Redis cluster集群：原理及搭建

1.为什么使用redis？

redis是一种典型的no-sql 即非关系数据库像python的字典一样存储key-value键值对工作在memory中

所以很适合用来充当整个互联网架构中各级之间的cache 比如lvs的4层转发层 nginx的7层代理层

尤其是lnmp架构应用层如php-fpm或者是Tomcat到mysql之间做一个cache 以减轻db的压力

因为有相当一部分的数据只是简单的key-value对应关系，而且在实际的业务中常常在短时间内迅速变动

如果用关系数据库mysql之类存储会大大增加对db的访问导致db的负担很重因为所有的require中的大部分最后都要汇聚到db

所以如果想要业务稳定那么解决db的压力就是关键所以现在大部分的解决方案就是在db层之上的各级使用多级的no-sql

像memcache redis 等来为db提供缓冲

2.为什么使用redis-cluster？

为了在大流量访问下提供稳定的业务，集群化是存储的必然形态

未来的发展趋势肯定是云计算和大数据的紧密结合只有分布式架构能满足要求

如果没有集群化何来的分布式？

3.顺带一提总结一波今天的redis原理之数据持久化

虽然redis这种no-sql一般都是作为cache来服务但是如果完全没有数据可持久化的方法那么显得有些单薄

就像memcache 由于这种no-sql是工作在memory的那么由于memory的实体是ram

所以如果重启或者宕机 memory中的数据就全没了数据的一致性的不到保障

但是 redis不同 redis有相对的数据持久化的方案由两种方式构成 aof & rdb

aof就像关系数据库中的binlog一样把每一次写操作以追加的形式记录在其中以文件的形式刷到磁盘里

并且可以使用不同的fsync策略无fsync,每秒fsync,每次写的时候fsync.

使用默认的每秒fsync策略,Redis的性能依然很好(fsync是由后台线程进行处理的,主线程会尽力处理客户端请求)

一旦出现故障，最多丢失1秒的数据.

但是缺点也随之而来那就是aof文件的大小会随着时间线性增长一段时间之后就会变得很大

如果要在一端以AOF的形式来恢复数据那么由于AOF文件的巨大体积可能会让进程如同假死一样十分的慢

rdb则是一种快照机制

redis工作在内存中 rdb就是每隔一段时间对内存中的数据做一次快照保存在rdb文件中

而且redis的主从同步可以实现异步也是由于rdb的机制他在做快照时会fork出一个子进程由子进程来做快照

父进程完全处理请求毫不影响很适合数据的备份

但是问题是如果数据量很大的话 rdb它要保存一个完整的数据集是一个大的工作如果时间间隔设置的太短

那么严重影响redis的性能但是按照常规设置的话如5分钟一次那么如果宕机或者重启就会基于上次做rdb的时间

从而丢失分钟级的数据

point：在redis4.0的新特性中采用了aof-rdb的混合方案来保障数据的持久性但是官方的说法是还不成熟

是一个长期的工作所以有待观察吧

集群实现方案：

关于redis的集群化方案目前有三种

（1）Twitter开发的twemproxy

（2）豌豆荚开发的codis

（3）redis官方的redis-cluster

简介：twemproxy架构简单就是用proxy对后端redis server进行代理但是由于代理层的消耗性能很低而且通常涉及多个key的操作都是不支持的而且本身不支持动态扩容和透明的数据迁移而且也失去维护 Twitter内部已经不使用了

redis-cluster是三个里性能最强大的因为他使用去中心化的思想使用hash slot方式将16348个hash slot 覆盖到所有节点上对于存储的每个key值使用CRC16(KEY)&16348=slot 得到他对应的hash slot 并在访问key时就去找他的hash slot在哪一个节点上然后由当前访问节点从实际被分配了这个hash slot的节点去取数据节点之间使用轻量协议通信减少带宽占用性能很高自动实现负载均衡与高可用自动实现failover 并且支持动态扩展官方已经玩到可以1000个节点实现的复杂度低总之个人比较喜欢这个架构因为他的去中心化思想免去了proxy的消耗是全新的思路

但是它也有一些不足例如官方没有提供图形化管理工具运维体验差全手工数据迁移并且自己对自己本身的redis命令支持也不完全等但是这些问题我觉得不能掩盖他关键的新思想所带来的的优势随着官方的推进这些问题应该都能在一定时间内得到解决那么这时候去中心化思想带来的高性能就会表现出他巨大的优势

codis使用的也是proxy思路但是做的比较好是这两种之间的一个中间级而且支持redis命令是最多的有图形化GUI管理和监控工具运维友好这个过段时间会详细另外写出来原理工作机制和搭建实现

Redis集群

基本介绍

Redis 集群是一个可以在多个 Redis 节点之间进行数据共享的设施installation。

Redis 集群不支持那些需要同时处理多个键的 Redis 命令，因为执行这些命令需要在多个 Redis 节点之间移动数据，并且在高负载的情况下，这些命令将降低Redis集群的性能，并导致不可预测的行为。

