redis集群

首先安装redis:

   make&&make install

   cp src/redis-trib.rb/usr/local/redis/redis-cluster

   cd /usr/local/redis/redis-cluster

   mkdir 7000 7001 7002 7003 7004 7005

把redis配置文件分别拷贝到7000—7005目录下

修改下面内容：

 port  7000   改成每个redis实例自己的端口

bind   本机ip

daemonize yes   设置为守护进程

pidfile /var/run/redis_7000.pid  改成自己相应的名字

cluster-enabled yes   开启集群

cluster-config-file nodes_7000.conf   集群的配置文件

cluster-node-timeout 15000   集群请求超时时间

appendonlyyes    开启aof持久化

 

 

 

根据不同的配置文件，依次启动不同的redis实例

 

安装ruby

去ruby官网下载安装包，然后安装

解压后：

./configure

make

make install

 

然后去下载redis-4.0.1.gem，自己选择版本

https://rubygems.org/gems/redis/网址

 

gem install -l redis-4.0.1.gem

 

若gem指令出错，自己百度，一般是ruby安装过程出错，会有..will not be intalled 这些提示。主要是linux环境有问题，包不全导致。建议别用centos nimi

 

然后执行

./redis-trib.rb  create  --replicas 1  192.168.94.130:7000 192.168.94.130:7001  192.168.94.130:7002 192.168.94.130:7003 192.168.94.130:7004 192.168.94.130:7005

 

 

中间要输入个yes，redis集群采用hash slot来分配，总共16384个slot,计算方法crc16(key)%16384,获取数据时也通过这个算法，所以下面指令会跳端口。

 

 

成功启动后登陆，-h指定bind的ip地址,-c表示连接到集群

redis-cli  -h  192.168.94.130  -p  7000  -c

 

//连接redis端口是7000的实例

[root@localhostredis-cluster]# redis-cli -h 192.168.94.130 -p 7000 -c

192.168.94.130:7000>set set3 sss

-> Redirected toslot [11167] located at 192.168.94.130:7002

OK

192.168.94.130:7002>set set4 bbb

-> Redirected toslot [7032] located at 192.168.94.130:7001

OK

192.168.94.130:7001>set set5 aaa

-> Redirected toslot [2905] located at 192.168.94.130:7000

OK

192.168.94.130:7000>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:6

cluster_my_epoch:1

cluster_stats_messages_ping_sent:481

cluster_stats_messages_pong_sent:450

cluster_stats_messages_sent:931

cluster_stats_messages_ping_received:445

cluster_stats_messages_pong_received:481

cluster_stats_messages_meet_received:5

cluster_stats_messages_received:931

 

 

 

 

//连接redis端口是7002的实例

[xuhaixing@localhost~]$ redis-cli -h 192.168.94.130 -p 7002 -c

192.168.94.130:7002>set set1 aa

-> Redirected toslot [3037] located at 192.168.94.130:7000

OK

192.168.94.130:7000>set set2 bb

-> Redirected toslot [15294] located at 192.168.94.130:7002

OK

192.168.94.130:7002>exit

[xuhaixing@localhost~]$ redis-cli -h 192.168.94.130 -p 7002 -c

192.168.94.130:7002>get set1

-> Redirected toslot [3037] located at 192.168.94.130:7000

"aa"

192.168.94.130:7000>get set2

-> Redirected toslot [15294] located at 192.168.94.130:7002

"bb"

192.168.94.130:7002>get set3

"sss"

192.168.94.130:7002>get set4

-> Redirected toslot [7032] located at 192.168.94.130:7001

"bbb"

192.168.94.130:7001>get set5

-> Redirected toslot [2905] located at 192.168.94.130:7000

"aaa"

192.168.94.130:7000>

展现奇迹的时刻就出现了：如上面代码所示，应用指令时各redis实例自己切换。

 

集群中每个节点都是平等的，每个节点都与其他节点连接，每个节点都保存自己的数据和整个集群的状态。这样就保证了连接集群中任何一个节点就能获取整个集群的数据。集群至少有三个主节点，当存活的主节点小于一半时，集群挂掉。