分布式锁场景
在分布式环境下多个操作需要以原子的方式执行
首先启一个springboot项目,再引入redis依赖包:
<!-- https://mvnrepository.com/artifa ... -starter-data-redis -->
<dependency>
<groupId>org.springframework.boot</groupId>
<artifactId>spring-boot-starter-data-redis</artifactId>
<version>2.2.2.RELEASE</version>
</dependency>
以下是一个扣减库存的接口作为例子:
@RestController
public class IndexController {
@Autowired
private StringRedisTemplate stringRedisTemplate;
@RequestMapping("/deduct_stock")
public Stirng deductStock() {
int stock = Integer.parseInt(stringRedisTemplate.opsForValue().get("stock"));//jedis.get(key)
if (stock > 0) {
int realStock = stock - 1;
stringRedisTemplate.opsForValue.set("stock",realStock+"");//jedis.set(key,value)
System.out.println(扣减成功,剩余库存:" + realStock + "");
} else {
System.out.println(扣减失败,库存不足!" );
}
return "end";
}
}
1.单实例应用场景
以上代码使用JMeter压测工具进行调用,设置参数为:
Number Of Threads[users]:100
Ramp Up Period[in seconds]:0
Loop Count:2
用单个web调用,结果出现并发问题:
解决方案:加入同步锁(synchronized)
@RestController
public class IndexController {
@Autowired
private StringRedisTemplate stringRedisTemplate;
@RequestMapping("/deduct_stock")
public Stirng deductStock() {
synchronized(this) {
int stock = Integer.parseInt(stringRedisTemplate.opsForValue().get("stock"));//jedis.get(key)
if (stock > 0) {
int realStock = stock - 1;
stringRedisTemplate.opsForValue.set("stock",realStock+"");//jedis.set(key,value)
System.out.println(扣减成功,剩余库存:" + realStock + "");
} else {
System.out.println(扣减失败,库存不足!" );
}
return "end";
}
}
}
2.多实例分布式场景
以上代码,比如有多个应用程序,用nginx做负载均衡,进行同时调用压测
两个程序存在同样的扣减,出现并发现象。
第一个应用扣减结果显示:
第二个应用扣减结果显示:
解决方案:redis的setnx方法(可参考SETNX的api)
多个线程setnx调用时,有且仅有一个线程会拿到这把锁,所以拿到锁的执行业务代码,最后释放掉锁,代码如下:
@RestController
public class IndexController {
@Autowired
private StringRedisTemplate stringRedisTemplate;
@RequestMapping("/deduct_stock")
public Stirng deductStock() {
String lockkey = "lockkey";
Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,"lockvalue");//jedis.setnx
if(!result) {
return "";
}
int stock = Integer.parseInt(stringRedisTemplate.opsForValue().get("stock"));//jedis.get(key)
if (stock > 0) {
int realStock = stock - 1;
stringRedisTemplate.opsForValue.set("stock",realStock+"");//jedis.set(key,value)
System.out.println(扣减成功,剩余库存:" + realStock + "");
} else {
System.out.println(扣减失败,库存不足!" );
}
springRedisTemplate.delete(lockkey);
return "end";
}
}
调用200次,压测结果显示还是有问题,只减掉了一部分:
这时,加大压测次数,结果正常了:
第一个应用扣减结果显示:
第二个应用扣减结果显示:
这个只是因为加大了调用次数,执行业务代码需要一点时间,这段时间拒绝了很多等待获取锁的请求。但是,还是有问题,假如redis服务挂掉了,抛出异常了,这时锁不会被释放掉,出现死锁问题,可以添加try catch处理,代码如下:
@RestController
public class IndexController {
@Autowired
private StringRedisTemplate stringRedisTemplate;
@RequestMapping("/deduct_stock")
public Stirng deductStock() {
String lockkey = "lockkey";
try{
Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,"lockvalue");//jedis.setnx
if(!result) {
return "";
}
int stock = Integer.parseInt(stringRedisTemplate.opsForValue().get("stock"));//jedis.get(key)
if (stock > 0) {
int realStock = stock - 1;
stringRedisTemplate.opsForValue.set("stock",realStock+"");//jedis.set(key,value)
System.out.println(扣减成功,剩余库存:" + realStock + "");
} else {
System.out.println(扣减失败,库存不足!" );
}
}finally{
springRedisTemplate.delete(lockkey);
}
return "end";
}
}
这时,Redis服务挂掉导致死锁的问题解决了,但是,如果服务器果宕机了,又会导致锁不能被释放的现象,所以可以设置超时时间为10s,代码如下:
@RestController
public class IndexController {
@Autowired
private StringRedisTemplate stringRedisTemplate;
