redis现在网站开发中经常用到 key-value存储系统 数据类型更加丰富 能做主从，集群 在redis3.0以前要做集群需要采用proxy的方法 在3.0（不过现在还在rc版本）自身提供了cluster的方法 这里没有用常用的一致性hash的方法，而是引入了新的概念叫hashslot（哈希
redis现在网站开发中经常用到
key-value存储系统
数据类型更加丰富
能做主从，集群
在redis3.0以前要做集群需要采用proxy的方法
在3.0（不过现在还在rc版本）自身提供了cluster的方法
这里没有用常用的一致性hash的方法，而是引入了新的概念叫hash slot（哈希槽），总共16384个 redis php 第一步：安装redis3.0，安装这里就不说了第二步：要做redis集群，至少需要3个master的服务器，我先测试的是具有子节点的，那个就需要6个服务器，3个主服务器，3个从服务器我们来按照端口的不同来进行划分，6379、6380、6381、6382、6383、6384创建这样6个目录，每个目录里有这样两个文件：redis-server、redis.confredis-server是从redis的安装目录中的src里copy过来就好redis.conf内容如下；#======================================#port 6379 #不同目录，这里端口对应不同cluster-enabled yes #开启集群功能cluster-config-file nodes.conf #节点配置文件，这个文件是服务启动时自己配置创建的cluster-node-timeout 5000 appendonly yes#======================================#那么6个服务器都配置好后则启动他们，方法如下：./redis-server ./redis.conf启动好以后查看一下是否都已经成功运行了 ps aux|grep redis第三步：创建集群./redis-trib.rb create --replicas 1 192.168.1.25:6379 192.168.1.25:6380 192.168.1.25:6381 192.168.1.25:6382 192.168.1.25:6383 192.168.1.25:6384这里的--replicas 1 表示每个主节点下有一个从节点操作结果如下>>> creating clusterconnecting to node 192.168.1.25:6379: okconnecting to node 192.168.1.25:6380: okconnecting to node 192.168.1.25:6381: okconnecting to node 192.168.1.25:6382: okconnecting to node 192.168.1.25:6383: okconnecting to node 192.168.1.25:6384: ok>>> performing hash slots allocation on 6 nodes...using 3 masters:192.168.1.25:6379192.168.1.25:6380192.168.1.25:6381adding replica 192.168.1.25:6382 to 192.168.1.25:6379adding replica 192.168.1.25:6383 to 192.168.1.25:6380adding replica 192.168.1.25:6384 to 192.168.1.25:6381说明了将前3个服作了主节点，后3个服务作为了从节点，并且是相对应的./redis-trib.rb是ruby程序，需要到rubygem去安装redis的包yum install rubygemsgem install redis #如果这样不能安装就下载到本地安装吧下面进行测试，使用redis-cli首先登入6380这个端口的redis服务./redis-cli -c -h 192.168.1.25 -p 6380192.168.1.25:6380> set name andyhua.chen退出后重新登入6379端口的redis服务./redis-cli -c -h 192.168.1.25 -p 6379192.168.1.25：6379> get name #如果没有集群的话，这里肯定是没有的，现在则出现如下情况-> redirected to slot [5798] located at 192.168.1.25:6380andyhua.chen192.168.1.25：6380> #可以看到get之后终端自动跳转到了6380端口下并取得了之前set过的name的值
./redis-trib.rb create --replicas 1 192.168.1.25:6379 192.168.1.25:6380 192.168.1.25:6381 192.168.1.25:6382 192.168.1.25:6383 192.168.1.25:6384#这里的--replicas 1 表示每个主节点下有一个从节点#分片情况如下：nodea: slots:0-5460 (5461 slots) masternodeb: slots:5461-10922 (5462 slots) masternodec: slots:10923-16383 (5461 slots) master
#如果不要从节点的方法如下./redis-trib.rb create --replicas 0 192.168.1.25:6379 192.168.1.25:6380 192.168.1.25:6381#分片情况如下：nodea: slots:0-5460 (5461 slots) masternodeb: slots:5461-10922 (5462 slots) masternodec: slots:10923-16383 (5461 slots) master
./redis-trib.rb check 192.168.1.25:6379#在创建集群里redis会自动执行这样的检测，当然你也可以自己运行这个check，端口（节点）也是任一其中一个---------------------------------------------------------------------------------connecting to node 192.168.1.25:6379: okconnecting to node 192.168.1.25:6381: okconnecting to node 192.168.1.25:6380: ok>>> performing cluster check (using node 192.168.1.25:6379)m: 5dfa7d9f2dbb4ab64b71f2713538125354c88462 192.168.1.25:6379 slots:0-5460 (5461 slots) master 0 additional replica(s)m: 47b1c1df820ff5e9057a526f69e9fe4fdd9800e9 192.168.1.25:6381 slots:10923-16383 (5461 slots) master 0 additional replica(s)m: c97f77ebcba3f48b9da07329b0e4e2943ded88c9 192.168.1.25:6380 slots:5461-10922 (5462 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.
