一、重点
代码讲解:7.3点赞功能-定时持久化到数据库-Java程序整合pipeline+lua_哔哩哔哩_bilibili
https://www.bilibili.com/video/BV1Lg4y1w7U9
代码:
blogLike_schedule/like08 · xin麒/XinQiUtilsOrDemo - 码云 - 开源中国 (gitee.com)
https://gitee.com/flowers-bloom-is-the-sea/XinQiUtilsOrDemo/tree/master/blogLike_schedule/like08
数据库表:
blogLike_schedule · xin麒/XinQiUtilsOrDemo - 码云 - 开源中国 (gitee.com)
二、实现过程:
这里主要是开2个定时任务,一个是平时的缓存数据持久化操作就是要pipeline技术;
另一个时间就使用lua脚本向redis缓存获取缓存数据再更新到数据库。
1、对lua脚本任务执行:
@Slf4j
public class LikeTaskAtNight extends QuartzJobBean {
@Autowired
BlogService blogService;
@Autowired
StringRedisTemplate stringRedisTemplate;
/**
* 用来从redis中的zset中获取点赞缓存的键值对,间隔时间存入mysql
* 这里就夜间执行吧,因为lua会阻塞redis其他操作,选择夜间流量低时进行执行
* @param context
* @throws JobExecutionException
*/
@Override
protected void executeInternal(JobExecutionContext context) throws JobExecutionException {
int hour = LocalTime.now().getHour();
// int minute = LocalTime.now().getMinute();
if (hour == 3){
// 要执行的内容就写在这个函数中
// blogService.updateAllLikeListToDatabase();//这个没使用pipeline技术,效果应该是明显低于updateAllLikeListToDatabaseByPipeline1
log.debug("LikeTaskAtNight - time is {}",System.currentTimeMillis());
// blogService.updateAllLikeListToDatabaseByPipeline1();//这个使用了pipeline技术,但是还有优化空间,比如书对sql的更新可以说批量更新
blogService.updateAllLikeListToDatabaseByLua();//这个使用了lua技术,但是还有优化空间,比如书对sql的更新可以说批量更新
}
}
}
配置文件:
import com.xinqi.task.LikeTask;
import com.xinqi.task.LikeTaskAtNight;
import org.quartz.*;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
@Configuration
public class QuartzConfig {
@Bean
public JobDetail quartzDetail_1(){
// withIdentity指定的是这个job的id
return JobBuilder.newJob(LikeTask.class)
.withIdentity("LIKE_TASK_IDENTITY")
.storeDurably()
.build();
}
@Bean
public Trigger quartzTrigger_1(){ //触发器
SimpleScheduleBuilder scheduleBuilder = SimpleScheduleBuilder.simpleSchedule()
.withIntervalInSeconds(20) //设置时间周期单位秒
// .withIntervalInHours(2) //两个小时执行一次
.repeatForever();
return TriggerBuilder.newTrigger().forJob(quartzDetail_1())
// .forJob(quartzDetail_2())
.withIdentity("LikeTask_TRENDS_TRIGGER")
.withSchedule(scheduleBuilder)
.build();
}
@Bean
public JobDetail quartzDetail_LikeTaskAtNight(){
// withIdentity指定的是这个job的id
return JobBuilder.newJob(LikeTaskAtNight.class)
.withIdentity("LikeTaskAtNight_IDENTITY")
.storeDurably()
.build();
}
@Bean
public Trigger quartzTrigger_2(){ //触发器
SimpleScheduleBuilder scheduleBuilder = SimpleScheduleBuilder.simpleSchedule()
// .withIntervalInSeconds(35) //设置时间周期单位秒
.withIntervalInHours(1) //两个小时执行一次
.repeatForever();
return TriggerBuilder.newTrigger().forJob(quartzDetail_LikeTaskAtNight())
// .forJob(quartzDetail_2())
.withIdentity("LikeTaskAtNight_USER_TRENDS_TRIGGER")
.withSchedule(scheduleBuilder)
.build();
}
}
updateAllLikeListToDatabaseByLua执行的内容:
@Override
public void updateAllLikeListToDatabaseByLua() {
String prefix = "BLOG_LIKED_KEY";
Map<Long, Map<String, String>> maps = getMapsByLuaUseJedis(prefix);
if (maps == null || maps.size() == 0) return;
for (Map.Entry<Long, Map<String, String>> entry : maps.entrySet()) {
Long blogId = entry.getKey();
Map<String, String> likeList = entry.getValue();
updateLikeListByBlogId(blogId, likeList);
}
}
public Map<Long, Map<String, String>> getMapsByLuaUseJedis(String prefix) {
Map<Long, Map<String, String>> maps = null;
String script = "local prefix = KEYS[1]; \n" +
"local redisKeys = redis.call('keys',prefix ..'*');\n" +
"\n" +
"if(not redisKeys) \n" +
" then \n" +
" \treturn (nil);\n" +
"end;\n" +
"\n" +
"local maps = {};\n" +
"\n" +
"for i, v in pairs(redisKeys) do\n" +
" local blogId = string.sub(v,string.len(prefix) + 1,string.len(v));\n" +
" local zset = redis.call('zrange',v,'0','-1','withscores');\n" +
" table.insert(maps,blogId);\n" +
" table.insert(maps,zset); \n" +
"end;\n" +
"\n" +
"return maps;";
List result = null;
result = getResultByLuaUseJedis(script, prefix);
maps = parseLuaResultToMaps(result);
return maps;
}
private List getResultByLuaUseJedis(String script, String prefix) {
Jedis jedis = null;
List result = null;
int failTime = 0;//执行失败的次数
boolean flag = true;//如果为true就说明报错
while (flag && failTime < 2) {
try {
jedis = jedisPool.getResource();
result = (List) jedis.eval(script, Arrays.asList(prefix), new ArrayList<>());//https://blog.csdn.net/EnjoyFight/article/details/127808971
System.out.println("result is " + result);//result is [1, [1, 1688218686914]]
/* log.debug("result is ",result);//不知道为什么无法显示
log.debug("result is null?{}",result==null);
String s = JSONUtil.toJsonStr(result);
log.debug("s is ",s);
log.debug("s is null? {}",s==null);
log.debug("s .len is {}",s.length());*/
flag = false;
} catch (Exception e) {
e.printStackTrace();
failTime++;
} finally {
if (jedis != null) jedis.close();
}
}
return result;
}
2、pipeline技术定时任务
看这篇文章吧:(108条消息) springboot+redis+mysql+quartz-通过Java操作jedis使用pipeline获取缓存数据定时更新数据库_xin麒的博客-CSDN博客
三、其他
本文章的具体内容基本上都是在视频里面了,就不多描述了。
其他类似的文章:
(108条消息) springboot+redis+mysql+quartz-通过Java操作redis的KEYS*命令获取缓存数据定时更新数据库_xin麒的博客-CSDN博客
(108条消息) springboot+redis+mysql+quartz-通过Java操作jedis定时使用lua脚本获取缓存数据并更新数据库_xin麒的博客-CSDN博客
(108条消息) lua脚本获取table类型-Java使用lua脚本操作redis获取zset元素的集合_xin麒的博客-CSDN博客
(108条消息) springboot+redis+mysql+quartz-通过Java操作jedis使用pipeline获取缓存数据定时更新数据库_xin麒的博客-CSDN博客
(108条消息) springboot+redis+mysql+quartz-通过Java操作jedis的scan命令获取缓存数据定时更新数据库_xin麒的博客-CSDN博客