@RequestMapping("/deduct_stock")
public Stirng deductStock() {
String lockkey = "lockkey";
try{
Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,"lockvalue",10,TimeUnit.SECONDS);//jedis.setnx
//Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,"lockvalue");//jedis.setnx
//stringRedisTemplate.expire(lockkey,10,TimeUnit.SECONDS);
if(!result) {
return "";
}
int stock = Integer.parseInt(stringRedisTemplate.opsForValue().get("stock"));//jedis.get(key)
if (stock > 0) {
int realStock = stock - 1;
stringRedisTemplate.opsForValue.set("stock",realStock+"");//jedis.set(key,value)
System.out.println(扣减成功,剩余库存:" + realStock + "");
} else {
System.out.println(扣减失败,库存不足!" );
}
}finally{
springRedisTemplate.delete(lockkey);
}
return "end";
}
}
这时,如果有一个线程执行需要15s,当执行到10s时第二个线程进来拿到这把锁,会出现多个线程拿到同一把锁执行,在第一个线程执行完时会释放掉第二个线程的锁,以此类推…就会导致锁的永久失效。所以,只能自己释放自己的锁,可以给当前线程取一个名字,代码如下:
@RestController
public class IndexController {
@Autowired
private StringRedisTemplate stringRedisTemplate;
@RequestMapping("/deduct_stock")
public Stirng deductStock() {
String lockkey = "lockkey";
String clientId = UUID.randomUUID().toString();
try{
Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,clientId ,10,TimeUnit.SECONDS);//jedis.setnx
//Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,"lockvalue");//jedis.setnx
//stringRedisTemplate.expire(lockkey,10,TimeUnit.SECONDS);
if(!result) {
return "";
}
int stock = Integer.parseInt(stringRedisTemplate.opsForValue().get("stock"));//jedis.get(key)
if (stock > 0) {
int realStock = stock - 1;
stringRedisTemplate.opsForValue.set("stock",realStock+"");//jedis.set(key,value)
System.out.println(扣减成功,剩余库存:" + realStock + "");
} else {
System.out.println(扣减失败,库存不足!" );
}
}finally{
springRedisTemplate.delete(lockkey);
}
return "end";
}
}
永久失效的问题解决了,但是,如果第一个线程执行15s,还是会存在多个线程拥有同一把锁的现象。所以,需要续期超时时间,当一个线程执行5s后对超时时间进行续期都10s,就可以解决了,续期设置可以借助redission工具。
Redission使用
Redission分布式锁实现原理:
pom.xml
<dependency>
<groupId>org.redisson</groupId>
<artifactId>redisson</artifactId>
<version>3.6.5</version>
</dependency>
Application.java启动类
@bean
public Redission redission {
//此为单机模式
Config config = new Config();
config.useSingleServer().setAddress("redis://120.0.0.1:6379").setDatabase(0);
return (Redission)Redission.creat(config);
}
最终解决以上所有问题的代码如下:
@RestController
public class IndexController {
@Autowired
private StringRedisTemplate stringRedisTemplate;
@Autowired
private Redissionredission;
@RequestMapping("/deduct_stock")
public Stirng deductStock() {
String lockkey = "lockkey";
//String clientId = UUID.randomUUID().toString();
RLock lock = redission.getLock();
try{
//Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,clientId ,10,TimeUnit.SECONDS);//jedis.setnx
//Boolean result = stringRedisTemplate.opsForValue.setIfAbsent(lockkey,"lockvalue");//jedis.setnx
//stringRedisTemplate.expire(lockkey,10,TimeUnit.SECONDS);
//加锁:redission默认超时时间为30s,每10s续期一次,也可以自己设置时间
lock.lock(60,TimeUnit.SECONDS);
int stock = Integer.parseInt(stringRedisTemplate.opsForValue().get("stock"));//jedis.get(key)
if (stock > 0) {
int realStock = stock - 1;
stringRedisTemplate.opsForValue.set("stock",realStock+"");//jedis.set(key,value)
System.out.println(扣减成功,剩余库存:" + realStock + "");
} else {
System.out.println(扣减失败,库存不足!" );
}
}finally{
lock.unlock();
//springRedisTemplate.delete(lockkey);
}
return "end";
}
}
高并发分布式锁的问题得到解决。
|
|