Redis 集群通过分区partition来提供一定程度的可用性availability：即使集群中有一部分节点失效或者无法进行通讯，集群也可以继续处理命令请求。

Redis集群提供了以下两个好处：

将数据自动切分split到多个节点的能力。
当集群中的一部分节点失效或者无法进行通讯时，仍然可以继续处理命令请求的能力。

集群原理

redis-cluster架构图

所有的redis节点彼此互联(PING-PONG机制),内部使用二进制协议优化传输速度和带宽。
节点的fail是通过集群中超过半数的节点检测失效时才生效。
客户端与redis节点直连,不需要中间proxy层.客户端不需要连接集群所有节点,连接集群中任何一个可用节点即可。
redis-cluster把所有的物理节点映射到[0-16383]slot上,cluster 负责维护node<->slot<->value

Redis集群中内置了 16384 个哈希槽，当需要在 Redis 集群中放置一个 key-value 时，redis 先对key 使用 crc16 算法算出一个结果，然后把结果对 16384 求余数，这样每个 key 都会对应一个编号在 0-16383 之间的哈希槽，redis 会根据节点数量大致均等的将哈希槽映射到不同的节点

redis-cluster投票:容错

投票过程是集群中所有master参与,如果半数以上master节点与master节点通信超时(cluster-node-timeout),认为当前master节点挂掉.
什么时候整个集群不可用(cluster_state:fail)?
- 如果集群任意master挂掉,且当前master没有slave.集群进入fail状态,也可以理解成集群的slot映射[0-16383]不完整时进入fail状态.
  - redis-3.0.0.rc1加入cluster-require-full-coverage参数,默认关闭,打开集群兼容部分失败.
- 如果集群超过半数以上master挂掉，无论是否有slave，集群进入fail状态.

Redis集群搭建

Redis单机版的安装见博客《redis入门——安装篇》，安装好之后，将redis复制成6份，注意要将.rdb和.aof后缀的文件删除，如果有的话。

Ruby环境

使用

yum -y install ruby
yum -y install rubygems

安装ruby环境。
网上下载redis-3.0.，执行gem install redis-3.0.安装。

redis配置文件修改

现在已经准备好了，6份干净的redis，如下所示

[root@localhost redis-cluster]# pwd
/usr/local/redis/redis-cluster
[root@localhost redis-cluster]# ll
total 72
drwxr-xr-x 2 root root  4096 Nov  2 00:17 redis1
drwxr-xr-x 2 root root  4096 Nov  2 00:25 redis2
drwxr-xr-x 2 root root  4096 Nov  2 00:25 redis3
drwxr-xr-x 2 root root  4096 Nov  2 00:25 redis4
drwxr-xr-x 2 root root  4096 Nov  2 00:25 redis5
drwxr-xr-x 2 root root  4096 Nov  2 00:25 redis6
-rwxr-xr-x 1 root root 48141 Nov  2 00:16 
[root@localhost redis-cluster]#

PS：注意，这里已经将redis源文件src目录下的文件拷贝过来了。
这个文件是redis集群的管理文件，ruby脚本。

将六个节点的配置文件按照如下进行修改

################################ GENERAL  #####################################

# By default Redis does not run as a daemon. Use 'yes' if you need it.
# Note that Redis will write a pid file in /var/run/ when daemonized.
daemonize yes

# Accept connections on the specified port, default is 6379.
# If port 0 is specified Redis will not listen on a TCP socket.
port *

################################ REDIS CLUSTER  ###############################
#
# ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
# WARNING EXPERIMENTAL: Redis Cluster is considered to be stable code, however
# in order to mark it as "mature" we need to wait for a non trivial percentage
# of users to deploy it in production.
# ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
#
# Normal Redis instances can't be part of a Redis Cluster; only nodes that are
# started as cluster nodes can. In order to start a Redis instance as a
# cluster node enable the cluster support uncommenting the following:
#
cluster-enabled yes

PS：端口号如果是同一台主机的话，必须不同。不同主机可以相同。

PS：我这里是使用一台主机，所以我将六个节点的端口号修改为7001-7006

编写集群启动脚本和停止脚本

启动脚本

cd redis1
./redis-server 
cd ..
cd redis2
./redis-server 
cd ..
cd redis3
./redis-server 
cd ..
cd redis4
./redis-server 
cd ..
cd redis5
./redis-server 
cd ..
cd redis6
./redis-server 
cd ..

