Redis集群搭建

目录

• 一、Redis Cluster 分布式集群
  • 1.什么是Redis Cluster
  • 2.Redis Cluster的特点
  • 3.槽的概念
  • 4.redis故障转移
• 二、redis集群搭建（方法一）
  • 须知
• 三、使用工具搭建redis集群（方法二）

一、Redis Cluster 分布式集群

1.什么是Redis Cluster

1.Redis集群是一个可以在多个Redis节点之间进行数据共享的设施（installation）
2.Redis集群不支持那些需要同时处理多个键的Redis命令，因为执行这些命令需要在多个Redis节点之间移动数据，并且在高负载的情况下，这些命令将降低Redis集群的性能，并导致不可预测的行为。（使用ack协议）
3.Redis集群通过分区（partition）来提供一定程度的可用性（availability）：即使集群中有一部分节点失效或者无法进行通讯，集群也可以继续处理命令请求。
4.Redis集群有将数据自动切分（split）到多个节点的能力。

2.Redis Cluster的特点

#高性能:
1.在多分片节点中，将16384个槽位，均匀分布到多个分片节点中
2.存数据时，将key做crc16（key），然后和16384进行取模，得出槽位值（0-16384之间）
3.根据计算得出的槽位值，找到相对应的分片节点的主节点，存储到相应槽位上
4.如果客户端当时连接的节点不是将来要存储的分片节点，分片集群会将客户端连接切换至真正存储节点进行数据存储
5.客户端与redis节点直连,不需要中间proxy层.客户端不需要连接集群所有节点,连接集群中任何一个可用节点即可
6.Redis Cluster解决了redis资源利用率的问题

#高可用
7.在搭建集群时，会为每一个分片的主节点，对应一个从节点，实现slaveof功能，同时当主节点down，实现类似于sentinel的自动failover的功能。

3.槽的概念

1.在集群中，会把所有节点分为16384个槽位
2.槽位的序号是 0 - 16383，序号不重要，数量才重要
3.每一个槽位分配到数据的概率是一样

4.redis故障转移

1.在集群里面，节点会对其他节点进行下线检测。
2.当一个主节点下线时，集群里面的其他主节点负责对下线主节点进行故障移。
3.换句话说，集群的节点集成了下线检测和故障转移等类似 Sentinel 的功能。

二、redis集群搭建（方法一）

须知

# 如下只是集群关联的一种方式，还有一种方式需要安装相应工具，但那种方式比较简单，但会自动进行主从，缺点也存在，会自动进行主从，如果不是想要的主从状态，将要自己手动修改。(详情见方法二)

1.环境准备

节点	IP	端口
节点1	172.16.1.51	6379,6380
节点2	172.16.1.52	6379,6380
节点3	172.16.1.53	6379,6380

2.搭建redis

#删除以前的redis数据
[root@db01 ~]# rm -rf /service/redis/*

#创建多实例目录
[root@db01 ~]# mkdir /service/redis/{6379,6380}
[root@db02 ~]# mkdir /service/redis/{6379,6380}
[root@db03 ~]# mkdir /service/redis/{6379,6380}

#配置所有redis
[root@db01 ~]# vim /service/redis/6379/redis.conf
bind 172.16.1.51 127.0.0.1
port 6379
daemonize yes
pidfile /service/redis/6379/redis.pid
loglevel notice
logfile /service/redis/6379/redis.log
dbfilename dump.rdb
dir /service/redis/6379
cluster-enabled yes						# 打开redis集群
cluster-config-file nodes.conf			                # 集群关联文件地址
cluster-node-timeout 5000				        # 节点互联超时阈值

[root@db01 ~]# vim /service/redis/6380/redis.conf
bind 172.16.1.51 127.0.0.1
port 6380
daemonize yes
pidfile /service/redis/6380/redis.pid
loglevel notice
logfile /service/redis/6380/redis.log
dbfilename dump.rdb
dir /service/redis/6380
cluster-enabled yes
cluster-config-file nodes.conf
cluster-node-timeout 5000

3.启动所有redis

[root@db01 ~]# redis-server /service/redis/6379/redis.conf 
[root@db01 ~]# redis-server /service/redis/6380/redis.conf
[root@db02 ~]# redis-server /service/redis/6379/redis.conf 
[root@db02 ~]# redis-server /service/redis/6380/redis.conf
[root@db03 ~]# redis-server /service/redis/6379/redis.conf 
[root@db03 ~]# redis-server /service/redis/6380/redis.conf

