由于业务需要，在公司用了Redis去存储Token，用于用户登录验证以及权限校验。原本的Redis方案是一个单体的Redis，存在当Redis节点down掉之后，整套分布式微服务都不可用的风险，于是进行Redis高可用方案的技术选型，综合考虑各种方案的可用性、并发性和复杂度，最后选择了使用Redis哨兵的方案。
1. Redis Sentinel原理
简单贴个图，就不细讲了，本篇主要是记录应用过程，不涉及太多原理讲解
2. Redis哨兵一主两从搭建
2.1 环境准备
Linux服务器，下载Redis安装包
a. wget http://download.redis.io/releases/redis-5.0.5.tar.gz b.解压 tar -xvf redis-5.0.5.tar.gz cd redis-5.0.5 make make install
2.2 Redis服务搭建
主redis配置文件
# redis_master.conf
protected-mode no
port 6380
requirepass "${redis密码}"
daemonize yes
# 先创建对应的文件夹
dir "/data/redis/redis_slave_one"
logfile "/data/redis/redis_slave_one.log"
masterauth "${redis密码}"
client-output-buffer-limit normal 0 0 0
# Generated by CONFIG REWRITE
client-output-buffer-limit replica 512mb 128mb 120
replica-read-only no
二从配置文件
# redis_slave_one.conf
protected-mode no
port 6381
requirepass ${redis密码}
daemonize yes
# 先创建对应的文件夹
dir "/data/redis/redis_slave_two"
logfile "/data/redis/redis_slave_two.log"
replicaof ${ip} 6380
masterauth ${redis密码}
client-output-buffer-limit replica 512mb 128mb 120
replica-read-only no
--------------------------------------------
# redis_slave_two.conf
protected-mode no
port 6382
requirepass "${redis密码}"
daemonize yes
# 先创建对应的文件夹
dir "/data/redis/redis_master"
logfile "/data/redis/redis_master.log"
masterauth "${redis密码}"
client-output-buffer-limit normal 0 0 0
# Generated by CONFIG REWRITE
replicaof ${ip} 6380
client-output-buffer-limit replica 512mb 128mb 120
replica-read-only no
启动主redis:/usr/local/redis-5.0.5/src/redis-server /etc/redis/redis_master.conf
启动两从：
/usr/local/redis-5.0.5/src/redis-server /etc/redis/redis_slave_one.conf
/usr/local/redis-5.05/src/redis-server /etc/redis/redis_slave_two.conf
2.3 Redis Sentinel搭建
Sentinel配置文件
# master的守护者文件配置
port 26380
# 先创建对应的文件夹
dir "/data/redis/master_sentinel"
# sentinel启动后日志文件
logfile "/data/redis/sentinel/master_sentinel.log"
# 启动sentinel是否是后台应用程序，默认是no，修改成yes后台启动
daemonize yes
# 格式：sentinel <option_name> <master_name> <option_value>；#该行的意思是：监控的master的名字叫做mymaster（自定义）,地址为127.0.0.1:6378 ，行尾最后的一个1代表在sentinel集群中，多少个sentinel认为masters死了，才能真正认为该master不可用了，该值需要小于当前到sentinel个数(基数个)，不然启动sentinel一直提示 +sentinel-address-switch 信息。
sentinel monitor mymaster ${ip} 6380 2
# sentinel会向master发送心跳PING来确认master是否存活，如果master在“一定时间范围”内不回应PONG 或者是回复了一个错误消息，那么这个sentinel会主观地(单方面地)认为这个master已经不可用了(subjectively down, 也简称为SDOWN)。而这个down-after-milliseconds就是用来指定这个“一定时间范围”的，单位是毫秒，默认30秒。
sentinel down-after-milliseconds mymaster 15000
# failover过期时间，当failover开始后，在此时间内仍然没有触发任何failover操作，当前sentinel将会认为此次failoer失败。默认180秒，即3分钟。
sentinel failover-timeout mymaster 120000
# 在发生failover主备切换时，这个选项指定了最多可以有多少个slave同时对新的master进行同步，这个数字越小，完成failover所需的时间就越长，但是如果这个数字越大，就意味着越多的slave因为replication而不可用。可以通过将这个值设为 1 来保证每次只有一个slave处于不能处理命令请求的状态。
sentinel deny-scripts-reconfig yes
sentinel auth-pass mymaster ${redis密码}
启动主Redis的哨兵：
/user/local/redis-5.0.5/src/redis-sentinel /etc/redis/redis_master_sentinel.conf
也可:
/user/local/redis-5.0.5/src/redis-server /etc/redis/redis_master_sentinel.conf --sentinel
启动两从Redis的哨兵：
/usr/local/redis-5.0.5/src/redis-sentinel /etc/redis/redis_slave_one_sentinel.conf