停止脚本

./redis1/redis-cli -p 7001 shutdown
./redis1/redis-cli -p 7002 shutdown
./redis1/redis-cli -p 7003 shutdown
./redis1/redis-cli -p 7004 shutdown
./redis1/redis-cli -p 7005 shutdown
./redis1/redis-cli -p 7006 shutdown

PS：两个脚本都放在如下所属目录

[root@localhost redis-cluster]# pwd
/usr/local/redis/redis-cluster
[root@localhost redis-cluster]# ll
total 80
drwxr-xr-x 2 root root  4096 Nov  2 00:52 redis1
drwxr-xr-x 2 root root  4096 Nov  2 00:51 redis2
drwxr-xr-x 2 root root  4096 Nov  2 00:53 redis3
drwxr-xr-x 2 root root  4096 Nov  2 00:53 redis4
drwxr-xr-x 2 root root  4096 Nov  2 00:53 redis5
drwxr-xr-x 2 root root  4096 Nov  2 00:53 redis6
-rwxr-xr-x 1 root root 48141 Nov  2 00:16 
-rw-r--r-- 1 root root   252 Nov  2 00:55 
-rw-r--r-- 1 root root   216 Nov  2 00:57 
[root@localhost redis-cluster]#

修改权限

[root@localhost redis-cluster]# chmod -u+x start-all.sh stop-all.sh

启动节点

[root@localhost redis-cluster]# ./ 
[root@localhost redis-cluster]# ps aux | grep redis
root      2924  0.8  0.1  33932  2048 ?        Ssl  Nov01   3:53 ./redis-server *:6379 [cluster]
root     11924  0.0  0.1  33936  1948 ?        Ssl  01:01   0:00 ./redis-server *:7001 [cluster]
root     11928  0.0  0.1  33936  1952 ?        Ssl  01:01   0:00 ./redis-server *:7002 [cluster]
root     11932  0.0  0.1  33936  1948 ?        Ssl  01:01   0:00 ./redis-server *:7003 [cluster]
root     11936  0.0  0.1  33936  1952 ?        Ssl  01:01   0:00 ./redis-server *:7004 [cluster]
root     11940  0.0  0.1  33936  1952 ?        Ssl  01:01   0:00 ./redis-server *:7005 [cluster]
root     11944  0.0  0.1  33936  1948 ?        Ssl  01:01   0:00 ./redis-server *:7006 [cluster]
root     11948  0.0  0.0   4360   748 pts/2    S+   01:01   0:00 grep redis
[root@localhost redis-cluster]#

执行创建集群命令

[root@localhost redis-cluster]# pwd
/usr/local/redis/redis-cluster
[root@localhost redis-cluster]# ./ create --replicas 1 192.168.37.131:7001 192.168.37.131:7002 192.168.37.131:7003 192.168.37.131:7004 192.168.37.131:7005  192.168.37.131:7006

成功启动信息

>>> Creating cluster
Connecting to node 192.168.37.131:7001: OK
Connecting to node 192.168.37.131:7002: OK
Connecting to node 192.168.37.131:7003: OK
Connecting to node 192.168.37.131:7004: OK
Connecting to node 192.168.37.131:7005: OK
Connecting to node 192.168.37.131:7006: OK
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
192.168.37.131:7001
192.168.37.131:7002
192.168.37.131:7003
Adding replica 192.168.37.131:7004 to 192.168.37.131:7001
Adding replica 192.168.37.131:7005 to 192.168.37.131:7002
Adding replica 192.168.37.131:7006 to 192.168.37.131:7003
M: 8b153503b52f83634e04b0077f32ef629ad91ee6 192.168.37.131:7001
   slots:0-5460 (5461 slots) master
M: f89799066dd8ecaaa1430559be4ce9c8c87055d8 192.168.37.131:7002
   slots:5461-10922 (5462 slots) master
M: 53d698ad56b09f89cfef34850213e2d0a44154dd 192.168.37.131:7003
   slots:10923-16383 (5461 slots) master
S: e73204399d08c14def1f71d0c5377cbc757dc4b8 192.168.37.131:7004
   replicates 8b153503b52f83634e04b0077f32ef629ad91ee6
S: 1d5dcc8d1ccb6bce55efc3e3aadc690dc77808d8 192.168.37.131:7005
   replicates f89799066dd8ecaaa1430559be4ce9c8c87055d8
S: e9458233cb85bd897ff694003e6d8a834eba2b44 192.168.37.131:7006
   replicates 53d698ad56b09f89cfef34850213e2d0a44154dd
Can I set the above configuration? (type 'yes' to accept): y
*** Aborting...
[root@localhost redis-cluster]# ./ create --replicas 1 192.168.37.131:7001 192.168.37.131:7002 192.168.37.131:7003 192.168.37.131:7004 192.168.37.131:7005  192.168.37.131:7006
>>> Creating cluster
Connecting to node 192.168.37.131:7001: OK
Connecting to node 192.168.37.131:7002: OK
Connecting to node 192.168.37.131:7003: OK
Connecting to node 192.168.37.131:7004: OK
Connecting to node 192.168.37.131:7005: OK
Connecting to node 192.168.37.131:7006: OK
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
192.168.37.131:7001
192.168.37.131:7002
192.168.37.131:7003
Adding replica 192.168.37.131:7004 to 192.168.37.131:7001
Adding replica 192.168.37.131:7005 to 192.168.37.131:7002
Adding replica 192.168.37.131:7006 to 192.168.37.131:7003
M: 8b153503b52f83634e04b0077f32ef629ad91ee6 192.168.37.131:7001
   slots:0-5460 (5461 slots) master
M: f89799066dd8ecaaa1430559be4ce9c8c87055d8 192.168.37.131:7002
   slots:5461-10922 (5462 slots) master
M: 53d698ad56b09f89cfef34850213e2d0a44154dd 192.168.37.131:7003
   slots:10923-16383 (5461 slots) master
S: e73204399d08c14def1f71d0c5377cbc757dc4b8 192.168.37.131:7004
   replicates 8b153503b52f83634e04b0077f32ef629ad91ee6
S: 1d5dcc8d1ccb6bce55efc3e3aadc690dc77808d8 192.168.37.131:7005
   replicates f89799066dd8ecaaa1430559be4ce9c8c87055d8
S: e9458233cb85bd897ff694003e6d8a834eba2b44 192.168.37.131:7006
   replicates 53d698ad56b09f89cfef34850213e2d0a44154dd
Can I set the above configuration? (type 'yes' to accept): yes
>>> Nodes configuration updated
>>> Assign a different config epoch to each node
>>> Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join.....
>>> Performing Cluster Check (using node 192.168.37.131:7001)
M: 8b153503b52f83634e04b0077f32ef629ad91ee6 192.168.37.131:7001
   slots:0-5460 (5461 slots) master
M: f89799066dd8ecaaa1430559be4ce9c8c87055d8 192.168.37.131:7002
   slots:5461-10922 (5462 slots) master
M: 53d698ad56b09f89cfef34850213e2d0a44154dd 192.168.37.131:7003
   slots:10923-16383 (5461 slots) master
M: e73204399d08c14def1f71d0c5377cbc757dc4b8 192.168.37.131:7004
   slots: (0 slots) master
   replicates 8b153503b52f83634e04b0077f32ef629ad91ee6
M: 1d5dcc8d1ccb6bce55efc3e3aadc690dc77808d8 192.168.37.131:7005
   slots: (0 slots) master
   replicates f89799066dd8ecaaa1430559be4ce9c8c87055d8
M: e9458233cb85bd897ff694003e6d8a834eba2b44 192.168.37.131:7006
   slots: (0 slots) master
   replicates 53d698ad56b09f89cfef34850213e2d0a44154dd
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
[root@localhost redis-cluster]#

