CacheCloud
CacheCloud提供一个Redis云管理平台：实现多种类型(Redis Standalone、Redis Sentinel、Redis Cluster)自动部署、解决Redis实例碎片化现象、提供完善统计、监控、运维功能、减少运维成本和误操作，提高机器的利用率，提供灵活的伸缩性，提供方便的接入客户端。
项目主页： https://github.com/sohutv/cac…

改造RedisConnectionFactory

/**
 * 根据缓存策略的不同，RedisConnectionFactory不同
 * 示例是单机模式。
 *
 * @return
 */
@Bean
public RedisConnectionFactory redisConnectionFactory() {
   while (true) {
        try {
            LOCK.tryLock(100, TimeUnit.MILLISECONDS);
            /**
             * 心跳返回的请求为空；
             */
            String response = HttpUtils.doGet("http://localhost:5005/cache/client/redis/standalone/10000.json?clientVersion=1.0-SNAPSHOT");
            if (response == null || response.isEmpty()) {
                continue;
            }
            JSONObject jsonObject = null;
            try {
                jsonObject = JSONObject.parseObject(response);
            } catch (Exception e) {
            }
            if (jsonObject == null) {
                continue;
            }
            /**
             * 从心跳中提取HostAndPort，构造JedisPool实例；
             */
            String instance = jsonObject.getString("standalone");
            String[] instanceArr = instance.split(":");
            if (instanceArr.length != 2) {
                continue;
            }

            //收集上报数据
            ClientDataCollectReportExecutor.getInstance("http://localhost:5005/cachecloud/client/reportData.json");

            String password = jsonObject.getString("password");
            String host = instanceArr[0];
            String port = instanceArr[1];

            JedisConnectionFactory jedisConnectionFactory = new JedisConnectionFactory();
            jedisConnectionFactory.setPassword(password);
            jedisConnectionFactory.setHostName(host);
            jedisConnectionFactory.setPort(Integer.parseInt(port));
            return jedisConnectionFactory;
        } catch (InterruptedException e) {
            logger.error("error in build().", e);
        }

    }
}

改造 jedis-2.9.0

Connection.java

/**
 * 命令捕获，异常保存
 * @param cmd
 * @param args
 */
public void sendCommand(final ProtocolCommand cmd, final byte[]... args) {
    try {
        //统计开始
        UsefulDataModel costModel = UsefulDataModel.getCostModel(threadLocal);
        costModel.setCommand(cmd.toString().toLowerCase());
        costModel.setStartTime(System.currentTimeMillis());
        connect();
        Protocol.sendCommand(outputStream, cmd, args);
    } catch (JedisConnectionException ex) {
        UsefulDataCollector.collectException(ex, getHostPort(), System.currentTimeMillis());
        broken = true;
        throw ex;
    }
}

JedisClusterCommand.java

private T runWithRetries(byte[] key, int attempts, boolean tryRandomNode, boolean asking) {
    if (attempts <= 0) {
        JedisClusterMaxRedirectionsException exception = new JedisClusterMaxRedirectionsException("Too many Cluster redirections? key=" + SafeEncoder.encode(key));
        //收集
        UsefulDataCollector.collectException(exception, "", System.currentTimeMillis(), ClientExceptionType.REDIS_CLUSTER);
        throw exception;
    }
}

更新spring-boot-starter-data-redis依赖

 <!--Redis-->
<dependency>
    <groupId>org.springframework.boot</groupId>
    <artifactId>spring-boot-starter-data-redis</artifactId>
    <exclusions>
        <exclusion>
            <artifactId>jedis</artifactId>
            <groupId>redis.clients</groupId>
        </exclusion>
    </exclusions>
</dependency>
<dependency>
    <groupId>com.sohu.tv</groupId>
    <artifactId>cachecloud-open-client-redis</artifactId>
    <version>1.0-SNAPSHOT</version>
    <exclusions>
        <exclusion>
            <artifactId>jedis</artifactId>
            <groupId>redis.clients</groupId>
        </exclusion>
    </exclusions>
</dependency>
<!--上步改造后编译的jar-->
<dependency>
    <groupId>com.github.pig</groupId>
    <artifactId>pig-cache-cloud-jedis</artifactId>
    <version>2.9.1</version>
</dependency>

部署服务war

这一步直接参考 cachecloud 的文档即可

总结

1. 源码，参考pig： https://gitee.com/log4j/pig
2. 关于pig， 这是一套微服务应用的实践基于Spring Cloud、Spring Security Oauth2.0开发企业级认证与授权，提供常见服务监控、链路追踪、日志分析、缓存管理、任务调度等实现
3. 改造写于2017-2018的跨年夜，文章整理与2018.1.1，真惨，不过我喜欢 ?
4. 2017 失去很多，2018本命年 ? 汪汪汪！