这篇文章主要介绍Redis集群的示例分析，文中介绍的非常详细，具有一定的参考价值，感兴趣的小伙伴们一定要看完！

Redis集群详解

Redis有三种集群模式，分别是：

*主从模式*Sentinel模式*Cluster模式

三种集群模式各有特点，关于Redis介绍可以参考这里：NoSQL（二)——Redis

Redis官网：https://redis.io/ ，最新版本5.0.4

主从模式

主从模式介绍

主从模式是三种模式中最简单的，在主从复制中，数据库分为两类：主数据库(master)和从数据库(slave)。

其中主从复制有如下特点：

*主数据库可以进行读写操作，当读写操作导致数据变化时会自动将数据同步给从数据库*从数据库一般都是只读的，并且接收主数据库同步过来的数据*一个master可以拥有多个slave，但是一个slave只能对应一个master*slave挂了不影响其他slave的读和master的读和写，重新启动后会将数据从master同步过来*master挂了以后，不影响slave的读，但redis不再提供写服务，master重启后redis将重新对外提供写服务*master挂了以后，不会在slave节点中重新选一个master

工作机制：

当slave启动后，主动向master发送SYNC命令。master接收到SYNC命令后在后台保存快照（RDB持久化）和缓存保存快照这段时间的命令，然后将保存的快照文件和缓存的命令发送给slave。slave接收到快照文件和命令后加载快照文件和缓存的执行命令。

复制初始化后，master每次接收到的写命令都会同步发送给slave，保证主从数据一致性。

安全设置：

当master节点设置密码后，

客户端访问master需要密码启动slave需要密码，在配置文件中配置即可客户端访问slave不需要密码

缺点：

从上面可以看出，master节点在主从模式中唯一，若master挂掉，则redis无法对外提供写服务。

主从模式搭建

环境准备：

master节点192.168.30.128slave节点192.168.30.129slave节点192.168.30.130

全部下载安装：

#cd/software#wgethttp://download.redis.io/releases/redis-5.0.4.tar.gz#tarzxfredis-5.0.4.tar.gz&&mvredis-5.0.4//usr/local/redis#cd/usr/local/redis&&make&&makeinstall#echo$?0

全部配置成服务：

服务文件

#vim/usr/lib/systemd/system/redis.service[Unit]Description=Redispersistentkey-valuedatabaseAfter=network.targetAfter=network-online.targetWants=network-online.target[Service]ExecStart=/usr/local/bin/redis-server/usr/local/redis/redis.conf--supervisedsystemdExecStop=/usr/libexec/redis-shutdownType=notifyUser=redisGroup=redisRuntimeDirectory=redisRuntimeDirectoryMode=0755[Install]WantedBy=multi-user.target

shutdown脚本

#vim/usr/libexec/redis-shutdown#!/bin/bash##Wrappertocloseproperlyredisandsentineltestx"$REDIS_DEBUG"!=x&&set-xREDIS_CLI=/usr/local/bin/redis-cli#RetrieveservicenameSERVICE_NAME="$1"if[-z"$SERVICE_NAME"];thenSERVICE_NAME=redisfi#GettheproperconfigfilebasedonservicenameCONFIG_FILE="/usr/local/redis/$SERVICE_NAME.conf"#Useawktoretrievehost,portfromconfigfileHOST=`awk'/^[[:blank:]]*bind/{print$2}'$CONFIG_FILE|tail-n1`PORT=`awk'/^[[:blank:]]*port/{print$2}'$CONFIG_FILE|tail-n1`PASS=`awk'/^[[:blank:]]*requirepass/{print$2}'$CONFIG_FILE|tail-n1`SOCK=`awk'/^[[:blank:]]*unixsocket\s/{print$2}'$CONFIG_FILE|tail-n1`#Justincase,usedefaulthost,portHOST=${HOST:-127.0.0.1}if["$SERVICE_NAME"=redis];thenPORT=${PORT:-6379}elsePORT=${PORT:-26739}fi#Setupadditionalparameters#e.gpassword-protectedredisinstances[-z"$PASS"]||ADDITIONAL_PARAMS="-a$PASS"#shutdowntheserviceproperlyif[-e"$SOCK"];then$REDIS_CLI-s$SOCK$ADDITIONAL_PARAMSshutdownelse$REDIS_CLI-h$HOST-p$PORT$ADDITIONAL_PARAMSshutdownfi

#chmod+x/usr/libexec/redis-shutdown#useradd-s/sbin/nologinredis#chown-Rredis:redis/usr/local/redis#chown-Rredis:redis/data/redis#yuminstall-ybash-completion&&source/etc/profile#命令补全#systemctldaemon-reload#systemctlenableredis