异常

>>> Creating cluster
Connecting to node 192.168.37.131:7001: OK
[ERR] Node 192.168.37.131:7001 is not empty. Either the node already knows other nodes (check with CLUSTER NODES) or contains some key in database 0.

解决方法是删除生成的配置文件，如果不行则说明现在创建的结点包括了旧集群的结点信息，需要删除redis的持久化文件后再重启redis，比如：、

首先，使用停止服务

[root@localhost redis-cluster]# ./
[root@localhost redis-cluster]# ps aux | grep redis
root      2924  0.8  0.1  33932  2048 ?        Ssl  Nov01   3:54 ./redis-server *:6379 [cluster]
root     11969  0.0  0.0   4360   744 pts/2    S+   01:10   0:00 grep redis
[root@localhost redis-cluster]#

然后每个节点中的、、。

[root@localhost redis-cluster]# rm -f redis*/
[root@localhost redis-cluster]# rm -f redis*/
[root@localhost redis-cluster]# rm -f redis*/

然后使用脚本启动，再启动集群管理服务。

[root@localhost redis-cluster]# ./ 
[root@localhost redis-cluster]# ps aux | grep redis
root      2924  0.8  0.1  33932  2048 ?        Ssl  Nov01   3:54 ./redis-server *:6379 [cluster]
root     11980  0.0  0.1  33936  1952 ?        Ssl  01:12   0:00 ./redis-server *:7001 [cluster]
root     11982  0.0  0.1  33936  1952 ?        Ssl  01:12   0:00 ./redis-server *:7002 [cluster]
root     11984  0.0  0.1  33936  1952 ?        Ssl  01:12   0:00 ./redis-server *:7003 [cluster]
root     11986  0.0  0.1  33936  1952 ?        Ssl  01:12   0:00 ./redis-server *:7004 [cluster]
root     11988  0.0  0.1  33936  1948 ?        Ssl  01:12   0:00 ./redis-server *:7005 [cluster]
root     11990  0.0  0.1  33936  1948 ?        Ssl  01:12   0:00 ./redis-server *:7006 [cluster]
root     12004  0.0  0.0   4360   748 pts/2    S+   01:12   0:00 grep redis
[root@localhost redis-cluster]# ./ create --replicas 1 192.168.37.131:7001 192.168.37.131:7002 192.168.37.131:7003 192.168.37.131:7004 192.168.37.131:7005  192.168.37.131:7006
[root@localhost redis-cluster]#

Redis集群节点的操作

查询集群信息

集群创建成功登陆任意redis结点查询集群中的节点情况。

客户端以集群方式登陆：

[root@localhost redis-cluster]# ./redis1/redis-cli -c -h 192.168.37.131 -p 7001
192.168.37.131:7001>