4.关联所有redis节点

1）登录所有节点

[root@db01 ~]# redis-cli -h 172.16.1.51 -p 6379
[root@db01 ~]# redis-cli -h 172.16.1.51 -p 6380
[root@db02 ~]# redis-cli -h 172.16.1.52 -p 6379
[root@db02 ~]# redis-cli -h 172.16.1.52 -p 6380
[root@db03 ~]# redis-cli -h 172.16.1.53 -p 6379
[root@db03 ~]# redis-cli -h 172.16.1.53 -p 6380

2）查看集群节点

#查看集群节 79 myself,master - 0 0 0 connected

3）关联所有节点

# 将需要加入集群中的主从服务器全部进行关联
172.16.1.51:6379> CLUSTER MEET 172.16.1.51 6380
OK
172.16.1.51:6379> CLUSTER MEET 172.16.1.52 6379
OK
172.16.1.51:6379> CLUSTER MEET 172.16.1.52 6380
OK
172.16.1.51:6379> CLUSTER MEET 172.16.1.53 6379
OK
172.16.1.51:6379> CLUSTER MEET 172.16.1.53 6380
OK

#查看集群状态，所有节点
172.16.1.51:6379> CLUSTER NODES
aee9f4e6e09a452fd44bca7639be442b5138f141 172.16.1.52:6380 master - 0 1596687131655 4 connected
777412c8d6554e3390e1083bf1f55002be08cf62 172.16.1.51:6380 master - 0 1596687131352 2 connected
ef18ab5bab6d8bc06917a0cf2dc9bffa8b431087 172.16.1.52:6379 master - 0 1596687132362 3 connected
f2747c92813ea06b25c3e9c8d5232b46b3cf9d3d 172.16.1.53:6379 master - 0 1596687131856 0 connected
25f735f08ac62b2f758c1e2c21e178cc46279087 172.16.1.53:6380 master - 0 1596687131251 5 connected
28faba09f4c0ec8cdb90d92e09636796427b7143 172.16.1.51:6379 myself,master - 0 0 1 connected

5.分配槽位

#查看集群状态
172.16.1.51:6379> CLUSTER INFO
cluster_state:fail
cluster_slots_assigned:0
cluster_slots_ok:0
cluster_slots_pfail:0
cluster_slots_fail:0
cluster_known_nodes:6
cluster_size:0
cluster_current_epoch:5
cluster_my_epoch:1
cluster_stats_messages_sent:1168
cluster_stats_messages_received:1168

#槽位规划
db01:     5462 个槽位  （0-5461）
db02:     5461 个槽位  （5462-10922）
db03:	  5461 个槽位  （10923-16383）

#分配槽位
[root@db01 ~]# redis-cli -p 6379 -h 172.16.1.51 CLUSTER ADDSLOTS {0..5461}
OK
[root@db02 ~]# redis-cli -p 6379 -h 172.16.1.52 CLUSTER ADDSLOTS {5462..10922}
OK
[root@db02 ~]# redis-cli -p 6379 -h 172.16.1.53 CLUSTER ADDSLOTS {10923..16383}

6.插入数据测试集群

#插入一条数据
172.16.1.51:6379> set k1 v1
(error) MOVED 12706 172.16.1.53:6379
#报错，该key的值只能插入到12706这个槽位

[root@db03 ~]# redis-cli -h 172.16.1.53
172.16.1.53:6379> set k1 v1
OK

#ASK协议，自动切换将数据添加到指定槽位（加一个-c参数即可）
[root@db03 ~]# redis-cli -h 172.16.1.53
172.16.1.53:6379> set k2 v2
(error) MOVED 449 172.16.1.51:6379
172.16.1.53:6379> quit
[root@db03 ~]# redis-cli -c -h 172.16.1.53
172.16.1.53:6379> set k2 v2
-> Redirected to slot [449] located at 172.16.1.51:6379
OK

#脚本插入数据测试
[root@db03 ~]# vim data.sh 
#!/bin/bash
for i in {1..1000};do
    redis-cli -c -p 6379 -h 172.16.1.51 set k${i} v${i}
done

