centos redis集群搭建

说明：

10.0.0.111部署6500，6501，6502三个主节点

10.0.0.222部署6500，6501，6502三个备份节点

1.安装redis:略

2.配置内核参数

# 配置 vm.overcommit_memory 为1，这可以避免数据被截断
sysctl -w vm.overcommit_memory=1

3.建文件夹

cd /usr/local/
mkdir redis-cluster
cd redis-cluster
mkdir 6500 6501 6502
cd /var/log/
mkdir redis

4.修改配置文件

cd /etc/redis/
cp redis.conf redis_bak.conf
vim /etc/redis/redis.conf

bind 192.168.2.247（需要不同服务器的节点连通，就不能设置为 127.0.0.1）
protected-mode no（需要不同服务器的节点连通，这个就要设置为 no）
daemonize yes（设置后台运行redis）
cluster-enabled yes
cluster-node-timeout 5000
appendonly yes

5.拷贝文件到不同目录，并做如下修改

cp redis.conf /usr/local/redis-cluster/6500

port 6500
pidfile /var/run/redis_6500.pid
logfile /var/log/redis/redis_6500.log
dbfilename dump_6500.rdb
appendfilename "appendonly_6500.aof"
cluster-config-file nodes_6500.conf

6.启动redis

7.测试远程连接

10.0.0.111跟10.0.0.222均进行1-8操作

8.创建集群

cd /root/redis-4.0.6/src/

./redis-trib.rb create --replicas 1 10.0.0.111:6500 10.0.0.111:6501 10.0.0.111:6502 10.0.0.222:6500 10.0.0.222:6501 10.0.0.222:6502 

yes(不能直接回车或者输入y,必须输yes)

9.连接测试

可连接任意节点，集群会重定向处理，连接需加参数-c(cluster首字母？)

redis-cli -h 10.0.0.222 -c -p 6500

10.模拟主节点挂掉

关掉主机的三个进程，连接备用节点，发现仍然可以取到a的值，因为在前面设了。

连接任意一个节点都能取到所有节点的数据（前提：ip+端口通，可以做个连接池，在一个ip+端口不通时切换到另一个）

11.重新开启111上的6500，6501，6502结点并加入集群(此处为新增节点的操作，如节点挂掉，只需重新启动进程就好，无需此操作，暂时不删，请忽略)

依此把三个节点加入集群：

语法 ./redis-trib.rb  add-node    new_host:new_port existing_host:existing_port


cd /root/redis-4.0.6/src/
./redis-trib.rb add-node 10.0.0.111:6500 10.0.0.222:6500


运行结果

>>> Adding node 10.0.0.111:6500 to cluster 10.0.0.222:6500
>>> Performing Cluster Check (using node 10.0.0.222:6500)
M: 16ed34de7815c7c13e5263a03685082fda1783a8 10.0.0.222:6500
slots:5461-10922 (5462 slots) master
0 additional replica(s)
M: a0f0745883e3eaec9e469db66ced83e9ce114629 10.0.0.222:6502
slots:10923-16383 (5461 slots) master
0 additional replica(s)
M: 458741ac8b916c95c8093573d0718af07da19832 10.0.0.222:6501
slots:0-5460 (5461 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.
>>> Send CLUSTER MEET to node 10.0.0.111:6500 to make it join the cluster.
[OK] New node added correctly.

连接111的6500，取上一步设的bbb，可以取到值，并且设置ccc的值

继续加入6501，6502

./redis-trib.rb add-node 10.0.0.111:6501 10.0.0.222:6501
./redis-trib.rb add-node 10.0.0.111:6502 10.0.0.222:6502

连接111的6501，取上一步设置的ccc值，可以取到

12.重新分配哈希槽测试（不然节点虽然能存取数据，但都是转向到其他节点做的）

E:\Program Files\Redis>redis-cli -h 10.0.0.111 -c -p 6500 cluster nodes
d64d42cc090440053eb464539f8a8bf0a93543d6 10.0.0.222:6502@16502 slave 21f1919531e456bdff078ab7351ed4e900287f9c 0 1534490272266 6 connected
21f1919531e456bdff078ab7351ed4e900287f9c 10.0.0.111:6501@16501 master - 0 1534490273000 2 connected 10923-16383
a6990cc5fa94c8e713bcc44536bf87d2e889dce8 10.0.0.222:6500@16500 master - 0 1534490273000 4 connected 5461-10922
9049ba4a344dbd1e75ee62f40e12dc07881f0a37 10.0.0.111:6500@16500 myself,master - 0 1534490273000 1 connected 0-5460
246b6da627b00c8dbc5e513592c227af6185a765 10.0.0.111:6502@16502 slave a6990cc5fa94c8e713bcc44536bf87d2e889dce8 0 1534490274269 4 connected
0957bbd8905c9d5a50d4545571304e175bc8f892 10.0.0.222:6501@16501 slave 9049ba4a344dbd1e75ee62f40e12dc07881f0a37 0 1534490273267 5 connected

找到master节点id: a6990cc5fa94c8e713bcc44536bf87d2e889dce8


redis-cli -h 10.0.0.111 -c -p 6501 cluster nodes
./redis-trib.rb reshard 10.0.0.111:6500
输入你要分配的数量 4000
填入id: a6990cc5fa94c8e713bcc44536bf87d2e889dce8
输入：all
输入：yes

13.集群重建测试

1.关掉所有节点的进程：pkill -9 redis

2.删掉相关文件(111,222均要执行)

cd /usr/local/redis-cluster/
rm *.rdb
rm *.conf

cd /root
rm *.rdb
rm *.conf

3.开启所有节点进程

[root@centos7 redis]# redis-server /usr/local/redis-cluster/6500/redis.conf
[root@centos7 redis]# redis-server /usr/local/redis-cluster/6501/redis.conf
[root@centos7 redis]# redis-server /usr/local/redis-cluster/6502/redis.conf

4.重新创建集群（步骤8）

不知道为什么没有按顺序前三个成为master节点。。

5.看看集群情况：

E:\Program Files\Redis>redis-cli -h 10.0.0.111 -c -p 6500 cluster nodes
d64d42cc090440053eb464539f8a8bf0a93543d6 10.0.0.222:6502@16502 slave 21f1919531e456bdff078ab7351ed4e900287f9c 0 1534490272266 6 connected
21f1919531e456bdff078ab7351ed4e900287f9c 10.0.0.111:6501@16501 master - 0 1534490273000 2 connected 10923-16383
a6990cc5fa94c8e713bcc44536bf87d2e889dce8 10.0.0.222:6500@16500 master - 0 1534490273000 4 connected 5461-10922
9049ba4a344dbd1e75ee62f40e12dc07881f0a37 10.0.0.111:6500@16500 myself,master - 0 1534490273000 1 connected 0-5460
246b6da627b00c8dbc5e513592c227af6185a765 10.0.0.111:6502@16502 slave a6990cc5fa94c8e713bcc44536bf87d2e889dce8 0 1534490274269 4 connected
0957bbd8905c9d5a50d4545571304e175bc8f892 10.0.0.222:6501@16501 slave 9049ba4a344dbd1e75ee62f40e12dc07881f0a37 0 1534490273267 5 connected