手动做redis分布式集群

Redis Cluster 分布式集群手动配置  

先停原来的环境，清掉配置，再启动

redis-cli -c -h 192.168.0.151 -p 7001

flushall

quit

ps -ef | grep redis-server | grep -v grep | awk '{print $2}' | xargs kill -9

rm -rf /mysql/data/redis/redis-cluster/nodes-7001/*.aof   #这里没有删除rdb文件，是因为，rdb文件是dump时的文件 每隔一段时间redis服务器程序会自动对数据库做一次遍历，把内存快照写在一个叫做“dump.rdb”的文件里，没删也不影响重新搭建数据库

rm -rf /mysql/data/redis/redis-cluster/nodes-7001/nodes-*

rm -rf /mysql/data/redis/redis-cluster/nodes-7002/*.aof

rm -rf /mysql/data/redis/redis-cluster/nodes-7002/nodes-*

rm -rf /mysql/data/redis/redis-cluster/nodes-7007/*.aof

rm -rf /mysql/data/redis/redis-cluster/nodes-7007/nodes-*

rm -rf /mysql/data/redis/redis-cluster/nodes-7003/*.aof

rm -rf /mysql/data/redis/redis-cluster/nodes-7003/nodes-*

rm -rf /mysql/data/redis/redis-cluster/nodes-7004/*.aof

rm -rf /mysql/data/redis/redis-cluster/nodes-7004/nodes-*

rm -rf /mysql/data/redis/redis-cluster/nodes-7008/*.aof

rm -rf /mysql/data/redis/redis-cluster/nodes-7008/nodes-*

rm -rf /mysql/data/redis/redis-cluster/nodes-7005/*.aof

rm -rf /mysql/data/redis/redis-cluster/nodes-7005/nodes-*

rm -rf /mysql/data/redis/redis-cluster/nodes-7006/*.aof

rm -rf /mysql/data/redis/redis-cluster/nodes-7006/nodes-*

 

 

重新启动后的集群结构：

主：192.168.0.151 7001  /mysql/data/redis/redis-cluster/nodes-7001/redis.conf

备：192.168.0.152 7003  /mysql/data/redis/redis-cluster/nodes-7003/redis.conf

主：192.168.0.153 7005  /mysql/data/redis/redis-cluster/nodes-7005/redis.conf

备：192.168.0.151 7002  /mysql/data/redis/redis-cluster/nodes-7002/redis.conf

主：192.168.0.152 7004  /mysql/data/redis/redis-cluster/nodes-7004/redis.conf

备：192.168.0.153 7006  /mysql/data/redis/redis-cluster/nodes-7005/redis.conf

 

各节点启动后： 使用 cluster meet ip port 与各节点握手， 是集群通信的第一步（关键步骤 1：集群搭建－与各节点握手）

redis-cli -c -h 192.168.0.151 -p 7001

cluster meet 192.168.0.151 7001

cluster meet 192.168.0.151 7002

cluster meet 192.168.0.152 7003

cluster meet 192.168.0.152 7004

cluster meet 192.168.0.153 7005

cluster meet 192.168.0.153 7006

 

握手成功后，使用 cluster nodes 可以看到各节点都可以互相查询到

192.168.0.151:7001> cluster nodes

全是 master节点

 

节点握手成功后，此时集群处理下线状态，所有读写都被禁止

192.168.0.151:7001> set fg 1

(error) CLUSTERDOWN Hash slot not served

使用 cluster info 命令获取集群当前状态

192.168.0.151:7001> cluster info

cluster_state:fail

 

 

redis 集群有 16384 个哈希槽，要把所有数据映射到 16384 槽，需要批量设置槽（关键步骤 2： 集群搭建－分配槽）

主：

redis-cli -h 192.168.0.151 -p 7001 cluster addslots {0..5461}

redis-cli -h 192.168.0.153 -p 7005 cluster addslots {5462..10922}

redis-cli -h 192.168.0.152 -p 7004 cluster addslots {10923..16383}

 

做主从映射(关键步骤 3：集群搭建－登录从节点，完成对主节点（下面写IP的点）的映射)，到此 redis 集群手动搭建完成

redis-cli -h 192.168.0.152 -p 7003

cluster replicate 08a9405c91c03bb6436f435f00e975aac58b8a98

quit

redis-cli -h 192.168.0.151 -p 7002

cluster replicate 438c70b8ddb309e782ff983cc594db5ae655f08c

quit

redis-cli -h 192.168.0.153 -p 7006

cluster replicate a4fb9a0ed113b2668b5cffbc51edd5c873a90d0e

 

192.168.0.151:7001>

cluster nodes

192.168.0.151:7001>

cluster info

cluster_state:ok

cluster_slots_assigned:16384

 

 

 -c 意思是连接到集群里面去

[redis@itpuxdb01 nodes-7002]$ redis-cli -c -h 192.168.0.153 -p 7006

192.168.0.153:7006> set key1 itpux1

-> Redirected to slot [9189] located at 192.168.0.153:7005

OK

192.168.0.153:7005> set key2 itpux2

-> Redirected to slot [4998] located at 192.168.0.151:7001

OK