springboot中redis的事务的研究

redis的事务类似于队列操作，执行过程分为三步：

1. 开启事务
2. 入队操作
3. 执行事务

使用到的几个命令如下：

命令	说明
multi	开启一个事务
exec	事务提交
discard	事务回滚
watch	监听key(s)：当监听一个key(s)时，如果在本次事务提交之前，有其他命令修改了该key的值，那么本地事务就会失效
unwatch	取消监听key(s)

下面我们使用一个springboot的代码操作来说明这几个命令的含义：

package com.test.spring;import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.boot.SpringApplication;
import org.springframework.boot.autoconfigure.SpringBootApplication;
import org.springframework.dao.DataAccessException;
import org.springframework.data.redis.RedisSystemException;
import org.springframework.data.redis.connection.RedisConnection;
import org.springframework.data.redis.core.RedisCallback;
import org.springframework.data.redis.core.StringRedisTemplate;
import org.springframework.web.bind.annotation.RequestMapping;
import org.springframework.web.bind.annotation.RestController;
@SpringBootApplication
@RestController
public class TestSpringApplication {@Autowiredprivate StringRedisTemplate stringRedisTemplate;@RequestMapping("/exec1")public String test1(){stringRedisTemplate.execute(new RedisCallback<Boolean>() {@Overridepublic Boolean doInRedis(RedisConnection connection) throws DataAccessException {connection.multi();connection.commands().set("k1".getBytes(),"1".getBytes());connection.commands().set("k2".getBytes(),"2".getBytes());connection.exec();return true;}});return "k1="+stringRedisTemplate.opsForValue().get("k1")+",k2="+stringRedisTemplate.opsForValue().get("k2");}@RequestMapping("/exec2")public String exec2(){try{stringRedisTemplate.execute(new RedisCallback<Boolean>() {private int i=0;@Overridepublic Boolean doInRedis(RedisConnection connection) throws DataAccessException {connection.multi();connection.commands().set("k1".getBytes(),"11".getBytes());connection.commands().set("k2".getBytes(),"22".getBytes());if(i==0){throw new RedisSystemException("一个异常",new RuntimeException("1"));}connection.exec();return true;}});}catch (Exception e){e.printStackTrace();}return "k1="+stringRedisTemplate.opsForValue().get("k1")+",k2="+stringRedisTemplate.opsForValue().get("k2");}@RequestMapping("/discard")public String discard(){try{stringRedisTemplate.execute(new RedisCallback<Boolean>() {@Overridepublic Boolean doInRedis(RedisConnection connection) throws DataAccessException {connection.multi();connection.commands().set("k3".getBytes(),"3".getBytes());connection.commands().set("k4".getBytes(),"4".getBytes());connection.discard();connection.exec();return true;}});}catch (Exception e){e.printStackTrace();}return "k3="+stringRedisTemplate.opsForValue().get("k3")+",k4="+stringRedisTemplate.opsForValue().get("k4");}@RequestMapping("/watch1")public String watch1(){//开启一个线程new Thread(()->{stringRedisTemplate.execute(new RedisCallback<Boolean>() {@Overridepublic Boolean doInRedis(RedisConnection connection) throws DataAccessException {try {connection.watch("k5".getBytes());connection.multi();connection.commands().set("k5".getBytes(), "5".getBytes());//休眠5秒钟在提交事务try {Thread.sleep(5000);} catch (InterruptedException e) {}connection.exec();}catch (Exception e){e.printStackTrace();}return true;}});}).start();//开启一个线程new Thread(()->{stringRedisTemplate.execute(new RedisCallback<Boolean>() {@Overridepublic Boolean doInRedis(RedisConnection connection) throws DataAccessException {try {connection.multi();connection.commands().set("k5".getBytes(), "55".getBytes());connection.exec();}catch (Exception e){e.printStackTrace();}return true;}});}).start();return "success";}@RequestMapping("/watch2")public String watch2(){return "k5="+stringRedisTemplate.opsForValue().get("k5");}public static void main(String[] args) {SpringApplication.run(TestSpringApplication.class, args);}}

1. exec1
   正常流程，使用curl进行测试会返回：k1=1,k2=2
2. exec2
   我们模拟在事务队列中发送异常，会发现这段设值不成功，测试返回：k1=1,k2=2
3. discard
   事务回滚，我们先回滚，再提交，后台会抛出：ERR EXEC without MULTI错误，说明设值失败
4. watch1、watch2
   这里我们模拟两个线程，第一个线程先监听key，然后等待5秒钟，但是第二个线程直接去修改这个key，当5秒结束时，第一个线程再去提交事务时，会发现已经失效了，然后我们再通过watch2去查询值，测试返回：k5=55，说明线程1事务失效

最后再说明一下unwatch，每次操作exec()后，底层会自动调用unwatch，所以我们可以不用显示去调用unwatch命令。