#查看集群节点的相关信息redis-cli -p 6379 -h 192.168.1.25 cluster nodes-------------------------------------------------------------------------------------47b1c1df820ff5e9057a526f69e9fe4fdd9800e9 192.168.1.25:6381 master - 0 1416900312414 3 connected 10923-16383c97f77ebcba3f48b9da07329b0e4e2943ded88c9 192.168.1.25:6380 master - 0 1416900313439 2 connected 5461-109225dfa7d9f2dbb4ab64b71f2713538125354c88462 192.168.1.25:6379 myself,master - 0 0 1 connected 0-5460--------------------------------------------------------------------------------------#官方文档里列表参数的说明：* node id* ip:port* flags: master, slave, myself, fail, ...* if it is a slave, the node id of the master* time of the last pending ping still waiting for a reply.* time of the last pong received.* configuration epoch for this node (see the cluster specification).* status of the link to this node.* slots served...
#添加节点的测试，需求：把6382这个节点作为slave添加到6379这个master节点上#命令如下：（两个命令都可行）./redis-trib.rb add-node --slave 192.168.1.25:6382 192.168.1.25:6379./redis-trib.rb add-node --slave master-id 5dfa7d9f2dbb4ab64b71f2713538125354c88462 192.168.1.25:6382 192.168.1.25:6379---------------------------------------------------------------------------------------#检测集群节点相关信息redis-cli -p 6379 -h 192.168.1.25 cluster nodes---------------------------------------------------------------------------------------c31fb6f027db72876cb36e606b93817736e0ce81 192.168.1.25:6382 slave 5dfa7d9f2dbb4ab64b71f2713538125354c88462 0 1416901290565 1 connected47b1c1df820ff5e9057a526f69e9fe4fdd9800e9 192.168.1.25:6381 master - 0 1416901293638 3 connected 10923-16383c97f77ebcba3f48b9da07329b0e4e2943ded88c9 192.168.1.25:6380 master - 0 1416901291590 2 connected 5461-109225dfa7d9f2dbb4ab64b71f2713538125354c88462 192.168.1.25:6379 myself,master - 0 0 1 connected 0-5460#根据官方文档列表参数说明的第4行说明#5dfa7d9f2dbb4ab64b71f2713538125354c88462就是master的nodeid#那么我们再看6379节点的nodeid正好也是5dfa7d9f2dbb4ab64b71f2713538125354c88462
#如果需要改变主从关系，比如将6383这个服务节点从master(6380)转变成master(6379)，那么如下redis-cli -p 6383 -h 192.168.1.25192.168.1.25:6383> cluster replicate 5dfa7d9f2dbb4ab64b71f2713538125354c88462#5dfa7d9f2dbb4ab64b71f2713538125354c88462是master(6379)的nodeid