说明：
./redis1/redis-cli -c -h 192.168.37.131 -p 7001 ，其中-c表示以集群方式连接redis，-h指定ip地址，-p指定端口号

cluster nodes 查询集群结点信息

192.168.37.131:7001> cluster nodes
8e1186475f87c928e8a146d3804d9a2697246ad0 192.168.37.131:7003 master - 0 1478085160899 3 connected 10923-16383
80de7003738f74134a3403fc939fed253b7774f2 192.168.37.131:7004 master - 0 1478085156858 8 connected 0-5460
652caf5daf7971135679951324eba7b50e99251a 192.168.37.131:7005 slave 3adba62fdcc331ce231ca580cd2c8701e047bc6d 0 1478085158876 9 connected
1cd6482fd7038d78ad556b52b0cb9e2590ad5598 192.168.37.131:7001 myself,slave 80de7003738f74134a3403fc939fed253b7774f2 0 0 1 connected
3adba62fdcc331ce231ca580cd2c8701e047bc6d 192.168.37.131:7002 master - 0 1478085159889 9 connected 5461-10922
cf23ca6d78cba3d3924065e1f7a394f6c51d4b28 192.168.37.131:7006 slave 8e1186475f87c928e8a146d3804d9a2697246ad0 0 1478085157867 6 connected
192.168.37.131:7001>

cluster info 查询集群状态信息

192.168.37.131:7001> cluster info
cluster_state:ok
cluster_slots_assigned:16384
cluster_slots_ok:16384
cluster_slots_pfail:0
cluster_slots_fail:0
cluster_known_nodes:6
cluster_size:3
cluster_current_epoch:9
cluster_my_epoch:8
cluster_stats_messages_sent:12727
cluster_stats_messages_received:10820
192.168.37.131:7001>

添加主节点

集群创建成功后可以向集群中添加节点，下面是添加一个master主节点。

首先，准备一个干净的redis节点。按上面集群版修改redis配置文件。开启该redis节点。

查看redis进程

[root@localhost redis-cluster]# !ps
ps aux | grep redis
root      2924  0.6  0.1  33932  1800 ?        Ssl  Nov01   4:08 ./redis-server *:6379 [cluster]
root     11980  0.0  0.2  33936  2216 ?        Ssl  01:12   0:09 ./redis-server *:7001 [cluster]
root     11982  0.0  0.2  33936  2244 ?        Ssl  01:12   0:09 ./redis-server *:7002 [cluster]
root     11984  0.0  0.2  33936  2220 ?        Ssl  01:12   0:09 ./redis-server *:7003 [cluster]
root     11986  0.0  0.2  33936  2216 ?        Ssl  01:12   0:09 ./redis-server *:7004 [cluster]
root     11988  0.0  0.2  33936  2228 ?        Ssl  01:12   0:09 ./redis-server *:7005 [cluster]
root     11990  0.0  0.2  33936  2212 ?        Ssl  01:12   0:09 ./redis-server *:7006 [cluster]
root     13913  0.0  0.1  33936  1952 ?        Ssl  04:21   0:00 ./redis7/redis-server *:7007 [cluster]   
root     13917  0.0  0.0   4360   728 pts/3    S+   04:21   0:00 grep redis
[root@localhost redis-cluster]#

执行下边命令：

[root@localhost redis-cluster]# ./ add-node 192.168.37.131:7007 192.168.37.131:7001
>>> Adding node 192.168.37.131:7007 to cluster 192.168.37.131:7001
Connecting to node 192.168.37.131:7001: OK
Connecting to node 192.168.37.131:7003: OK
Connecting to node 192.168.37.131:7004: OK
Connecting to node 192.168.37.131:7005: OK
Connecting to node 192.168.37.131:7002: OK
Connecting to node 192.168.37.131:7006: OK
>>> Performing Cluster Check (using node 192.168.37.131:7001)
S: 1cd6482fd7038d78ad556b52b0cb9e2590ad5598 192.168.37.131:7001
   slots: (0 slots) slave
   replicates 80de7003738f74134a3403fc939fed253b7774f2
M: 8e1186475f87c928e8a146d3804d9a2697246ad0 192.168.37.131:7003
   slots:10923-16383 (5461 slots) master
   1 additional replica(s)
M: 80de7003738f74134a3403fc939fed253b7774f2 192.168.37.131:7004
   slots:0-5460 (5461 slots) master
   1 additional replica(s)
S: 652caf5daf7971135679951324eba7b50e99251a 192.168.37.131:7005
   slots: (0 slots) slave
   replicates 3adba62fdcc331ce231ca580cd2c8701e047bc6d
M: 3adba62fdcc331ce231ca580cd2c8701e047bc6d 192.168.37.131:7002
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
S: cf23ca6d78cba3d3924065e1f7a394f6c51d4b28 192.168.37.131:7006
   slots: (0 slots) slave
   replicates 8e1186475f87c928e8a146d3804d9a2697246ad0
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
Connecting to node 192.168.37.131:7007: OK
>>> Send CLUSTER MEET to node 192.168.37.131:7007 to make it join the cluster.
[OK] New node added correctly.
[root@localhost redis-cluster]#

