背景
本篇主要讲Redis Cluster的安装，让我们先用起来，感受一下到底是怎么回事，后面再继续学习有关如何Cluster如何failover,添加节点，删除节点，迁移slots等功能。
Redis Cluster特点
(1)Redis Cluster 共有16384(0-16383)个hash slots,数据写入时，根据CRC16('key')%16384 hash slots分配到不同的节点上；
(2)当整个集群部分节点crash不影响继续使用，如有A,B,C三个master和A1,B1,C1三个slave组成的cluster,如果B宕机，则B1会自动提成为master,当B恢复之后，也会自动成为B1的slave,保证集群的高可用性；
(3)当集群负载比较高，不足以支撑现在的业务，可以添加空节点，然后rebalance slot或者reshard slot；
安装Redis Cluster
(1)环境准备
Redis版本都是3.2.9,有6个redis实例分别在6台不同的主机上运行，其中3个作为Master,另外3个分别作为它们的Slave存在，端口统一使用默认的端口6379；
Master1: sht-sgmhadoopcm-01(172.16.101.54)
Slave1: sht-sgmhadoopdn-01(172.16.101.58)
Master2: sht-sgmhadoopnn-01(172.16.101.55)
Slave2: sht-sgmhadoopdn-02(172.16.101.59)
Master3: sht-sgmhadoopnn-02(172.16.101.56)
Slave3: sht-sgmhadoopdn-03(172.16.101.60)
#修改配置文件，每个节点都需要执行，修改其他节点对应的IP地址
[root@sht-sgmhadoopcm-01 redis]# vim redis.conf
port 6379
bind 172.16.101.54
appendonly yes
cluster-enabled yes
cluster-config-file nodes-6379.conf
cluster-node-timeout 15000
(2)在需要ruby的节点上安装ruby
[root@sht-sgmhadoopcm-01 redis]# yum install ruby -y
#下面这一步可能报错，因为CentOS7.3默认安装的ruby2.0.0，版本太低
[root@sht-sgmhadoopcm-01 redis]# gem install redis
Fetching: redis-4.0.1.gem (100%)
Successfully installed redis-4.0.1
Parsing documentation for redis-4.0.1
Installing ri documentation for redis-4.0.1
Done installing documentation for redis after 2 seconds
1 gem installed
(3)启动所有节点
[root@sht-sgmhadoopcm-01 redis]# src/redis-server redis.conf
[root@sht-sgmhadoopnn-02 redis]# src/redis-server redis.conf
[root@sht-sgmhadoopnn-01 redis]# src/redis-server redis.conf
[root@sht-sgmhadoopdn-01 redis]# src/redis-server redis.conf
[root@sht-sgmhadoopdn-02 redis]# src/redis-server redis.conf
[root@sht-sgmhadoopdn-03 redis]# src/redis-server redis.conf
[root@sht-sgmhadoopnn-02 redis]# ps -ef|grep redis
root 1716 1 0 12:40 ? 00:00:00 src/redis-server 172.16.101.56:6379 [cluster]
root 1720 1504 0 12:41 pts/2 00:00:00 grep --color=auto redis
(4)使用redis-trib.rb脚本创建cluster
Create的过程中已经给三个master分配好了hash slots,并且制定好Slave;
选项--replicas 1 表示为每个创建的主服务器节点创建一个从服务器节点;
[root@sht-sgmhadoopcm-01 redis]# ruby src/redis-trib.rb create --replicas 1 172.16.101.54:6379 172.16.101.55:6379 172.16.101.56:6379 172.16.101.58:6379 172.16.101.59:6379 172.16.101.60:6379
>>> Creating cluster
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
172.16.101.54:6379
172.16.101.55:6379
172.16.101.56:6379
Adding replica 172.16.101.58:6379 to 172.16.101.54:6379
Adding replica 172.16.101.59:6379 to 172.16.101.55:6379
Adding replica 172.16.101.60:6379 to 172.16.101.56:6379
M: 610aa83831404be545b25cc7f7322e987da1dd33 172.16.101.54:6379
slots:0-5460 (5461 slots) master