修改配置：

192.168.30.128

#mkdir-p/data/redis#vim/usr/local/redis/redis.confbind192.168.30.128#监听ip，多个ip用空格分隔daemonizeyes#允许后台启动logfile"/usr/local/redis/redis.log"#日志路径dir/data/redis#数据库备份文件存放目录masterauth123456#slave连接master密码，master可省略requirepass123456#设置master连接密码，slave可省略appendonlyyes#在/data/redis/目录生成appendonly.aof文件，将每一次写操作请求都追加到appendonly.aof文件中#echo'vm.overcommit_memory=1'>>/etc/sysctl.conf#sysctl-p

192.168.30.129

#mkdir-p/data/redis#vim/usr/local/redis/redis.confbind192.168.30.129daemonizeyeslogfile"/usr/local/redis/redis.log"dir/data/redisreplicaof192.168.30.1286379masterauth123456requirepass123456appendonlyyes#echo'vm.overcommit_memory=1'>>/etc/sysctl.conf#sysctl-p

192.168.30.130

#mkdir-p/data/redis#vim/usr/local/redis/redis.confbind192.168.30.130daemonizeyeslogfile"/usr/local/redis/redis.log"dir/data/redisreplicaof192.168.30.1286379masterauth123456requirepass123456appendonlyyes#echo'vm.overcommit_memory=1'>>/etc/sysctl.conf#sysctl-p

全部启动redis：

#systemctlstartredis

查看集群状态：

#redis-cli-h192.168.30.128-a123456Warning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.192.168.30.128:6379>inforeplication#Replicationrole:masterconnected_slaves:2slave0:ip=192.168.30.129,port=6379,state=online,offset=168,lag=1slave1:ip=192.168.30.130,port=6379,state=online,offset=168,lag=1master_replid:fb4941e02d5032ad74c6e2383211fc58963dbe90master_replid2:0000000000000000000000000000000000000000master_repl_offset:168second_repl_offset:-1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:1repl_backlog_histlen:168

#redis-cli-h192.168.30.129-a123456inforeplicationWarning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.#Replicationrole:slavemaster_host:192.168.30.128master_port:6379master_link_status:upmaster_last_io_seconds_ago:1master_sync_in_progress:0slave_repl_offset:196slave_priority:100slave_read_only:1connected_slaves:0master_replid:fb4941e02d5032ad74c6e2383211fc58963dbe90master_replid2:0000000000000000000000000000000000000000master_repl_offset:196second_repl_offset:-1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:1repl_backlog_histlen:196

数据演示：

192.168.30.128:6379>keys*(emptylistorset)192.168.30.128:6379>setkey1100OK192.168.30.128:6379>setkey2lzxOK192.168.30.128:6379>keys*1)"key1"2)"key2"

#redis-cli-h192.168.30.129-a123456Warning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.192.168.30.129:6379>keys*1)"key2"2)"key1"192.168.30.129:6379>CONFIGGETdir1)"dir"2)"/data/redis"192.168.30.129:6379>CONFIGGETdbfilename1)"dbfilename"2)"dump.rdb"192.168.30.129:6379>getkey1"100"192.168.30.129:6379>getkey2"lzx"192.168.30.129:6379>setkey3aaa(error)READONLYYoucan'twriteagainstareadonlyreplica.

#redis-cli-h192.168.30.130-a123456Warning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.192.168.30.130:6379>keys*1)"key2"2)"key1"192.168.30.130:6379>CONFIGGETdir1)"dir"2)"/data/redis"192.168.30.130:6379>CONFIGGETdbfilename1)"dbfilename"2)"dump.rdb"192.168.30.130:6379>getkey1"100"192.168.30.130:6379>getkey2"lzx"192.168.30.130:6379>setkey3aaa(error)READONLYYoucan'twriteagainstareadonlyreplica.

可以看到，在master节点写入的数据，很快就同步到slave节点上，而且在slave节点上无法写入数据。

Sentinel模式Sentinel模式介绍

主从模式的弊端就是不具备高可用性，当master挂掉以后，Redis将不能再对外提供写入操作，因此sentinel应运而生。

sentinel中文含义为哨兵，顾名思义，它的作用就是监控redis集群的运行状况，特点如下：

*sentinel模式是建立在主从模式的基础上，如果只有一个Redis节点，sentinel就没有任何意义*当master挂了以后，sentinel会在slave中选择一个做为master，并修改它们的配置文件，其他slave的配置文件也会被修改，比如slaveof属性会指向新的master*当master重新启动后，它将不再是master而是做为slave接收新的master的同步数据*sentinel因为也是一个进程有挂掉的可能，所以sentinel也会启动多个形成一个sentinel集群*多sentinel配置的时候，sentinel之间也会自动监控*当主从模式配置密码时，sentinel也会同步将配置信息修改到配置文件中，不需要担心*一个sentinel或sentinel集群可以管理多个主从Redis，多个sentinel也可以监控同一个redis*sentinel最好不要和Redis部署在同一台机器，不然Redis的服务器挂了以后，sentinel也挂了