查看集群结点发现7007已添加到集群中：

[root@localhost redis-cluster]# ./redis1/redis-cli -c -h 192.168.37.131 -p 7001
192.168.37.131:7001> cluster nodes
8e1186475f87c928e8a146d3804d9a2697246ad0 192.168.37.131:7003 master - 0 1478085870038 3 connected 10923-16383
80de7003738f74134a3403fc939fed253b7774f2 192.168.37.131:7004 master - 0 1478085868020 8 connected 0-5460
652caf5daf7971135679951324eba7b50e99251a 192.168.37.131:7005 slave 3adba62fdcc331ce231ca580cd2c8701e047bc6d 0 1478085874075 9 connected
1cd6482fd7038d78ad556b52b0cb9e2590ad5598 192.168.37.131:7001 myself,slave 80de7003738f74134a3403fc939fed253b7774f2 0 0 1 connected
3adba62fdcc331ce231ca580cd2c8701e047bc6d 192.168.37.131:7002 master - 0 1478085873064 9 connected 5461-10922
cf23ca6d78cba3d3924065e1f7a394f6c51d4b28 192.168.37.131:7006 slave 8e1186475f87c928e8a146d3804d9a2697246ad0 0 1478085875086 6 connected
5d6c61ecff23bff3b0fb01a86c66d882f2d402a0 192.168.37.131:7007 master - 0 1478085872056 0 connected
192.168.37.131:7001>

hash槽重新分配

添加完主节点需要对主节点进行hash槽分配这样该主节才可以存储数据。
redis集群有16384个槽，集群中的每个master结点分配一些槽，通过查看集群结点可以看到槽占用情况。

给刚添加的7007结点分配槽：

第一步：连接上集群
./ reshard 192.168.37.131:7001（连接集群中任意一个可用结点都行）

[root@localhost redis-cluster]# ./ reshard 192.168.37.131:7001
Connecting to node 192.168.37.131:7001: OK
Connecting to node 192.168.37.131:7003: OK
Connecting to node 192.168.37.131:7004: OK
Connecting to node 192.168.37.131:7005: OK
Connecting to node 192.168.37.131:7002: OK
Connecting to node 192.168.37.131:7006: OK
Connecting to node 192.168.37.131:7007: OK
>>> Performing Cluster Check (using node 192.168.37.131:7001)
S: 1cd6482fd7038d78ad556b52b0cb9e2590ad5598 192.168.37.131:7001
   slots: (0 slots) slave
   replicates 80de7003738f74134a3403fc939fed253b7774f2
M: 8e1186475f87c928e8a146d3804d9a2697246ad0 192.168.37.131:7003
   slots:10923-16383 (5461 slots) master
   1 additional replica(s)
M: 80de7003738f74134a3403fc939fed253b7774f2 192.168.37.131:7004
   slots:0-5460 (5461 slots) master
   1 additional replica(s)
S: 652caf5daf7971135679951324eba7b50e99251a 192.168.37.131:7005
   slots: (0 slots) slave
   replicates 3adba62fdcc331ce231ca580cd2c8701e047bc6d
M: 3adba62fdcc331ce231ca580cd2c8701e047bc6d 192.168.37.131:7002
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
S: cf23ca6d78cba3d3924065e1f7a394f6c51d4b28 192.168.37.131:7006
   slots: (0 slots) slave
   replicates 8e1186475f87c928e8a146d3804d9a2697246ad0
M: 5d6c61ecff23bff3b0fb01a86c66d882f2d402a0 192.168.37.131:7007
   slots: (0 slots) master
   0 additional replica(s)
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
How many slots do you want to move (from 1 to 16384)?

第二步：输入要分配的槽数量

How many slots do you want to move (from 1 to 16384)? 1000
What is the receiving node ID?

输入1000表示要分配1000个槽

第三步：输入接收槽的结点id

这里准备给7007分配槽，通过cluster nodes查看7007结点id为5d6c61ecff23bff3b0fb01a86c66d882f2d402a0

输入：5d6c61ecff23bff3b0fb01a86c66d882f2d402a0

What is the receiving node ID? 5d6c61ecff23bff3b0fb01a86c66d882f2d402a0
Please enter all the source node IDs.
  Type 'all' to use all the nodes as source nodes for the hash slots.
  Type 'done' once you entered all the source nodes IDs.
Source node #1:

第四步：输入源结点id
输入源结点id，槽将从源结点中拿，分配后的槽在源结点中就不存在了。
输入all表示从所有源结点中获取槽。
输入done取消分配。

这里输入all

Source node #1:all

省略了很多

    Moving slot 11253 from 8e1186475f87c928e8a146d3804d9a2697246ad0
    Moving slot 11254 from 8e1186475f87c928e8a146d3804d9a2697246ad0
    Moving slot 11255 from 8e1186475f87c928e8a146d3804d9a2697246ad0
Do you want to proceed with the proposed reshard plan (yes/no)?