M: 4c7348ac575a1a3bd4357e937a8f94f1654193cf 172.16.101.55:6379
slots:5461-10922 (5462 slots) master
M: ece3a80f6989fa7f8316d73ee30ceea84340f24e 172.16.101.56:6379
slots:10923-16383 (5461 slots) master
S: 28e8d5c8fd12533622d8110f8d262cb50120ca02 172.16.101.58:6379
replicates 610aa83831404be545b25cc7f7322e987da1dd33
S: b50ba506525ee547823b3d9a7e5a095933e3ff42 172.16.101.59:6379
replicates 4c7348ac575a1a3bd4357e937a8f94f1654193cf
S: c79de1092011ed395ee772969274c0567c2b5dd1 172.16.101.60:6379
replicates ece3a80f6989fa7f8316d73ee30ceea84340f24e
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 172.16.101.54:6379)
M: 610aa83831404be545b25cc7f7322e987da1dd33 172.16.101.54:6379
slots:0-5460 (5461 slots) master
1 additional replica(s)
S: b50ba506525ee547823b3d9a7e5a095933e3ff42 172.16.101.59:6379
slots: (0 slots) slave
replicates 4c7348ac575a1a3bd4357e937a8f94f1654193cf
S: 28e8d5c8fd12533622d8110f8d262cb50120ca02 172.16.101.58:6379
slots: (0 slots) slave
replicates 610aa83831404be545b25cc7f7322e987da1dd33
M: 4c7348ac575a1a3bd4357e937a8f94f1654193cf 172.16.101.55:6379
slots:5461-10922 (5462 slots) master
1 additional replica(s)
M: ece3a80f6989fa7f8316d73ee30ceea84340f24e 172.16.101.56:6379
slots:10923-16383 (5461 slots) master
1 additional replica(s)
S: c79de1092011ed395ee772969274c0567c2b5dd1 172.16.101.60:6379
slots: (0 slots) slave
replicates ece3a80f6989fa7f8316d73ee30ceea84340f24e
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
(5)查看Cluster基本信息
[root@sht-sgmhadoopcm-01 redis]# src/redis-cli -c -h 172.16.101.54 -p 6379
172.16.101.54:6379> info replication
# Replication
role:master
connected_slaves:1
slave0:ip=172.16.101.58,port=6379,state=online,offset=2591,lag=0
master_repl_offset:2591
repl_backlog_active:1
repl_backlog_size:1048576
repl_backlog_first_byte_offset:2
repl_backlog_histlen:2590
172.16.101.54:6379> 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:4462
cluster_stats_messages_received:4462
172.16.101.54:6379> cluster nodes
b50ba506525ee547823b3d9a7e5a095933e3ff42 172.16.101.59:6379 slave 4c7348ac575a1a3bd4357e937a8f94f1654193cf 0 1532497136876 5 connected
610aa83831404be545b25cc7f7322e987da1dd33 172.16.101.54:6379 myself,master - 0 0 1 connected 0-5460
28e8d5c8fd12533622d8110f8d262cb50120ca02 172.16.101.58:6379 slave 610aa83831404be545b25cc7f7322e987da1dd33 0 1532497135872 4 connected
4c7348ac575a1a3bd4357e937a8f94f1654193cf 172.16.101.55:6379 master - 0 1532497133867 2 connected 5461-10922
ece3a80f6989fa7f8316d73ee30ceea84340f24e 172.16.101.56:6379 master - 0 1532497134870 3 connected 10923-16383