工作机制：

*每个sentinel以每秒钟一次的频率向它所知的master，slave以及其他sentinel实例发送一个PING命令*如果一个实例距离最后一次有效回复PING命令的时间超过down-after-milliseconds选项所指定的值，则这个实例会被sentinel标记为主观下线。*如果一个master被标记为主观下线，则正在监视这个master的所有sentinel要以每秒一次的频率确认master的确进入了主观下线状态*当有足够数量的sentinel（大于等于配置文件指定的值）在指定的时间范围内确认master的确进入了主观下线状态，则master会被标记为客观下线*在一般情况下，每个sentinel会以每10秒一次的频率向它已知的所有master，slave发送INFO命令*当master被sentinel标记为客观下线时，sentinel向下线的master的所有slave发送INFO命令的频率会从10秒一次改为1秒一次*若没有足够数量的sentinel同意master已经下线，master的客观下线状态就会被移除；若master重新向sentinel的PING命令返回有效回复，master的主观下线状态就会被移除

当使用sentinel模式的时候，客户端就不要直接连接Redis，而是连接sentinel的ip和port，由sentinel来提供具体的可提供服务的Redis实现，这样当master节点挂掉以后，sentinel就会感知并将新的master节点提供给使用者。

Sentinel模式搭建

环境准备：

master节点192.168.30.128sentinel端口：26379slave节点192.168.30.129sentinel端口：26379slave节点192.168.30.130sentinel端口：26379

修改配置：

前面已经下载安装了redis，这里省略，直接修改sentinel配置文件。

192.168.30.128

#vim/usr/local/redis/sentinel.confdaemonizeyeslogfile"/usr/local/redis/sentinel.log"dir"/usr/local/redis/sentinel"#sentinel工作目录sentinelmonitormymaster192.168.30.12863792#判断master失效至少需要2个sentinel同意，建议设置为n/2+1，n为sentinel个数sentinelauth-passmymaster123456sentineldown-after-millisecondsmymaster30000#判断master主观下线时间，默认30s

这里需要注意，sentinel auth-pass mymaster 123456需要配置在sentinel monitor mymaster 192.168.30.128 6379 2下面，否则启动报错：

#/usr/local/bin/redis-sentinel/usr/local/redis/sentinel.conf***FATALCONFIGFILEERROR***Readingtheconfigurationfile,atline104>>>'sentinelauth-passmymaster123456'Nosuchmasterwithspecifiedname.

全部启动sentinel：

#mkdir/usr/local/redis/sentinel&&chown-Rredis:redis/usr/local/redis#/usr/local/bin/redis-sentinel/usr/local/redis/sentinel.conf

任一主机查看日志：

#tail-f/usr/local/redis/sentinel.log21574:X09May201915:32:04.298#SentinelIDis30c417116a8edbab09708037366c4a7471beb77021574:X09May201915:32:04.298#+monitormastermymaster192.168.30.1286379quorum221574:X09May201915:32:04.299*+slaveslave192.168.30.129:6379192.168.30.1296379@mymaster192.168.30.128637921574:X09May201915:32:04.300*+slaveslave192.168.30.130:6379192.168.30.1306379@mymaster192.168.30.128637921574:X09May201915:32:16.347*+sentinelsentinel79b8d61626afd4d059fb5a6a63393e9a1374e78f192.168.30.12926379@mymaster192.168.30.128637921574:X09May201915:32:31.584*+sentinelsentineld7b429dcba792103ef0d80827dd0910bd9284d21192.168.30.13026379@mymaster192.168.30.1286379

Sentinel模式下的几个事件：

·+reset-master：主服务器已被重置。·+slave：一个新的从服务器已经被Sentinel识别并关联。·+failover-state-reconf-slaves：故障转移状态切换到了reconf-slaves状态。·+failover-detected：另一个Sentinel开始了一次故障转移操作，或者一个从服务器转换成了主服务器。·+slave-reconf-sent：领头（leader）的Sentinel向实例发送了[SLAVEOF](/commands/slaveof.html)命令，为实例设置新的主服务器。·+slave-reconf-inprog：实例正在将自己设置为指定主服务器的从服务器，但相应的同步过程仍未完成。·+slave-reconf-done：从服务器已经成功完成对新主服务器的同步。·-dup-sentinel：对给定主服务器进行监视的一个或多个Sentinel已经因为重复出现而被移除——当Sentinel实例重启的时候，就会出现这种情况。·+sentinel：一个监视给定主服务器的新Sentinel已经被识别并添加。·+sdown：给定的实例现在处于主观下线状态。·-sdown：给定的实例已经不再处于主观下线状态。·+odown：给定的实例现在处于客观下线状态。·-odown：给定的实例已经不再处于客观下线状态。·+new-epoch：当前的纪元（epoch）已经被更新。·+try-failover：一个新的故障迁移操作正在执行中，等待被大多数Sentinel选中（waitingtobeelectedbythemajority）。·+elected-leader：赢得指定纪元的选举，可以进行故障迁移操作了。·+failover-state-select-slave：故障转移操作现在处于select-slave状态——Sentinel正在寻找可以升级为主服务器的从服务器。·no-good-slave：Sentinel操作未能找到适合进行升级的从服务器。Sentinel会在一段时间之后再次尝试寻找合适的从服务器来进行升级，又或者直接放弃执行故障转移操作。·selected-slave：Sentinel顺利找到适合进行升级的从服务器。·failover-state-send-slaveof-noone：Sentinel正在将指定的从服务器升级为主服务器，等待升级功能完成。·failover-end-for-timeout：故障转移因为超时而中止，不过最终所有从服务器都会开始复制新的主服务器（slaveswilleventuallybeconfiguredtoreplicatewiththenewmasteranyway）。·failover-end：故障转移操作顺利完成。所有从服务器都开始复制新的主服务器了。·+switch-master：配置变更，主服务器的IP和地址已经改变。这是绝大多数外部用户都关心的信息。·+tilt：进入tilt模式。·-tilt：退出tilt模式。

