,具体如下图所示

实现一

package com.rjzheng.delay4;
​
import redis.clients.jedis.Jedis;
import redis.clients.jedis.JedisPool;
import redis.clients.jedis.Tuple;
​
import java.util.Calendar;
import java.util.Set;
​
public class AppTest {
​
    private static final String ADDR = "127.0.0.1";
​
    private static final int PORT = 6379;
​
    private static JedisPool jedisPool = new JedisPool(ADDR, PORT);
​
    public static Jedis getJedis() {
        return jedisPool.getResource();
    }
​
    //生产者,生成5个订单放进去
    public void productionDelayMessage() {
        for (int i = 0; i < 5; i++) {
            //延迟3秒
            Calendar cal1 = Calendar.getInstance();
            cal1.add(Calendar.SECOND, 3);
            int second3later = (int) (cal1.getTimeInMillis() / 1000);
            AppTest.getJedis().zadd("OrderId", second3later, "OID0000001" + i);
            System.out.println(System.currentTimeMillis() + "ms:redis生成了一个订单任务:订单ID为" + "OID0000001" + i);
        }
    }
​
    //消费者,取订单
​
    public void consumerDelayMessage() {
        Jedis jedis = AppTest.getJedis();
        while (true) {
            Set<Tuple> items = jedis.zrangeWithScores("OrderId", 0, 1);
            if (items == null || items.isEmpty()) {
                System.out.println("当前没有等待的任务");
                try {
                    Thread.sleep(500);
                } catch (InterruptedException e) {
                    e.printStackTrace();
                }
                continue;
            }
            int score = (int) ((Tuple) items.toArray()[0]).getScore();
            Calendar cal = Calendar.getInstance();
            int nowSecond = (int) (cal.getTimeInMillis() / 1000);
            if (nowSecond >= score) {
                String orderId = ((Tuple) items.toArray()[0]).getElement();
                jedis.zrem("OrderId", orderId);
                System.out.println(System.currentTimeMillis() + "ms:redis消费了一个任务:消费的订单OrderId为" + orderId);
            }
        }
    }
​
    public static void main(String[] args) {
        AppTest appTest = new AppTest();
        appTest.productionDelayMessage();
        appTest.consumerDelayMessage();
    }
​
}
复制代码

此时对应输出如下

可以看到,几乎都是 3 秒之后,消费订单。

然而,这一版存在一个致命的硬伤,在高并发条件下,多消费者会取到同一个订单号,我们上测试代码 ThreadTest

package com.rjzheng.delay4;
​
import java.util.concurrent.CountDownLatch;
​
public class ThreadTest {
​
    private static final int threadNum = 10;
    private static CountDownLatch cdl = newCountDownLatch(threadNum);
​
    static class DelayMessage implements Runnable {
        public void run() {
            try {
                cdl.await();
            } catch (InterruptedException e) {
                e.printStackTrace();
            }
            AppTest appTest = new AppTest();
            appTest.consumerDelayMessage();
        }
    }
​
    public static void main(String[] args) {
        AppTest appTest = new AppTest();
        appTest.productionDelayMessage();
        for (int i = 0; i < threadNum; i++) {
            new Thread(new DelayMessage()).start();
            cdl.countDown();
        }
    }
​
}
复制代码

输出如下所示

显然,出现了多个线程消费同一个资源的情况。

解决方案

(1)用分布式锁,但是用分布式锁,性能下降了,该方案不细说。

(2)对 ZREM 的返回值进行判断,只有大于 0 的时候,才消费数据,于是将 consumerDelayMessage()方法里的

if(nowSecond >= score){
    String orderId = ((Tuple)items.toArray()[0]).getElement();
    jedis.zrem("OrderId", orderId);
    System.out.println(System.currentTimeMillis()+"ms:redis消费了一个任务:消费的订单OrderId为"+orderId);
}
复制代码

修改为

if (nowSecond >= score) {
    String orderId = ((Tuple) items.toArray()[0]).getElement();
    Long num = jedis.zrem("OrderId", orderId);
    if (num != null && num > 0) {
        System.out.println(System.currentTimeMillis() + "ms:redis消费了一个任务:消费的订单OrderId为" + orderId);
    }
}
复制代码

在这种修改后,重新运行 ThreadTest 类,发现输出正常了

思路二

该方案使用 redis 的 Keyspace Notifications,中文翻译就是键空间机制,就是利用该机制可以在 key 失效之后,提供一个回调,实际上是 redis 会给客户端发送一个消息。是需要 redis 版本 2.8 以上。

实现二

在 redis.conf 中,加入一条配置

notify-keyspace-events Ex

运行代码如下

package com.rjzheng.delay5;
​
import redis.clients.jedis.JedisPool;
import redis.clients.jedis.JedisPubSub;
​
public class RedisTest {
​
    private static final String ADDR = "127.0.0.1";
    private static final int PORT = 6379;
    private static JedisPool jedis = new JedisPool(ADDR, PORT);
    private static RedisSub sub = new RedisSub();
​
    public static void init() {
        new Thread(new Runnable() {
            public void run() {
                jedis.getResource().subscribe(sub, "__keyevent@0__:expired");
            }
        }).start();
    }
​
    public static void main(String[] args) throws InterruptedException {
        init();
        for (int i = 0; i < 10; i++) {
            String orderId = "OID000000" + i;
            jedis.getResource().setex(orderId, 3, orderId);
            System.out.println(System.currentTimeMillis() + "ms:" + orderId + "订单生成");
        }
    }
​
    static class RedisSub extends JedisPubSub {
        @Override
        public void onMessage(String channel, String message) {
            System.out.println(System.currentTimeMillis() + "ms:" + message + "订单取消");
​
        }
    }
}
复制代码

输出如下

可以明显看到 3 秒过后,订单取消了

ps:redis 的 pub/sub 机制存在一个硬伤,官网内容如下

原:Because Redis Pub/Sub is fire and forget currently there is no way to use this feature if your application demands reliable notification of events, that is, if your Pub/Sub client disconnects, and reconnects later, all the events delivered during the time the client was disconnected are lost.

翻: Redis 的发布/订阅目前是即发即弃(fire and forget)模式的,因此无法实现事件的可靠通知。也就是说,如果发布/订阅的客户端断链之后又重连,则在客户端断链期间的所有事件都丢失了。因此,方案二不是太推荐。当然,如果你对可靠性要求不高,可以使用。

优点

(1) 由于使用 Redis 作为消息通道,消息都存储在 Redis 中。如果发送程序或者任务处理程序挂了,重启之后,还有重新处理数据的可能性。

(2) 做集群扩展相当方便

(3) 时间准确度高

缺点

需要额外进行 redis 维护

方案 5:使用消息队列

思路

我们可以采用 rabbitMQ 的延时队列。RabbitMQ 具有以下两个特性,可以实现延迟队列

RabbitMQ 可以针对 Queue 和 Message 设置 x-message-tt,来控制消息的生存时间,如果超时,则消息变为 dead letter

lRabbitMQ 的 Queue 可以配置 x-dead-letter-exchange 和 x-dead-letter-routing-key(可选)两个参数,用来控制队列内出现了 deadletter,则按照这两个参数重新路由。结合以上两个特性,就可以模拟出延迟消息的功能,具体的,我改天再写一篇文章,这里再讲下去,篇幅太长。

优点

Efficient, you can use the distributed nature of rabbitmq to easily expand horizontally, and message support persistence increases reliability.

shortcoming

The ease of use itself depends on the operation and maintenance of rabbitMq. Because rabbitMq is referenced, the complexity and cost become higher.

Guess you like

Origin juejin.im/post/7181297729979547705