Redis集群搭建
一、搭建集群
Redis Cluster集群要求必须有3个master才构成一个集群，所以我这里采用3主3从的方式，实际生产中的高可用至少3主6从
（1）使用docker-compose启动6个redis容器，使用 host network，不然jedis内部重定向ip地址不正确。
version: '3.7'
services:
redis7001:
image: 'redis:6.2.4'
container_name: redis7001
network_mode: "host"
command:
["redis-server","/etc/redis/redis.conf"]
volumes:
- /opt/docker/redis/7001/redis.conf:/etc/redis/redis.conf
- /opt/docker/redis/7001/data:/data
ports:
- "7001:7001"
- "17001:17001"
environment:
- TZ=Asia/Shanghai
redis7002:
image: 'redis:6.2.4'
container_name: redis7002
network_mode: "host"
command:
["redis-server", "/etc/redis/redis.conf"]
volumes:
- /opt/docker/redis/7002/redis.conf:/etc/redis/redis.conf
- /opt/docker/redis/7002/data:/data
ports:
- "7002:7002"
- "17002:17002"
environment:
- TZ=Asia/Shanghai
redis7003:
image: 'redis:6.2.4'
container_name: redis7003
network_mode: "host"
command:
["redis-server", "/etc/redis/redis.conf"]
volumes:
- /opt/docker/redis/7003/redis.conf:/etc/redis/redis.conf
- /opt/docker/redis/7003/data:/data
ports:
- "7003:7003"
- "17003:17003"
environment:
- TZ=Asia/Shanghai
redis7004:
image: 'redis:6.2.4'
container_name: redis7004
network_mode: "host"
command:
["redis-server", "/etc/redis/redis.conf"]
volumes:
- /opt/docker/redis/7004/redis.conf:/etc/redis/redis.conf
- /opt/docker/redis/7004/data:/data
ports:
- "7004:7004"
- "17004:17004"
environment:
- TZ=Asia/Shanghai
redis7005:
image: 'redis:6.2.4'
container_name: redis7005
network_mode: "host"
command:
["redis-server", "/etc/redis/redis.conf"]
volumes:
- /opt/docker/redis/7005/redis.conf:/etc/redis/redis.conf
- /opt/docker/redis/7005/data:/data
ports:
- "7005:7005"
- "17005:17005"
environment:
- TZ=Asia/Shanghai
redis7006:
image: 'redis:6.2.4'
container_name: redis7006
network_mode: "host"
command:
["redis-server", "/etc/redis/redis.conf"]
volumes:
- /opt/docker/redis/7006/redis.conf:/etc/redis/redis.conf
- /opt/docker/redis/7006/data:/data
ports:
- "7006:7006"
- "17006:17006"
environment:
- TZ=Asia/Shanghai
[root@shang redis]# docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
e3018a083b37 redis:6.2.4 "docker-entrypoint.s…" 13 seconds ago Up 4 seconds 0.0.0.0:7003->7003/tcp, :::7003->7003/tcp, 6379/tcp, 0.0.0.0:17003->17003/tcp, :::17003->17003/tcp redis7003
911dc36fe35a redis:6.2.4 "docker-entrypoint.s…" 13 seconds ago Up 4 seconds 0.0.0.0:7005->7005/tcp, :::7005->7005/tcp, 6379/tcp, 0.0.0.0:17005->17005/tcp, :::17005->17005/tcp redis7005
efabccd7769c redis:6.2.4 "docker-entrypoint.s…" 13 seconds ago Up 6 seconds 0.0.0.0:7004->7004/tcp, :::7004->7004/tcp, 6379/tcp, 0.0.0.0:17004->17004/tcp, :::17004->17004/tcp redis7004
db48c12796cb redis:6.2.4 "docker-entrypoint.s…" 13 seconds ago Up 4 seconds 0.0.0.0:7002->7002/tcp, :::7002->7002/tcp, 6379/tcp, 0.0.0.0:17002->17002/tcp, :::17002->17002/tcp redis7002
c717621d2c7b redis:6.2.4 "docker-entrypoint.s…" 13 seconds ago Up 4 seconds 0.0.0.0:7006->7006/tcp, :::7006->7006/tcp, 6379/tcp, 0.0.0.0:17006->17006/tcp, :::17006->17006/tcp redis7006
10fd3e24f10f redis:6.2.4 "docker-entrypoint.s…" 13 seconds ago Up 4 seconds 0.0.0.0:7001->7001/tcp, :::7001->7001/tcp, 6379/tcp, 0.0.0.0:17001->17001/tcp, :::17001->17001/tcp redis7001