#查看数据分配
172.16.1.51:6379> DBSIZE
(integer) 341
172.16.1.52:6379> DBSIZE
(integer) 332
172.16.1.53:6379> DBSIZE
(integer) 327

7.添加副本节点

1）查看节点

172.16.1.51:6379> CLUSTER NODES
5a7f0cf95e1850b5b5ae81d873c4c76fd366d604 172.16.1.51:6380 master - 0 1596763193422 4 connected
5eb9e5356534ff4acda736d13f0dc9fc3d40049b 172.16.1.52:6379 master - 0 1596763192412 5 connected 5462-10922
50878ef6a4d8141c8dbca3e2bf7c84ed48a73ee2 172.16.1.53:6380 master - 0 1596763192512 3 connected
acc3a4d0e6e43fc74630c1f0714865fdcbdaf677 172.16.1.53:6379 master - 0 1596763191908 0 connected 10923-16383
2325be6f1f9c1c9f57d5a033fc05e0d798ea823a 172.16.1.51:6379 myself,master - 0 0 1 connected 0-5461
381b54584572e8013becdae2eeaff48bf6eb5450 172.16.1.52:6380 master - 0 1596763193925 2 connected

2）配置主从

#db01的6380做db02的6379的从库
172.16.1.51:6380> CLUSTER REPLICATE 5eb9e5356534ff4acda736d13f0dc9fc3d40049b
OK

#db02的6380做db03的6379的从库
172.16.1.52:6380> CLUSTER REPLICATE acc3a4d0e6e43fc74630c1f0714865fdcbdaf677
OK

#db03的6380做db01的6379的从库
172.16.1.53:6380> CLUSTER REPLICATE 2325be6f1f9c1c9f57d5a033fc05e0d798ea823a
OK

3）再次查看节点信息

172.16.1.51:6379> CLUSTER NODES
5a7f0cf95e1850b5b5ae81d873c4c76fd366d604 172.16.1.51:6380 slave 5eb9e5356534ff4acda736d13f0dc9fc3d40049b 0 1596763362696 5 connected
5eb9e5356534ff4acda736d13f0dc9fc3d40049b 172.16.1.52:6379 master - 0 1596763363202 5 connected 5462-10922
50878ef6a4d8141c8dbca3e2bf7c84ed48a73ee2 172.16.1.53:6380 slave 2325be6f1f9c1c9f57d5a033fc05e0d798ea823a 0 1596763362192 3 connected
acc3a4d0e6e43fc74630c1f0714865fdcbdaf677 172.16.1.53:6379 master - 0 1596763363203 0 connected 10923-16383
2325be6f1f9c1c9f57d5a033fc05e0d798ea823a 172.16.1.51:6379 myself,master - 0 0 1 connected 0-5461
381b54584572e8013becdae2eeaff48bf6eb5450 172.16.1.52:6380 slave acc3a4d0e6e43fc74630c1f0714865fdcbdaf677 0 1596763364211 2 connected

8.故障演示

#停掉一台节点
[root@db03 ~]# reboot

#到另一台机器查看集群状态，发现集群是正常的
172.16.1.51:6379> 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_sent:327031
cluster_stats_messages_received:326973

#查看节点信息，副本被提升为主库
172.16.1.51:6379> CLUSTER NODES
5a7f0cf95e1850b5b5ae81d873c4c76fd366d604 172.16.1.51:6380 slave 5eb9e5356534ff4acda736d13f0dc9fc3d40049b 0 1596763771309 5 connected
5eb9e5356534ff4acda736d13f0dc9fc3d40049b 172.16.1.52:6379 master - 0 1596763771310 5 connected 5462-10922
50878ef6a4d8141c8dbca3e2bf7c84ed48a73ee2 172.16.1.53:6380 slave,fail 2325be6f1f9c1c9f57d5a033fc05e0d798ea823a 1596763736458 1596763734245 3 disconnected
acc3a4d0e6e43fc74630c1f0714865fdcbdaf677 172.16.1.53:6379 master,fail - 1596763736458 1596763735246 0 disconnected
2325be6f1f9c1c9f57d5a033fc05e0d798ea823a 172.16.1.51:6379 myself,master - 0 0 1 connected 0-5461
381b54584572e8013becdae2eeaff48bf6eb5450 172.16.1.52:6380 master - 0 1596763772319 6 connected 10923-16383