c79de1092011ed395ee772969274c0567c2b5dd1 172.16.101.60:6379 slave ece3a80f6989fa7f8316d73ee30ceea84340f24e 0 1532497137878 6 connected
[root@sht-sgmhadoopcm-01 redis]# cat nodes-6379.conf
b50ba506525ee547823b3d9a7e5a095933e3ff42 172.16.101.59:6379 slave 4c7348ac575a1a3bd4357e937a8f94f1654193cf 0 1532493960094 5 connected
610aa83831404be545b25cc7f7322e987da1dd33 172.16.101.54:6379 myself,master - 0 0 1 connected 0-5460
28e8d5c8fd12533622d8110f8d262cb50120ca02 172.16.101.58:6379 slave 610aa83831404be545b25cc7f7322e987da1dd33 0 1532493955077 4 connected
4c7348ac575a1a3bd4357e937a8f94f1654193cf 172.16.101.55:6379 master - 0 1532493958087 2 connected 5461-10922
ece3a80f6989fa7f8316d73ee30ceea84340f24e 172.16.101.56:6379 master - 0 1532493956080 3 connected 10923-16383
c79de1092011ed395ee772969274c0567c2b5dd1 172.16.101.60:6379 slave ece3a80f6989fa7f8316d73ee30ceea84340f24e 0 1532493959091 6 connected
vars currentEpoch 6 lastVoteEpoch 0
FAQ
Error1: ruby2.0.0版本太低
[root@sht-sgmhadoopcm-01 redis]# gem install redis
Fetching: redis-4.0.1.gem (100%)
ERROR: Error installing redis:
redis requires Ruby version >= 2.2.2.
解决方法：
[root@sht-sgmhadoopcm-01 redis]# ruby --version
ruby 2.0.0p648 (2015-12-16) [x86_64-linux]
[root@sht-sgmhadoopcm-01 redis]# gpg --keyserver hkp://keys.gnupg.net --recv-keys 409B6B1796C275462A1703113804BB82D39DC0E3 7D2BAF1CF37B13E2069D6956105BD0E739499BDB
[root@sht-sgmhadoopcm-01 redis]# curl -sSLhttps://get.rvm.io | bash -s stable
[root@sht-sgmhadoopcm-01 redis]# find / -name rvm -print
/usr/local/rvm
/usr/local/rvm/src/rvm
/usr/local/rvm/src/rvm/bin/rvm
/usr/local/rvm/src/rvm/lib/rvm
/usr/local/rvm/src/rvm/scripts/rvm
/usr/local/rvm/bin/rvm
/usr/local/rvm/lib/rvm
/usr/local/rvm/scripts/rvm
[root@sht-sgmhadoopcm-01 redis]# source /usr/local/rvm/scripts/rvm
[root@sht-sgmhadoopcm-01 redis]# rvm list known
[root@sht-sgmhadoopcm-01 redis]# rvm install 2.4.1
[root@sht-sgmhadoopcm-01 redis]# rvm use 2.4.1
Using /usr/local/rvm/gems/ruby-2.4.1
[root@sht-sgmhadoopcm-01 redis]# rvm use 2.4.1 --default
Using /usr/local/rvm/gems/ruby-2.4.1
[root@sht-sgmhadoopcm-01 redis]# rvm remove 2.3.4
ruby-2.3.4 - #already gone
Using /usr/local/rvm/gems/ruby-2.4.1
[root@sht-sgmhadoopcm-01 redis]# ruby --version
ruby 2.4.1p111 (2017-03-22 revision 58053) [x86_64-linux]
Error2：添加节点数据不为空
[root@sht-sgmhadoopcm-01 redis]# src/redis-trib.rb add-node 172.16.101.66:6379 172.16.101.54:6379
......
[ERR] Node 172.16.101.66:6379 is not empty. Either the node already knows other nodes (check with CLUSTER NODES) or contains some key in database 0.
解决方法：
(1)删除aof和rdb文件
(2)删除nodes-6379.conf配置文件
(3)执行flushdb and flushall
参考链接
https://redis.io/topics/cluster-tutorial