第五步：输入yes开始移动槽到目标结点id

省略了很多

Moving slot 11253 from 192.168.37.131:7003 to 192.168.37.131:7007: 
Moving slot 11254 from 192.168.37.131:7003 to 192.168.37.131:7007: 
Moving slot 11255 from 192.168.37.131:7003 to 192.168.37.131:7007: 
[root@localhost redis-cluster]#

第六步：查看结点信息

[root@localhost redis-cluster]# ./redis1/redis-cli -c -h 192.168.37.131 -p 7001
192.168.37.131:7001> cluster nodes
8e1186475f87c928e8a146d3804d9a2697246ad0 192.168.37.131:7003 master - 0 1478086754466 3 connected 11256-16383
80de7003738f74134a3403fc939fed253b7774f2 192.168.37.131:7004 master - 0 1478086758509 8 connected 333-5460
652caf5daf7971135679951324eba7b50e99251a 192.168.37.131:7005 slave 3adba62fdcc331ce231ca580cd2c8701e047bc6d 0 1478086756490 9 connected
1cd6482fd7038d78ad556b52b0cb9e2590ad5598 192.168.37.131:7001 myself,slave 80de7003738f74134a3403fc939fed253b7774f2 0 0 1 connected
3adba62fdcc331ce231ca580cd2c8701e047bc6d 192.168.37.131:7002 master - 0 1478086757500 9 connected 5795-10922
cf23ca6d78cba3d3924065e1f7a394f6c51d4b28 192.168.37.131:7006 slave 8e1186475f87c928e8a146d3804d9a2697246ad0 0 1478086755477 6 connected
5d6c61ecff23bff3b0fb01a86c66d882f2d402a0 192.168.37.131:7007 master - 0 1478086759518 10 connected 0-332 5461-5794 10923-11255
192.168.37.131:7001>

查看结点信息，可以发现7007结点已经从三个主节点中获取了slot。

添加从节点

集群创建成功后可以向集群中添加节点，下面是添加一个slave从节点。
添加7008从结点，将7008作为7007的从结点。

首先，准备一个干净的redis节点。按上面集群版修改redis配置文件。开启该redis节点。

查看redis进程

[root@localhost redis-cluster]# ps aux | grep redis
root      2924  0.5  0.1  33932  1800 ?        Ssl  Nov01   4:12 ./redis-server *:6379 [cluster]
root     11980  0.0  0.2  33936  2308 ?        Ssl  01:12   0:13 ./redis-server *:7001 [cluster]
root     11982  0.0  0.2  33936  2288 ?        Ssl  01:12   0:13 ./redis-server *:7002 [cluster]
root     11984  0.0  0.2  33936  2236 ?        Ssl  01:12   0:13 ./redis-server *:7003 [cluster]
root     11986  0.0  0.2  33936  2288 ?        Ssl  01:12   0:13 ./redis-server *:7004 [cluster]
root     11988  0.0  0.2  33936  2248 ?        Ssl  01:12   0:13 ./redis-server *:7005 [cluster]
root     11990  0.0  0.2  33936  2244 ?        Ssl  01:12   0:13 ./redis-server *:7006 [cluster]
root     13913  0.1  0.2  33936  2092 ?        Ssl  04:21   0:04 ./redis7/redis-server *:7007 [cluster]   
root     14000  0.1  0.1  33936  1948 ?        Ssl  05:24   0:00 ./redis-server *:7008 [cluster]
root     14006  0.0  0.0   4360   732 pts/3    S+   05:24   0:00 grep redis
[root@localhost redis-cluster]#

