使用lettuce和redisTemplate操作redis cluster踩坑日记

环境：虚拟机3主3从

1、关闭slave是对读写都没有影响的

2、关键就是关闭master,读写都会失败

虽说可以设置lettuce的拓扑自动更新，但是redis的slave变成master是需要时间的
在这段时间内的读写都无法进行,异常-->【connection refused】
等到slave变成了master,读写才会恢复,但是依然会报警告,无法连接到xxx，哎...

找了半天发现有个方法可以避免读的失败，就是优先从slave读
    StatefulRedisClusterConnection对象之
    setReadFrom(ReadFrom.REPLICA_PREFERRED);
通过测试发现只有REPLICA_PREFERRED有效，其他无效
这个选项意思是优先从副本读取,master宕机或者slave宕机都不会对读造成任何影响
NICE啊。。。

3、使用lettuce操作cluster

    |--工具类

/**
 * lettuce获取cluster连接
 */
public class ClusterUtil {
    private static final List<RedisURI> redisURIList = new ArrayList<>();
    private static final String address =
                    "192.168.8.131:6379," +
                    "192.168.8.129:6379," +
                    "192.168.8.130:6379," +
                    "192.168.8.132:6379," +
                    "192.168.8.134:6379," +
                    "192.168.8.133:6379";
    private static final String auth = "******";
    private static RedisClusterClient client;
    private static ThreadLocal<StatefulRedisClusterConnection> box = new ThreadLocal<>();
    static {
        for (String x : address.split(",")) {
            StringBuilder sb = new StringBuilder();
            sb.append("redis://").append(auth).append("@").append(x);
            redisURIList.add(RedisURI.create(sb.toString()));
        }
        client = RedisClusterClient.create(redisURIList);
        //拓扑更新设置
        ClusterTopologyRefreshOptions.Builder topoBuilder = ClusterTopologyRefreshOptions.builder();
        topoBuilder.enableAllAdaptiveRefreshTriggers();
        topoBuilder.enablePeriodicRefresh(true);
        topoBuilder.dynamicRefreshSources(true);
        topoBuilder.closeStaleConnections(true);
        //client选项设置
        ClusterClientOptions.Builder clientOptionsBuilder = ClusterClientOptions.builder();
        clientOpti