/usr/local/redis-5.0.5/src/redis-sentinel /etc/redis/redis_slave_two_sentinel.conf
看看结果：
首先是服务有没有正常启动：
然后是一主两从：
最后是哨兵是否监控到了redis：
ok，都没有问题，接着下一步。
2.4 测试主从复制
主库添加一个key
查看从库是否有记录：
ok，没有问题。
用的redis连接工具是RedisInsight，需要的话可以去官网下载一个 Redis连接工具下载
2.5 测试自动切换主从库
验证完了，都没什么问题，继续下一步(记得重启一下6380服务)。
3. SpringBoot集成
<!--版本号我是根据自己的springboot版本来的，如果有问题的话可以看一下是不是和自己的版本不匹配--> <dependency> <groupId>org.springframework.boot</groupId> <artifactId>spring-boot-starter-data-redis</artifactId> <version>2.2.8.RELEASE</version> </dependency> <dependency> <groupId>org.apache.commons</groupId> <artifactId>commons-pool2</artifactId> <version>2.6.1</version> </dependency>
spring:
data:
redis:
repositories:
enabled: false
redis:
#集群模式的配置start
password: ${password}
sentinel:
master: mymaster
nodes: ${ip}:26382,${ip}:26380,${ip}:26381 # 哨兵的IP:Port列表
#集群模式的配置stop
#单机模式的配置start
# host: ${ip}
# password: ${password}
# port: 6379
#单机模式的配置stop
timeout: 5000
database: 0
lettuce:
pool:
max-active: 8
max-wait: -1
max-idle: 8
min-idle: 0
myconfig:
time-to-live: 86400
package xx;
import com.fasterxml.jackson.annotation.JsonAutoDetect;
import com.fasterxml.jackson.annotation.PropertyAccessor;
import com.fasterxml.jackson.databind.DeserializationFeature;
import com.fasterxml.jackson.databind.ObjectMapper;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.springframework.beans.factory.annotation.Autowired;
import org.springframework.cache.CacheManager;
import org.springframework.cache.annotation.CachingConfigurerSupport;
import org.springframework.cache.annotation.EnableCaching;
import org.springframework.cache.interceptor.KeyGenerator;
import org.springframework.context.annotation.Bean;
import org.springframework.context.annotation.Configuration;
import org.springframework.context.annotation.Primary;
import org.springframework.data.redis.cache.RedisCacheConfiguration;
import org.springframework.data.redis.cache.RedisCacheManager;
import org.springframework.data.redis.connection.RedisConnectionFactory;
import org.springframework.data.redis.core.*;
import org.springframework.data.redis.serializer.GenericJackson2JsonRedisSerializer;
import org.springframework.data.redis.serializer.RedisSerializationContext;
import org.springframework.data.redis.serializer.RedisSerializer;
import org.springframework.data.redis.serializer.StringRedisSerializer;
import javax.annotation.Resource;
import java.lang.reflect.Method;
import java.lang.reflect.Proxy;
import java.time.Duration;
/**
* 缓存配置类
* @author xxx
* @Date 20xx-xx-xx xx:xx
**/
@Configuration
@EnableCaching
public class RedisConfig extends CachingConfigurerSupport {
private static final Logger logger = LoggerFactory.getLogger(RedisConfig.class);
@Resource
private RedisConnectionFactory factory;
/**
* 默认两小时
*/
private static final long DURATION_SECOND_7200 = 7200L;
private static final long DURATION_SECOND_300 = 300L;
@Override
@Bean
public KeyGenerator keyGenerator() {
return new KeyGenerator() {
@SuppressWarnings("rawtypes")
@Override
public Object generate(Object target, Method method, Object... params) {