master宕机演示：

192.168.30.128

#systemctlstopredis#tail-f/usr/local/redis/sentinel.log22428:X09May201915:51:29.287#+sdownmastermymaster192.168.30.128637922428:X09May201915:51:29.371#+odownmastermymaster192.168.30.1286379#quorum2/222428:X09May201915:51:29.371#+new-epoch122428:X09May201915:51:29.371#+try-failovermastermymaster192.168.30.128637922428:X09May201915:51:29.385#+vote-for-leader30c417116a8edbab09708037366c4a7471beb770122428:X09May201915:51:29.403#d7b429dcba792103ef0d80827dd0910bd9284d21votedfor30c417116a8edbab09708037366c4a7471beb770122428:X09May201915:51:29.408#79b8d61626afd4d059fb5a6a63393e9a1374e78fvotedfor30c417116a8edbab09708037366c4a7471beb770122428:X09May201915:51:29.451#+elected-leadermastermymaster192.168.30.128637922428:X09May201915:51:29.451#+failover-state-select-slavemastermymaster192.168.30.128637922428:X09May201915:51:29.528#+selected-slaveslave192.168.30.129:6379192.168.30.1296379@mymaster192.168.30.128637922428:X09May201915:51:29.528*+failover-state-send-slaveof-nooneslave192.168.30.129:6379192.168.30.1296379@mymaster192.168.30.128637922428:X09May201915:51:29.594*+failover-state-wait-promotionslave192.168.30.129:6379192.168.30.1296379@mymaster192.168.30.128637922428:X09May201915:51:30.190#+promoted-slaveslave192.168.30.129:6379192.168.30.1296379@mymaster192.168.30.128637922428:X09May201915:51:30.190#+failover-state-reconf-slavesmastermymaster192.168.30.128637922428:X09May201915:51:30.258*+slave-reconf-sentslave192.168.30.130:6379192.168.30.1306379@mymaster192.168.30.128637922428:X09May201915:51:30.511#-odownmastermymaster192.168.30.128637922428:X09May201915:51:31.233*+slave-reconf-inprogslave192.168.30.130:6379192.168.30.1306379@mymaster192.168.30.128637922428:X09May201915:51:31.233*+slave-reconf-doneslave192.168.30.130:6379192.168.30.1306379@mymaster192.168.30.128637922428:X09May201915:51:31.297#+failover-endmastermymaster192.168.30.128637922428:X09May201915:51:31.297#+switch-mastermymaster192.168.30.1286379192.168.30.129637922428:X09May201915:51:31.298*+slaveslave192.168.30.130:6379192.168.30.1306379@mymaster192.168.30.129637922428:X09May201915:51:31.298*+slaveslave192.168.30.128:6379192.168.30.1286379@mymaster192.168.30.129637922428:X09May201915:52:31.307#+sdownslave192.168.30.128:6379192.168.30.1286379@mymaster192.168.30.1296379

从日志中可以看到，master已经从192.168.30.128转移到192.168.30.129上

192.168.30.129上查看集群信息

#/usr/local/bin/redis-cli-h192.168.30.129-p6379-a123456Warning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.192.168.30.129:6379>inforeplication#Replicationrole:masterconnected_slaves:1slave0:ip=192.168.30.130,port=6379,state=online,offset=291039,lag=1master_replid:757aff269236ed2707ba584a86a40716c1c76d74master_replid2:47a862fc0ff20362be29096ecdcca6d432070ee9master_repl_offset:291182second_repl_offset:248123repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:1repl_backlog_histlen:291182192.168.30.129:6379>setkey4linuxOK