（2）查看redis集群 [cluster]表示为集群，但此时它们之间并没有建立联系，也没有分配slot。
[root@shang 7001]# ps -ef | grep redis
polkitd 3888 3867 0 09:10 ? 00:00:02 redis-server *:7001 [cluster]
polkitd 4143 4120 0 09:14 ? 00:00:01 redis-server *:7002 [cluster]
polkitd 4257 4235 0 09:17 ? 00:00:01 redis-server *:7003 [cluster]
polkitd 4373 4352 0 09:20 ? 00:00:00 redis-server *:7004 [cluster]
polkitd 4490 4469 0 09:22 ? 00:00:00 redis-server *:7005 [cluster]
polkitd 4604 4581 0 09:23 ? 00:00:00 redis-server *:7006 [cluster]
root 4667 1576 0 09:24 pts/0 00:00:00 grep --color=auto redis
[root@shang 7001]# docker exec -it redis7001 redis-cli -p 7001 # 进入端口为7001的redis7001容器
127.0.0.1:7001> cluster nodes # 查看集群节点信息
87346232e7666c772d304518fa4215804be3f576 :7001@17001 myself,master - 0 0 0 connected
（3）建立联系,并分配slot
# 查看帮助命令
[root@shang 7001]# docker exec -it redis7001 redis-cli --cluster help
# meet
create host1:port1 ... hostN:portN
--cluster-replicas <arg>
docker exec -it redis7001 redis-cli --cluster create 192.168.16.3:7001 192.168.16.3:7002 192.168.16.3:7003 192.168.16.3:7004 192.168.16.3:7005 192.168.16.3:7006 --cluster-replicas 1
建立cluster联系，按照 1:1 的方式（3:3），前三个为master，后三个为slave。
如果--cluster-replicas 2，表示按照 1:2 的方式（3:6)，前三个为master，后三个为slave
（4）查看是否搭建成功
[root@shang 7001]# docker exec -it redis7001 redis-cli -p 7001 # 进入端口为7001的redis7001容器
127.0.0.1:7001> cluster nodes # 查看集群节点信息
二、集群扩容
docker exec -it redis7001 redis-cli --cluster help # 查看帮助命令
add-node new_host:new_port existing_host:existing_port # 新redis 集群中任一redis
--cluster-slave # 标志为从机
--cluster-master-id <arg> # 指定从属于哪个主机，master-id
# 扩容一个slave
docker exec -it redis7001 redis-cli --cluster add-node 192.168.16.3:7007 192.168.16.3:7001 --cluster-slave --cluster-master-id 87346232e7666c772d304518fa4215804be3f576
# 扩容一个master，并分配slot
docker exec -it redis7001 redis-cli --cluster add-node 192.168.16.3:7008 192.168.16.3:7001
reshard host:port # 集群任一节点
--cluster-from <arg> # 带有slot的master节点id，多个以 , 隔开
--cluster-to <arg> # 分配给哪个master节点id，多个以 , 隔开
--cluster-slots <arg> # 总共分配多少slot
--cluster-yes # 跳过确认
--cluster-timeout <arg> #
--cluster-pipeline <arg> # 控制每次批量迁移键的数量，默认为10
--cluster-replace
docker exec -it redis7001 redis-cli --cluster reshard 192.168.16.3:7007 --cluster-from 87346232e7666c772d304518fa4215804be3f576,4979d1b009f30576c4ba7587c01176275cf81628 --cluster-to 90afcac7725cac92e9b064f3a71348b9913c7eb3 --cluster-slots 1000 --cluster-yes
三、集群缩容
先移除某一个master节点的所有slave，避免选举为master。
移除master之前，要将master上的slot分配给其它存活的master，然后再删除当前master，避免数据丢失。
# 删除slave
del-node host:port node_id # 任一集群节点 需要删除的节点id
docker exec -it redis7001 redis-cli --cluster del-node 192.168.16.3:7001 3a29bd1d4e885cdfe5e03185183b88da2cfd5115
# 分离slot
docker exec -it redis7001 redis-cli --cluster reshard 192.168.16.3:7001 --cluster-from 90afcac7725cac92e9b064f3a71348b9913c7eb3 --cluster-to 05abce1703ec1aca5e38548ca2c252c4713c5840 --cluster-slots 1000 --cluster-yes
docker exec -it redis7001 redis-cli --cluster del-node 192.168.16.3:7001 90afcac7725cac92e9b064f3a71348b9913c7eb3 # 移除master 7008
四、记录点
由于宿主机器（192.168.xx.xx）和Docker容器（170.17.0.x）的网络不同，导致Jedis操作Redis集群无法获取连接。根本原因在于JedisFactory中的makeObject中的HostAndPort被修改为docker网络，而不是我们宿主机器的外网ip，所以无法创建连接，获取失败。