StringBuilder sb = new StringBuilder(RedisAutoCacheValue.AUTO_KEY_PREFIX);
if(target instanceof Proxy) {
//如果是代理类
Class[] i = target.getClass().getInterfaces();
if(i != null && i.length > 0) {
//取第一个即可
sb.append(i[0].getName());
}else {
sb.append(target.getClass().getName());
}
} else if(target instanceof org.springframework.cglib.proxy.Factory){
//如果是cglib代理，需要手动去除 $$ 后面的
String className = target.getClass().getName();
sb.append(className, 0, className.indexOf("$$"));
} else {
sb.append(target.getClass().getName());
}
sb.append(".");
sb.append(method.getName());
sb.append("_");
for (Object obj : params) {
if (obj != null) {
Class cls = obj.getClass();
if (cls.isArray()) {
//对于基础数据处理
logger.info("keyGenerator : {}", cls.getComponentType());
if (cls.isAssignableFrom(long.class)) {
long[] ay = (long[]) obj;
for (long o : ay) {
sb.append(o).append("");
}
} else if (cls.isAssignableFrom(int.class)) {
int[] ay = (int[]) obj;
for (int o : ay) {
sb.append(o).append("");
}
} else if (cls.isAssignableFrom(float.class)) {
float[] ay = (float[]) obj;
for (float o : ay) {
sb.append(o).append("");
}
} else if (cls.isAssignableFrom(double.class)) {
double[] ay = (double[]) obj;
for (double o : ay) {
sb.append(o).append("");
}
} else if (cls.isAssignableFrom(String.class)) {
String[] ay = (String[]) obj;
for (String o : ay) {
sb.append(o).append("");
}
} else {
sb.append(obj.toString());
}
//TODO 对其他类型数组处理
} else {
sb.append(obj.toString());
}
} else {
sb.append("null");
}
sb.append("_");
//sb.append(obj == null ? "null" : obj.toString());
}
sb.delete(sb.length()-1, sb.length());
return sb.toString();
}
};
}
/**
* 默认的缓存管理，存放时效较长的缓存
* @param redisTemplate
* @return
*/
@SuppressWarnings({"rawtypes", "Duplicates"})
@Primary
@Bean
public CacheManager cacheManager(RedisTemplate redisTemplate) {
RedisCacheConfiguration config = RedisCacheConfiguration
.defaultCacheConfig()
//过期时间
.entryTtl(Duration.ofSeconds(DURATION_SECOND_7200))
//不缓存null值
//.disableCachingNullValues()
//明确manager中的序列化与template一样，防止莫名其妙的问题
.serializeKeysWith(RedisSerializationContext.SerializationPair.fromSerializer(this.keySerializer()))
.serializeValuesWith(RedisSerializationContext.SerializationPair.fromSerializer(this.valueSerializer()));
RedisCacheManager rcm = RedisCacheManager.builder(redisTemplate.getConnectionFactory()).cacheDefaults(config).transactionAware().build();
return rcm;
}
/**
* 存放时效较短的缓存（5分钟）
* @param redisTemplate
* @return
*/
@SuppressWarnings({"rawtypes", "Duplicates"})
@Bean
public CacheManager cacheManagerIn5Minutes(RedisTemplate redisTemplate) {
RedisCacheConfiguration config = RedisCacheConfiguration
.defaultCacheConfig()
//过期时间
.entryTtl(Duration.ofSeconds(DURATION_SECOND_300))
//不缓存null值
//.disableCachingNullValues()
//明确manager中的序列化与template一样，防止莫名其妙的问题
.serializeKeysWith(RedisSerializationContext.SerializationPair.fromSerializer(this.keySerializer()))
.serializeValuesWith(RedisSerializationContext.SerializationPair.fromSerializer(this.valueSerializer()));
RedisCacheManager rcm = RedisCacheManager.builder(redisTemplate.getConnectionFactory()).cacheDefaults(config).transactionAware().build();