当前集群中只有一个slave——192.168.30.130，master是192.168.30.129，且192.168.30.129具有写权限。

192.168.30.130上查看redis的配置文件也可以看到replicaof 192.168.30.129 6379，这是sentinel在选举master是做的修改。

重新把192.168.30.128上进程启动

#systemctlstartredis#tail-f/usr/local/redis/sentinel.log22428:X09May201915:51:31.297#+switch-mastermymaster192.168.30.1286379192.168.30.129637922428:X09May201915:51:31.298*+slaveslave192.168.30.130:6379192.168.30.1306379@mymaster192.168.30.129637922428:X09May201915:51:31.298*+slaveslave192.168.30.128:6379192.168.30.1286379@mymaster192.168.30.129637922428:X09May201915:52:31.307#+sdownslave192.168.30.128:6379192.168.30.1286379@mymaster192.168.30.129637922428:X09May201916:01:24.872#-sdownslave192.168.30.128:6379192.168.30.1286379@mymaster192.168.30.1296379

查看集群信息

#/usr/local/bin/redis-cli-h192.168.30.128-p6379-a123456Warning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.192.168.30.128:6379>inforeplication#Replicationrole:slavemaster_host:192.168.30.129master_port:6379master_link_status:upmaster_last_io_seconds_ago:0master_sync_in_progress:0slave_repl_offset:514774slave_priority:100slave_read_only:1connected_slaves:0master_replid:757aff269236ed2707ba584a86a40716c1c76d74master_replid2:0000000000000000000000000000000000000000master_repl_offset:514774second_repl_offset:-1repl_backlog_active:1repl_backlog_size:1048576repl_backlog_first_byte_offset:376528repl_backlog_histlen:138247192.168.30.128:6379>getkey4"linux"192.168.30.128:6379>setkey5(error)ERRwrongnumberofargumentsfor'set'command

即使192.168.30.128重新启动redis服务，也是作为slave加入redis集群，192.168.30.129仍然是master。

Cluster模式Cluster模式介绍

sentinel模式基本可以满足一般生产的需求，具备高可用性。但是当数据量过大到一台服务器存放不下的情况时，主从模式或sentinel模式就不能满足需求了，这个时候需要对存储的数据进行分片，将数据存储到多个Redis实例中。cluster模式的出现就是为了解决单机Redis容量有限的问题，将Redis的数据根据一定的规则分配到多台机器。

cluster可以说是sentinel和主从模式的结合体，通过cluster可以实现主从和master重选功能，所以如果配置两个副本三个分片的话，就需要六个Redis实例。因为Redis的数据是根据一定规则分配到cluster的不同机器的，当数据量过大时，可以新增机器进行扩容。

使用集群，只需要将redis配置文件中的cluster-enable配置打开即可。每个集群中至少需要三个主数据库才能正常运行，新增节点非常方便。

cluster集群特点：

*多个redis节点网络互联，数据共享*所有的节点都是一主一从（也可以是一主多从），其中从不提供服务，仅作为备用*不支持同时处理多个key（如MSET/MGET），因为redis需要把key均匀分布在各个节点上，并发量很高的情况下同时创建key-value会降低性能并导致不可预测的行为*支持在线增加、删除节点*客户端可以连接任何一个主节点进行读写Cluster模式搭建

环境准备：

三台机器，分别开启两个redis服务（端口）192.168.30.128端口：7001,7002192.168.30.129端口：7003,7004192.168.30.130端口：7005,7006

修改配置文件：

192.168.30.128

#mkdir/usr/local/redis/cluster#cp/usr/local/redis/redis.conf/usr/local/redis/cluster/redis_7001.conf#cp/usr/local/redis/redis.conf/usr/local/redis/cluster/redis_7002.conf#chown-Rredis:redis/usr/local/redis#mkdir-p/data/redis/cluster/{redis_7001,redis_7002}&&chown-Rredis:redis/data/redis

#vim/usr/local/redis/cluster/redis_7001.confbind192.168.30.128port7001daemonizeyespidfile"/var/run/redis_7001.pid"logfile"/usr/local/redis/cluster/redis_7001.log"dir"/data/redis/cluster/redis_7001"#replicaof192.168.30.1296379masterauth123456requirepass123456appendonlyyescluster-enabledyescluster-config-filenodes_7001.confcluster-node-timeout15000

#vim/usr/local/redis/cluster/redis_7002.confbind192.168.30.128port7002daemonizeyespidfile"/var/run/redis_7002.pid"logfile"/usr/local/redis/cluster/redis_7002.log"dir"/data/redis/cluster/redis_7002"#replicaof192.168.30.1296379masterauth"123456"requirepass"123456"appendonlyyescluster-enabledyescluster-config-filenodes_7002.confcluster-node-timeout15000

