当前位置:网站首页>Redis seckill demo
Redis seckill demo
2022-07-01 15:35:00 【y_ w_ x_ k】
Scheme 1 : Method plus synchronized that will do , Thread safety scheme , But the efficiency of execution is low , Not suitable for large concurrent occasions
Option two : Use optimistic lock
controller:
@RestController
public class SecondKill {
@Autowired
SecondKillSev secondKillSev;
@RequestMapping(value = "/buy",method = RequestMethod.POST)
public Mes buy(@RequestParam("proid") String proid){
String usedId=new Random().nextInt(1000)+"";
Mes mes = secondKillSev.doSecondKill(usedId, proid);
return mes;
}
}
service:
@Service
public class SecondKillSev {
Mes message;
public SecondKillSev(){
}
@Autowired
StringRedisTemplate stringRedisTemplate;
/**
* Execute seckill
* Scheme 1 :synchronized, Thread safety , But the efficiency of execution is low , Not suitable for large concurrent occasions
* Option two : Monitor transactions with optimistic locks
*/
public Mes doSecondKill(String uid, String proid){
//1. Judge uid and proid Non empty
if(uid==null || proid==null){
return Mes.getMes("400","uid and proid Can't be empty ");
}
//2. Connect redis
ValueOperations<String, String> ops = stringRedisTemplate.opsForValue();
//3. Splicing key, stock key, Successful user key
String kcKey="kcKey:"+proid;
String userKey="userKey:"+proid;
// Optimistic lock solves oversold problem
//stringRedisTemplate The default is not to open things
stringRedisTemplate.setEnableTransactionSupport(true);
List<Object> txResults = stringRedisTemplate.execute(new SessionCallback<List<Object>>() {
public List<Object> execute(RedisOperations operations) throws DataAccessException {
// Monitor inventory
operations.watch(kcKey);
//4. Get inventory , If the inventory is null, The second kill has not yet started
String kc=ops.get(kcKey);
if(kc==null){
message= Mes.getMes("400"," The stock is null, The second kill hasn't started yet ");
return null;
}
//5. Judge whether the user repeats the second kill
if(operations.opsForSet().isMember(userKey, uid)){
message=Mes.getMes("400"," You can't repeat the second kill ");
return null;
}
//6. Judge the quantity of goods , If the quantity is less than 1, Then the second kill is over
if(Integer.parseInt(kc)<=0){
message=Mes.getMes("400"," The second kill is over ");
return null;
}
//7. Second kill process , stock -1, Successful user added to manifest
// With a transaction
operations.multi();
operations.opsForValue().decrement(kcKey);
operations.opsForSet().add(userKey,uid);
List exec = operations.exec();
if(exec==null || exec.size()==0){
message=Mes.getMes("400"," Seckill failure ");
return null;
}else{
return exec;
}
}
});
if(txResults==null ||txResults.size()==0){
return message;
}
return Mes.getMes("200"," Seckill success ");
}
}
But this scheme also has a disadvantage , That is, if the quantity of goods is 500, But the number of threads is 1000, That is, when the number of visits is not very large , There will be the problem that the inventory is not sold out , That is, inventory leftover problems , So it can be used lua Script to solve ;
Option three :lua Script
local userid=KEYS[1];
local prodid=KEYS[2];
local qtkey="sk:"..prodid..":qt";
local userskey="sk:"..prodid..":user";
local userExists=redis.call("sismember",userskey,userid);
if tonumber(userExists)==1 then
return 2;
end
local num=redis.call("get",qtkey);
if tonumber(num)<=0 then
return 0;
else
redis.call("decr",qtkey);
redis.call("sadd",userskey,userid);
end
return 1
Code implementation :
@Service
public class SeckillByScript {
static String secKillScript1 = "local userid=KEYS[1];\n" +
"local prodid=KEYS[2];\n" +
"local qtkey=\"sk:\"..prodid..\":qt\";\n" +
"local userskey=\"sk:\"..prodid..\":user\";\n" +
"local userExists=redis.call(\"sismember\",userskey,userid);" +
"if tonumber(userExists)==1 then\n" +
" return 2;\n" +
"end\n" +
"local num=redis.call(\"get\",qtkey);\n" +
"if tonumber(num)<=0 then\n" +
" return 0;\n" +
"else\n" +
" redis.call(\"decr\",qtkey);\n" +
" redis.call(\"sadd\",userskey,userid);\n" +
" end\n" +
" return 1";
public boolean doSecKill(String userid, String prodid) {
JedisPool jedisPool = JedisPoolUtil.getJedisPool();
Jedis jedis = jedisPool.getResource();
String sha1 = jedis.scriptLoad(secKillScript1);
Object result = jedis.evalsha(sha1, 2, userid, prodid);
String reString = String.valueOf(result);
if ("0".equals(reString)) {
System.out.println(" Empty !");
return false;
} else if ("1".equals(reString)) {
System.out.println(" Panic buying ");
return true;
} else if ("2".equals(reString)) {
System.out.println(" The user has robbed !");
return false;
} else {
System.out.println(" Panic buying exception ");
return false;
}
}
}
边栏推荐
- TS reports an error don't use 'object' as a type The `object` type is currently hard to use
- Redis秒杀demo
- [Cloudera][ImpalaJDBCDriver](500164)Error initialized or created transport for authentication
- The solution to turn the newly created XML file into a common file in idea
- The difference between arrow function and ordinary function in JS
- Phpcms background upload picture button cannot be clicked
- 22-06-26周总结
- Basic use process of cmake
- STM32ADC模拟/数字转换详解
- Wechat applet 03 - text is displayed from left to right, and the block elements in the line are centered
猜你喜欢
Qt+pcl Chapter 6 point cloud registration ICP Series 2
[leetcode] 16. The sum of the nearest three numbers
异常检测中的浅层模型与深度学习模型综述(A Unifying Review of Deep and Shallow Anomaly Detection)
Photoshop plug-in HDR (II) - script development PS plug-in
6.2 normalization 6.2.6 BC normal form (BCNF) 6.2.9 normalization summary
Zhang Chi Consulting: household appliance enterprises use Six Sigma projects to reduce customers' unreasonable return cases
Stm32f4-tft-spi timing logic analyzer commissioning record
The solution to turn the newly created XML file into a common file in idea
《QT+PCL第六章》点云配准icp系列6
Filter & (login interception)
随机推荐
Summary of point cloud reconstruction methods I (pcl-cgal)
Microservice tracking SQL (support Gorm query tracking under isto control)
[advanced ROS] lesson 5 TF coordinate transformation in ROS
智能运维实战:银行业务流程及单笔交易追踪
Qt+pcl Chapter 6 point cloud registration ICP Series 2
张驰课堂:六西格玛数据的几种类型与区别
Tableapi & SQL and MySQL grouping statistics of Flink
选择在长城证券上炒股开户可以吗?安全吗?
Qt+pcl Chapter 6 point cloud registration ICP series 4
[antenna] [3] some shortcut keys of CST
远程办公经验?来一场自问自答的介绍吧~ | 社区征文
Research on manually triggering automatic decision of SAP CRM organization model with ABAP code
《QT+PCL第六章》点云配准icp系列3
张驰咨询:家电企业用六西格玛项目减少客户非合理退货案例
Qt+pcl Chapter 9 point cloud reconstruction Series 2
Phpcms background upload picture button cannot be clicked
ABAP-屏幕切换时,刷新上一个屏幕
Introduction to MySQL audit plug-in
MySQL审计插件介绍
Beilianzhuguan joined the dragon lizard community to jointly promote carbon neutralization