[集群自动化三]三结点Redis集群

Redis 5.0 Cluster in CoreOS

---

基本结构

	前端 : VIP:6379  haproxy_stats:9091
Node1	Node2	Node3
master1:6380	master2:6380	master3:6380
slave2:6381	slave3:6381	slave1:6381
phpredmin:8080	phpredmin:8080	phpredmin:8080
redisadmin:8000
分片slots(0-5461)	slots(5462-10922)	slots(10923-16383)

Download

---

• ha.tgz

• redis.tgz

• redismon.tgz

• git clone https://github.com/Thomas-YangHT/redisHa.git;cp *tgz redisHa; cd redisHa

vim CONFIG

---

NODE1_NAME=node1
NODE2_NAME=node2
NODE3_NAME=node3
NODE1_IP=192.168.253.41
NODE2_IP=192.168.253.42
NODE3_IP=192.168.253.43
VIP_IP=192.168.253.40
VR_ID=53
VPORT=6379
MasterPort=6380
SlavePort=6381
REMOTE_USER=core
masterauth=123321
requirepass=123321 
NODES=$NODE1_IP,$NODE2_IP,$NODE3_IP#for kvm clone shKVM_NAME=$NODE1_NAME,$NODE2_NAME,$NODE3_NAMEGATEWAY=192.168.253.125  
DNS1=192.168.253.110
DNS2=114.114.114.114
PJ=redis

INSTALL

---

sh -x install.sh all and waiting for a moment...
see video

install video

Check STATUS

sh install.sh status  and see all nodes connected

cluster info & nodes status

Check Result by browse svc-redis.html

---

svc-redis.html

phpredmin带统计和命令终端

RedisAdmin支持多主机管理

haproxy_stats

More about install.sh:

usage: install.sh [OPTION]OPTIONS:
        p|prepare      :prepare redis.tgz .etc.
        p1|prepareMon  :prepare redismon.tgz
        p2|haprepare   :prepare ha.tgz.
        keepalived     :install keepalived.
        haproxy        :install haproxy.
        status         :get redis status.
        reset          :when reinstall, delete all.
        start          :start all redis container.
        redis          :start redis container.
        redisMeet      :handshake nodes.
        addSlots       :seprate 16384 slots to three nodes
        nodeInfo       :get nodeInfo
        addSlave       :add slave to master
        redisAdmin     :install phpredisadmin on node1
        phpRedmin      :install phpredmin on all nodes
        repaireRedis   :fix aof file & restart failed redis
        all            :install all components.
        reboot         :reboot all nodes in CEPH cluster.

Q&A

• 集群模式的限制

  • Redis集群模式只支持一个库，即db0

  • 分片存储，Master1(0-5461slots) Master2(5462-10922) Master3(10923-16383)，keys *将只返回连接结点的KEY

  • redis-cli -c 需要加-c 选项，其它常用选项：-p $MasterPort -a $requirepass -h $node[1-3]

  • 可通过VIP：VPORT，或任一NODEIP：MasterPort访问集群

  • 集群通讯是通过通讯端口（数据端口+10000）互连

作者：LinuxMan_yang
链接：https://www.jianshu.com/p/ee83230a4cfb