return rcm;
}
/*@SuppressWarnings({"rawtypes", "unchecked"})
@Bean
public RedisTemplate<String, Object> redisTemplate(RedisConnectionFactory factory) {
//factory = connectionFactory(3,"172.20.11.134",6379,"123456",2000,100,1,1000,2000);
StringRedisTemplate template = new StringRedisTemplate(factory);
template.setKeySerializer(keySerializer());
template.setHashKeySerializer(keySerializer());
template.setValueSerializer(valueSerializer());
template.setHashValueSerializer(valueSerializer());
template.afterPropertiesSet();
return template;
}
*/
private RedisSerializer<String> keySerializer() {
return new StringRedisSerializer();
}
private RedisSerializer<Object> valueSerializer() {
// Jackson2JsonRedisSerializer jackson2JsonRedisSerializer = new Jackson2JsonRedisSerializer(Object.class);
// ObjectMapper om = new ObjectMapper();
// om.setVisibility(PropertyAccessor.ALL, JsonAutoDetect.Visibility.ANY);
// om.enableDefaultTyping(ObjectMapper.DefaultTyping.NON_FINAL);
// //略过不匹配的属性
// om.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
//
// jackson2JsonRedisSerializer.setObjectMapper(om);
// return jackson2JsonRedisSerializer;
ObjectMapper om = new ObjectMapper();
om.setVisibility(PropertyAccessor.ALL, JsonAutoDetect.Visibility.ANY);
om.enableDefaultTyping(ObjectMapper.DefaultTyping.NON_FINAL);
//略过不匹配的属性
om.configure(DeserializationFeature.FAIL_ON_UNKNOWN_PROPERTIES, false);
GenericJackson2JsonRedisSerializer genericJackson2JsonRedisSerializer = new GenericJackson2JsonRedisSerializer(om);
return genericJackson2JsonRedisSerializer;
}
// ==========注解：以上是用来适配Cacheable缓存注解的配置，自定义缓存类型和时长=============================================
// ==========手动：以下是用来适配原来的Redis的配置，用于手动添加Redis缓存，现在把Redis做成新版的缓存配置==================
@Bean
public RedisTemplate<String, Object> redisTemplate() {
RedisTemplate<String, Object> redisTemplate = new RedisTemplate<>();
redisTemplate.setKeySerializer(new StringRedisSerializer());
redisTemplate.setHashKeySerializer(new StringRedisSerializer());
redisTemplate.setHashValueSerializer(new StringRedisSerializer());
redisTemplate.setValueSerializer(new StringRedisSerializer());
redisTemplate.setConnectionFactory(factory);
return redisTemplate;
}
@Bean
public HashOperations<String, String, Object> hashOperations(RedisTemplate<String, Object> redisTemplate) {
return redisTemplate.opsForHash();
}
@Bean
public ValueOperations<String, Object> valueOperations(RedisTemplate<String, Object> redisTemplate) {
return redisTemplate.opsForValue();
}
@Bean
public ListOperations<String, Object> listOperations(RedisTemplate<String, Object> redisTemplate) {
return redisTemplate.opsForList();
}
@Bean
public SetOperations<String, Object> setOperations(RedisTemplate<String, Object> redisTemplate) {
return redisTemplate.opsForSet();
}
@Bean
public ZSetOperations<String, Object> zSetOperations(RedisTemplate<String, Object> redisTemplate) {
return redisTemplate.opsForZSet();
}
}
然后调用该接口，看一下redis中是否存入该数据，再次调用，查看前后结果是否一致，如果一致，说明第二次调用直接从缓存中拿的数据，而不是重新生成一个随机数。
Over！！！