其它两台机器配置与192.168.30.128一致，此处省略

启动redis服务：

#redis-server/usr/local/redis/cluster/redis_7001.conf#tail-f/usr/local/redis/cluster/redis_7001.log#redis-server/usr/local/redis/cluster/redis_7002.conf#tail-f/usr/local/redis/cluster/redis_7002.log

其它两台机器启动与192.168.30.128一致，此处省略

安装ruby并创建集群（低版本）：

如果redis版本比较低，则需要安装ruby。任选一台机器安装ruby即可

#yum-ygroupinstall"DevelopmentTools"#yuminstall-ygdbm-devellibdb4-devellibffi-devellibyamllibyaml-develncurses-developenssl-develreadline-develtcl-devel#mkdir-p~/rpmbuild/{BUILD,BUILDROOT,RPMS,SOURCES,SPECS,SRPMS}#wgethttp://cache.ruby-lang.org/pub/ruby/2.2/ruby-2.2.3.tar.gz-P~/rpmbuild/SOURCES#wgethttp://raw.githubusercontent.com/tjinjin/automate-ruby-rpm/master/ruby22x.spec-P~/rpmbuild/SPECS#rpmbuild-bb~/rpmbuild/SPECS/ruby22x.spec#rpm-ivh~/rpmbuild/RPMS/x86_64/ruby-2.2.3-1.el7.x86_64.rpm#geminstallredis#目的是安装这个，用于配置集群

#cp/usr/local/redis/src/redis-trib.rb/usr/bin/#redis-trib.rbcreate--replicas1192.168.30.128:7001192.168.30.128:7002192.168.30.129:7003192.168.30.129:7004192.168.30.130:7005192.168.30.130:7006

创建集群：

我这里是redis5.0.4，所以不需要安装ruby，直接创建集群即可

#redis-cli-a123456--clustercreate192.168.30.128:7001192.168.30.128:7002192.168.30.129:7003192.168.30.129:7004192.168.30.130:7005192.168.30.130:7006--cluster-replicas1Warning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.>>>Performinghashslotsallocationon6nodes...Master[0]->Slots0-5460Master[1]->Slots5461-10922Master[2]->Slots10923-16383Addingreplica192.168.30.129:7004to192.168.30.128:7001Addingreplica192.168.30.130:7006to192.168.30.129:7003Addingreplica192.168.30.128:7002to192.168.30.130:7005M:80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001slots:[0-5460](5461slots)masterS:b4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002replicates6788453ee9a8d7f72b1d45a9093838efd0e501f1M:4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003slots:[5461-10922](5462slots)masterS:b6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004replicates80c80a3f3e33872c047a8328ad579b9bea001ad8M:6788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005slots:[10923-16383](5461slots)masterS:277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006replicates4d74ec66e898bf09006dac86d4928f9fad81f373CanIsettheaboveconfiguration?(type'yes'toaccept):yes#输入yes，接受上面配置>>>Nodesconfigurationupdated>>>Assignadifferentconfigepochtoeachnode>>>SendingCLUSTERMEETmessagestojointhecluster

可以看到，

192.168.30.128:7001是master，它的slave是192.168.30.129:7004；192.168.30.129:7003是master，它的slave是192.168.30.130:7006；192.168.30.130:7005是master，它的slave是192.168.30.128:7002

自动生成nodes.conf文件：

#ls/data/redis/cluster/redis_7001/appendonly.aofdump.rdbnodes-7001.conf#vim/data/redis/cluster/redis_7001/nodes-7001.conf6788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574544063125connected10923-16383277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574544070006connectedb4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574544083715connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001myself,master-015574544060001connected0-5460b6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574544073664connected4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574544070003connected5461-10922varscurrentEpoch6lastVoteEpoch0集群操作

登录集群：

#redis-cli-c-h192.168.30.128-p7001-a123456#-c，使用集群方式登录

查看集群信息：

192.168.30.128:7001>CLUSTERINFO#集群状态cluster_state:okcluster_slots_assigned:16384cluster_slots_ok:16384cluster_slots_pfail:0cluster_slots_fail:0cluster_known_nodes:6cluster_size:3cluster_current_epoch:6cluster_my_epoch:1cluster_stats_messages_ping_sent:580cluster_stats_messages_pong_sent:551cluster_stats_messages_sent:1131cluster_stats_messages_ping_received:546cluster_stats_messages_pong_received:580cluster_stats_messages_meet_received:5cluster_stats_messages_received:1131

列出节点信息：

192.168.30.128:7001>CLUSTERNODES#列出节点信息6788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574551760005connected10923-16383277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574551740006connectedb4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574551750005connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001myself,master-015574551750001connected0-5460b6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574551749894connected4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574551759953connected5461-10922