命令格式为：

./redis-trib.rb add-node --slave --master-id 主节点id 添加节点的ip和端口 集群中已存在节点ip和端口

执行如下命令：

[root@localhost redis-cluster]# ./ add-node --slave --master-id 5d6c61ecff23bff3b0fb01a86c66d882f2d402a0 192.168.37.131:7008 192.168.37.131:7001
>>> Adding node 192.168.37.131:7008 to cluster 192.168.37.131:7001
Connecting to node 192.168.37.131:7001: OK
Connecting to node 192.168.37.131:7003: OK
Connecting to node 192.168.37.131:7004: OK
Connecting to node 192.168.37.131:7005: OK
Connecting to node 192.168.37.131:7002: OK
Connecting to node 192.168.37.131:7006: OK
Connecting to node 192.168.37.131:7007: OK
>>> Performing Cluster Check (using node 192.168.37.131:7001)
S: 1cd6482fd7038d78ad556b52b0cb9e2590ad5598 192.168.37.131:7001
   slots: (0 slots) slave
   replicates 80de7003738f74134a3403fc939fed253b7774f2
M: 8e1186475f87c928e8a146d3804d9a2697246ad0 192.168.37.131:7003
   slots:11256-16383 (5128 slots) master
   1 additional replica(s)
M: 80de7003738f74134a3403fc939fed253b7774f2 192.168.37.131:7004
   slots:333-5460 (5128 slots) master
   1 additional replica(s)
S: 652caf5daf7971135679951324eba7b50e99251a 192.168.37.131:7005
   slots: (0 slots) slave
   replicates 3adba62fdcc331ce231ca580cd2c8701e047bc6d
M: 3adba62fdcc331ce231ca580cd2c8701e047bc6d 192.168.37.131:7002
   slots:5795-10922 (5128 slots) master
   1 additional replica(s)
S: cf23ca6d78cba3d3924065e1f7a394f6c51d4b28 192.168.37.131:7006
   slots: (0 slots) slave
   replicates 8e1186475f87c928e8a146d3804d9a2697246ad0
M: 5d6c61ecff23bff3b0fb01a86c66d882f2d402a0 192.168.37.131:7007
   slots:0-332,5461-5794,10923-11255 (1000 slots) master
   0 additional replica(s)
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
Connecting to node 192.168.37.131:7008: OK
>>> Send CLUSTER MEET to node 192.168.37.131:7008 to make it join the cluster.
Waiting for the cluster to join.
>>> Configure node as replica of 192.168.37.131:7007.
[OK] New node added correctly.
[root@localhost redis-cluster]#

5d6c61ecff23bff3b0fb01a86c66d882f2d402a0是7007结点的id，可以通过cluster nodes查看。

查看集群中的结点，刚添加的7008为7007的从节点：

[root@localhost redis-cluster]# ./redis1/redis-cli -p 7001 -c                               127.0.0.1:7001> cluster nodes
8e1186475f87c928e8a146d3804d9a2697246ad0 192.168.37.131:7003 master - 0 1478089964769 3 connected 11256-16383
80de7003738f74134a3403fc939fed253b7774f2 192.168.37.131:7004 master - 0 1478089966584 8 connected 333-5460
652caf5daf7971135679951324eba7b50e99251a 192.168.37.131:7005 slave 3adba62fdcc331ce231ca580cd2c8701e047bc6d 0 1478089963748 9 connected
1cd6482fd7038d78ad556b52b0cb9e2590ad5598 192.168.37.131:7001 myself,slave 80de7003738f74134a3403fc939fed253b7774f2 0 0 1 connected
3adba62fdcc331ce231ca580cd2c8701e047bc6d 192.168.37.131:7002 master - 0 1478089966787 9 connected 5795-10922
cf23ca6d78cba3d3924065e1f7a394f6c51d4b28 192.168.37.131:7006 slave 8e1186475f87c928e8a146d3804d9a2697246ad0 0 1478089962729 6 connected
5c97e8eab019c40ea3df4925c7400fe7df1846bb 192.168.37.131:7008 slave 5d6c61ecff23bff3b0fb01a86c66d882f2d402a0 0 1478089961713 10 connected
5d6c61ecff23bff3b0fb01a86c66d882f2d402a0 192.168.37.131:7007 master - 0 1478089965777 10 connected 0-332 5461-5794 10923-11255
127.0.0.1:7001>

删除结点

./redis-trib.rb del-node 192.168.37.131:7007 5d6c61ecff23bff3b0fb01a86c66d882f2d402a0

删除已经占有hash槽的结点会失败，报错如下：

[ERR] Node 192.168.37.131:7007 is not empty! Reshard data away and try again.

需要将该结点占用的hash槽分配出去，请参考《hash槽重新分配》这段内容。

客户端对Redis集群的使用方法

使用redis命令行客户端连接

[root@localhost redis-cluster]# ./redis1/redis-cli -p 7001 -c
127.0.0.1:7001> get a
-> Redirected to slot [15495] located at 192.168.37.131:7003
(nil)
192.168.37.131:7003>

一定要加-c参数,节点之间就可以互相跳转

使用jedis连接

package ;

import ;
import ;

import ;
import ;

/**
 * Jedis集群测试
 * 
 * @author Switch
 * @data 2017年2月11日
 * @version V1.0
 */
public class JedisClusterTest {
    public static void main(String[] args) {
        // 创建并填充节点信息
        Set<HostAndPort> nodes = new HashSet<>();
        (new HostAndPort("192.168.37.131", 7001));
        (new HostAndPort("192.168.37.131", 7002));
        (new HostAndPort("192.168.37.131", 7003));
        (new HostAndPort("192.168.37.131", 7004));
        (new HostAndPort("192.168.37.131", 7005));
        (new HostAndPort("192.168.37.131", 7006));

        // 创建JedisCluster对象
        JedisCluster jedisCluster = new JedisCluster(nodes);

        // 使用jedisCluster操作redis
        String key = "jedisCluster";
        String setResult = (key, "hello redis!");
        (setResult);

        String getResult = (key);
        (getResult);

        // 关闭jedisCluster（程序执行完后才能关闭，内部封装了连接池）
        ();
    }
}

——参考《Redis官方文档》

秒客网