9.节点恢复

#修复机器
[root@db03 ~]# redis-server /service/redis/6379/redis.conf 
[root@db03 ~]# redis-server /service/redis/6380/redis.conf

#再次查看节点信息
172.16.1.51:6379> CLUSTER NODES
5a7f0cf95e1850b5b5ae81d873c4c76fd366d604 172.16.1.51:6380 slave 5eb9e5356534ff4acda736d13f0dc9fc3d40049b 0 1596764061287 5 connected
5eb9e5356534ff4acda736d13f0dc9fc3d40049b 172.16.1.52:6379 master - 0 1596764060781 5 connected 5462-10922
50878ef6a4d8141c8dbca3e2bf7c84ed48a73ee2 172.16.1.53:6380 slave 2325be6f1f9c1c9f57d5a033fc05e0d798ea823a 0 1596764059770 3 connected
acc3a4d0e6e43fc74630c1f0714865fdcbdaf677 172.16.1.53:6379 slave 381b54584572e8013becdae2eeaff48bf6eb5450 0 1596764062094 6 connected
2325be6f1f9c1c9f57d5a033fc05e0d798ea823a 172.16.1.51:6379 myself,master - 0 0 1 connected 0-5461
381b54584572e8013becdae2eeaff48bf6eb5450 172.16.1.52:6380 master - 0 1596764061789 6 connected 10923-16383

#原主节点修复后变为从节点

三、使用工具搭建redis集群（方法二）

1.环境准备

节点	IP	端口
节点1	172.16.1.51	6379,6380
节点2	172.16.1.52	6379,6380
节点3	172.16.1.53	6379,6380

2.搭建redis

3.启动所有redis

4.安装集群插件

#EPEL源安装ruby支持
[root@db01 ~]# yum install ruby rubygems -y

#查看gem源
[root@db01 ~]# gem sources -l
*** CURRENT SOURCES ***

http://rubygems.org/

#添加阿里云的gem源
[root@db01 ~]# gem sources -a http://mirrors.aliyun.com/rubygems/
http://mirrors.aliyun.com/rubygems/ added to sources 

#删除国外gem源
[root@db01 ~]# gem sources  --remove https://rubygems.org/
http://rubygems.org/ removed from sources

#再次查看gem源
[root@db01 ~]# gem sources -l

#使用gem安装redis的ruby插件
[root@db01 ~]# gem install redis -v 3.3.3
Successfully installed redis-3.3.3
1 gem installed
Installing ri documentation for redis-3.3.3...
Installing RDoc documentation for redis-3.3.3...

5. redis-trib.rb命令

[root@db01 ~]# redis-trib.rb 
create			#创建一个集群
check			#检查集群
info			#集群状态
fix		        #修复集群
reshard			#重新分配槽位
rebalance		#平衡槽位数量
add-node		#添加节点
del-node		#删除节点
set-timeout		#设置超时时间
call			#向集群所有机器输入命令
import			#导入数据
help                    #帮助