这里与nodes.conf文件内容相同

写入数据：

192.168.30.128:7001>setkey111aaa->Redirectedtoslot[13680]locatedat192.168.30.130:7005#说明数据到了192.168.30.130:7005上OK192.168.30.130:7005>setkey222bbb->Redirectedtoslot[2320]locatedat192.168.30.128:7001#说明数据到了192.168.30.128:7001上OK192.168.30.128:7001>setkey333ccc->Redirectedtoslot[7472]locatedat192.168.30.129:7003#说明数据到了192.168.30.129:7003上OK192.168.30.129:7003>getkey111->Redirectedtoslot[13680]locatedat192.168.30.130:7005"aaa"192.168.30.130:7005>getkey333->Redirectedtoslot[7472]locatedat192.168.30.129:7003"ccc"192.168.30.129:7003>

可以看出redis cluster集群是去中心化的，每个节点都是平等的，连接哪个节点都可以获取和设置数据。

当然，平等指的是master节点，因为slave节点根本不提供服务，只是作为对应master节点的一个备份。

增加节点：

192.168.30.129上增加一节点：

#cp/usr/local/redis/cluster/redis_7003.conf/usr/local/redis/cluster/redis_7007.conf#vim/usr/local/redis/cluster/redis_7007.confbind192.168.30.129port7007daemonizeyespidfile"/var/run/redis_7007.pid"logfile"/usr/local/redis/cluster/redis_7007.log"dir"/data/redis/cluster/redis_7007"#replicaof192.168.30.1296379masterauth"123456"requirepass"123456"appendonlyyescluster-enabledyescluster-config-filenodes_7007.confcluster-node-timeout15000#mkdir/data/redis/cluster/redis_7007#chown-Rredis:redis/usr/local/redis&&chown-Rredis:redis/data/redis#redis-server/usr/local/redis/cluster/redis_7007.conf

192.168.30.130上增加一节点：

#cp/usr/local/redis/cluster/redis_7005.conf/usr/local/redis/cluster/redis_7008.conf#vim/usr/local/redis/cluster/redis_7007.confbind192.168.30.130port7008daemonizeyespidfile"/var/run/redis_7008.pid"logfile"/usr/local/redis/cluster/redis_7008.log"dir"/data/redis/cluster/redis_7008"#replicaof192.168.30.1306379masterauth"123456"requirepass"123456"appendonlyyescluster-enabledyescluster-config-filenodes_7008.confcluster-node-timeout15000#mkdir/data/redis/cluster/redis_7008#chown-Rredis:redis/usr/local/redis&&chown-Rredis:redis/data/redis#redis-server/usr/local/redis/cluster/redis_7008.conf

集群中增加节点：

192.168.30.129:7003>CLUSTERMEET192.168.30.1297007OK192.168.30.129:7003>CLUSTERNODES4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003myself,master-015574573610003connected5461-1092280c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001master-015574573647461connected0-5460277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574573620006connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574573630004connectedb4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574573620005connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574573627290connected6788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574573637395connected10923-16383

192.168.30.129:7003>CLUSTERMEET192.168.30.1307008OK192.168.30.129:7003>CLUSTERNODES4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003myself,master-015574574890003connected5461-1092280c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001master-015574574890001connected0-5460277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574574890006connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574574880004connectedb4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574574894725connected1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008master-015574574892590connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574574890000connected6788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574574904755connected10923-16383

可以看到，新增的节点都是以master身份加入集群的

更换节点身份：

将新增的192.168.30.130:7008节点身份改为192.168.30.129:7007的slave

#redis-cli-c-h192.168.30.130-p7008-a123456clusterreplicatee51ab166bc0f33026887bcf8eba0dff3d5b0bf14

cluster replicate后面跟node_id，更改对应节点身份。也可以登入集群更改

#redis-cli-c-h192.168.30.130-p7008-a123456192.168.30.130:7008>CLUSTERREPLICATEe51ab166bc0f33026887bcf8eba0dff3d5b0bf14OK192.168.30.130:7008>CLUSTERNODES277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574583168813connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001master-015574583148641connected0-54604d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574583160003connected5461-109226788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574583158725connected10923-16383b4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574583178905connected1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008myself,slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574583150007connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574583150001connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574583140000connected

查看相应的nodes.conf文件，可以发现有更改，它记录当前集群的节点信息

#cat/data/redis/cluster/redis_7001/nodes-7001.conf1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574582361697connected6788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574582350005connected10923-16383277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574582341036connectedb4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574582351295connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001myself,master-015574582340001connected0-5460b6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574582360004connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574582360000connected4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574582330893connected5461-10922varscurrentEpoch7lastVoteEpoch0

删除节点：

192.168.30.130:7008>CLUSTERFORGET1a1c7f02fce87530bd5abdfc98df1cffce4f1767(error)ERRItriedhardbutIcan'tforgetmyself...#无法删除登录节点192.168.30.130:7008>CLUSTERFORGETe51ab166bc0f33026887bcf8eba0dff3d5b0bf14(error)ERRCan'tforgetmymaster!#不能删除自己的master节点192.168.30.130:7008>CLUSTERFORGET6788453ee9a8d7f72b1d45a9093838efd0e501f1OK#可以删除其它的master节点192.168.30.130:7008>CLUSTERNODES277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574588873283connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001master-015574588870001connected0-54604d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574588860003connected5461-10922b4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave-015574588883515connected1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008myself,slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574588850007connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574588832891connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574588853100connected192.168.30.130:7008>CLUSTERFORGETb4d3eb411a7355d4767c6c23b4df69fa183ef8bcOK#可以删除其它的slave节点192.168.30.130:7008>CLUSTERNODES277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574590313973connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001master-015574590324071connected0-54604d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574590354343connected5461-109226788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574590340005connected10923-163831a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008myself,slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574590320007connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574590340001connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574590344270connected

保存配置：

192.168.30.130:7008>CLUSTERSAVECONFIG#将节点配置信息保存到硬盘OK#cat/data/redis/cluster/redis_7001/nodes-7001.conf1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574582361697connected6788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574582350005connected10923-16383277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574582341036connectedb4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574582351295connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001myself,master-015574582340001connected0-5460b6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574582360004connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574582360000connected4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574582330893connected5461-10922varscurrentEpoch7lastVoteEpoch0#redis-cli-c-h192.168.30.130-p7008-a123456Warning:Usingapasswordwith'-a'or'-u'optiononthecommandlineinterfacemaynotbesafe.192.168.30.130:7008>CLUSTERNODES277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574595007413connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001master-015574595000001connected0-54604d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574595010003connected5461-109226788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574595000005connected10923-16383b4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574594997375connected1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008myself,slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574594990007connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004slave80c80a3f3e33872c047a8328ad579b9bea001ad8015574595017501connectede51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574594980000connected

可以看到，之前删除的节点又恢复了，这是因为对应的配置文件没有删除，执行CLUSTER SAVECONFIG恢复。

模拟master节点挂掉：

192.168.30.128

#netstat-lntp|grep7001tcp00192.168.30.128:170010.0.0.0:*LISTEN6701/redis-server1tcp00192.168.30.128:70010.0.0.0:*LISTEN6701/redis-server1#kill6701

192.168.30.130:7008>CLUSTERNODES277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574611780003connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001master,fail-155746095048315574609471451disconnected4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574611749223connected5461-109226788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574611810035connected10923-16383b4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574611799935connected1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008myself,slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574611760007connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004master-015574611789818connected0-5460e51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574611790000connected

对应7001的一行可以看到，master fail，状态为disconnected；而对应7004的一行，slave已经变成master。

重新启动7001节点：

#redis-server/usr/local/redis/cluster/redis_7001.conf192.168.30.130:7008>CLUSTERNODES277daeb8660d5273b7c3e05c263f861ed5f17b92192.168.30.130:7006@17006slave4d74ec66e898bf09006dac86d4928f9fad81f373015574613070003connected80c80a3f3e33872c047a8328ad579b9bea001ad8192.168.30.128:7001@17001slaveb6331cbc986794237c83ed2d5c30777c1551546e015574613054418connected4d74ec66e898bf09006dac86d4928f9fad81f373192.168.30.129:7003@17003master-015574613079623connected5461-109226788453ee9a8d7f72b1d45a9093838efd0e501f1192.168.30.130:7005@17005master-015574613049355connected10923-16383b4d3eb411a7355d4767c6c23b4df69fa183ef8bc192.168.30.128:7002@17002slave6788453ee9a8d7f72b1d45a9093838efd0e501f1015574613060005connected1a1c7f02fce87530bd5abdfc98df1cffce4f1767192.168.30.130:7008@17008myself,slavee51ab166bc0f33026887bcf8eba0dff3d5b0bf14015574613050007connectedb6331cbc986794237c83ed2d5c30777c1551546e192.168.30.129:7004@17004master-015574613089728connected0-5460e51ab166bc0f33026887bcf8eba0dff3d5b0bf14192.168.30.129:7007@17007master-015574613070000connected

可以看到，7001节点启动后为slave节点，并且是7004的slave节点。即master节点如果挂掉，它的slave节点变为新master节点继续对外提供服务，而原来的master节点如果重启，则变为新master节点的slave节点。

另外，如果这里是拿7007节点做测试的话，会发现7008节点并不会切换，这是因为7007节点上根本没数据。集群数据被分为三份，采用哈希槽 (hash slot)的方式来分配16384个slot的话，它们三个节点分别承担的slot 区间是：

节点7004覆盖0－5460节点7003覆盖5461－10922节点7005覆盖10923－16383

以上是“Redis集群的示例分析”这篇文章的所有内容，感谢各位的阅读！希望分享的内容对大家有帮助，更多相关知识，欢迎关注亿速云行业资讯频道！