6.关联所有节点

[root@db01 ~]# redis-trib.rb create --replicas 1 172.16.1.51:6379 172.16.1.52:6379 172.16.1.53:6379 172.16.1.52:6380 172.16.1.53:6380 172.16.1.51:6380 
# --replicas参数指定集群中每个主节点配备几个从节点，这里设置为1。
>>> Creating cluster
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
172.16.1.51:6379
172.16.1.52:6379
172.16.1.53:6379
Adding replica 172.16.1.52:6380 to 172.16.1.51:6379
Adding replica 172.16.1.51:6380 to 172.16.1.52:6379
Adding replica 172.16.1.53:6380 to 172.16.1.53:6379
M: 5ad7bd957133eac9c3a692b35f8ae72258cf0ece 172.16.1.51:6379
   slots:0-5460 (5461 slots) master
M: 7c79559b280db9d9c182f3a25c718efe9e934fc7 172.16.1.52:6379
   slots:5461-10922 (5462 slots) master
M: d27553035a3e91c78d375208c72b756e9b2523d4 172.16.1.53:6379
   slots:10923-16383 (5461 slots) master
S: fee551a90c8646839f66fa0cd1f6e5859e9dd8e0 172.16.1.52:6380
   replicates 5ad7bd957133eac9c3a692b35f8ae72258cf0ece
S: e4794215d9d3548e9c514c10626ce618be19ebfb 172.16.1.53:6380
   replicates d27553035a3e91c78d375208c72b756e9b2523d4
S: 1d10edbc5ed08f85d2afc21cd338b023b9dd61b4 172.16.1.51:6380
   replicates 7c79559b280db9d9c182f3a25c718efe9e934fc7
Can I set the above configuration? (type 'yes' to accept): yes   
>>> Nodes configuration updated
>>> Assign a different config epoch to each node
>>> Sending CLUSTER MEET messages to join the cluster
Waiting for the cluster to join...
>>> Performing Cluster Check (using node 172.16.1.51:6379)
M: 5ad7bd957133eac9c3a692b35f8ae72258cf0ece 172.16.1.51:6379
   slots:0-5460 (5461 slots) master
   1 additional replica(s)
S: e4794215d9d3548e9c514c10626ce618be19ebfb 172.16.1.53:6380
   slots: (0 slots) slave
   replicates d27553035a3e91c78d375208c72b756e9b2523d4
M: d27553035a3e91c78d375208c72b756e9b2523d4 172.16.1.53:6379
   slots:10923-16383 (5461 slots) master
   1 additional replica(s)
S: fee551a90c8646839f66fa0cd1f6e5859e9dd8e0 172.16.1.52:6380
   slots: (0 slots) slave
   replicates 5ad7bd957133eac9c3a692b35f8ae72258cf0ece
S: 1d10edbc5ed08f85d2afc21cd338b023b9dd61b4 172.16.1.51:6380
   slots: (0 slots) slave
   replicates 7c79559b280db9d9c182f3a25c718efe9e934fc7
M: 7c79559b280db9d9c182f3a25c718efe9e934fc7 172.16.1.52:6379
   slots:5461-10922 (5462 slots) master
   1 additional replica(s)
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.

7.查看集群状态

[root@db01 ~]# redis-cli -h 172.16.1.51 -p 6379 CLUSTER NODES
e4794215d9d3548e9c514c10626ce618be19ebfb 172.16.1.53:6380 slave d27553035a3e91c78d375208c72b756e9b2523d4 0 1596767315453 5 connected
d27553035a3e91c78d375208c72b756e9b2523d4 172.16.1.53:6379 master - 0 1596767315453 3 connected 10923-16383
5ad7bd957133eac9c3a692b35f8ae72258cf0ece 172.16.1.51:6379 myself,master - 0 0 1 connected 0-5460
fee551a90c8646839f66fa0cd1f6e5859e9dd8e0 172.16.1.52:6380 slave 5ad7bd957133eac9c3a692b35f8ae72258cf0ece 0 1596767313429 4 connected
1d10edbc5ed08f85d2afc21cd338b023b9dd61b4 172.16.1.51:6380 slave 7c79559b280db9d9c182f3a25c718efe9e934fc7 0 1596767313935 6 connected
7c79559b280db9d9c182f3a25c718efe9e934fc7 172.16.1.52:6379 master - 0 1596767314949 2 connected 5461-10922

8.重新做主从

#由于使用工具，始终有一台机器为主库本机的从库，所以要重新分配主从
172.16.1.52:6380> CLUSTER REPLICATE d27553035a3e91c78d375208c72b756e9b2523d4
OK
172.16.1.53:6380> CLUSTER REPLICATE 5ad7bd957133eac9c3a692b35f8ae72258cf0ece
OK

9.插入数据测试

[root@db01 ~]# redis-trib.rb info 172.16.1.52:6379
172.16.1.52:6379 (7c79559b...) -> 332 keys | 5462 slots | 1 slaves.
172.16.1.51:6379 (5ad7bd95...) -> 341 keys | 5461 slots | 1 slaves.
172.16.1.53:6379 (d2755303...) -> 327 keys | 5461 slots | 1 slaves.
[OK] 1000 keys in 3 masters.
0.06 keys per slot on average.

[root@db01 ~]# redis-trib.rb info 172.16.1.52:6379
172.16.1.52:6379 (7c79559b...) -> 661 keys | 5462 slots | 1 slaves.
172.16.1.51:6379 (5ad7bd95...) -> 674 keys | 5461 slots | 1 slaves.
172.16.1.53:6379 (d2755303...) -> 665 keys | 5461 slots | 1 slaves.
[OK] 2000 keys in 3 masters.
0.12 keys per slot on average.