Linux Redis集群搭建与集群客户端实现
硬件环境
本文适用的硬件环境如下
Linux版本：CentOS release 6.7 (Final) Redis版本：3.2.1
Redis已经成功安装，安装路径为/home/idata/yangfan/local/redis-3.2.1。
我们要在单台机器上搭建Redis集群，方式是通过不同的TCP端口启动多个实例，然后组成集群。
1、启动Redis多个实例
我们在Redis安装目录下创建目录cluster，并编写7000.conf~7005.conf 6个配置文件，这6个配置文件用来启动6个实例，后面将使用这6个实例组成集群。
以7000.conf为例，配置文件需要填写如下几项。
port 7000 //端口7000,7002,7003
bind 10.93.84.53 //默认ip为127.0.0.1 需要改为其他节点机器可访问的ip 否则创建集群时无法访问对应的端口，无法创建集群
daemonize yes //redis后台运行
pidfile ./redis_7000.pid //pidfile文件对应7000,7001,7002
cluster-enabled yes //开启集群 把注释#去掉
cluster-config-file nodes_7000.conf //集群的配置 配置文件首次启动自动生成 7000,7001,7002
cluster-node-timeout 15000 //请求超时 默认15秒，可自行设置
appendonly yes //aof日志开启 有需要就开启，它会每次写操作都记录一条日志
分别启动6个实例
./bin/redis-server cluster/conf/7000.conf
./bin/redis-server cluster/conf/7001.conf
./bin/redis-server cluster/conf/7002.conf
./bin/redis-server cluster/conf/7003.conf
./bin/redis-server cluster/conf/7004.conf
./bin/redis-server cluster/conf/7005.conf
启动成功后，看一下进程
# ps -ef | grep redis | grep cluster
idata 15711 22329 0 18:40 pts/10 00:00:00 ./bin/redis-server 10.93.84.53:7000 [cluster]
idata 15740 22329 0 18:40 pts/10 00:00:00 ./bin/redis-server 10.93.84.53:7001 [cluster]
idata 15810 22329 0 18:40 pts/10 00:00:00 ./bin/redis-server 10.93.84.53:7002 [cluster]
idata 17023 22329 0 18:42 pts/10 00:00:00 ./bin/redis-server 10.93.84.53:7003 [cluster]
idata 17030 22329 0 18:42 pts/10 00:00:00 ./bin/redis-server 10.93.84.53:7004 [cluster]
idata 17035 22329 0 18:42 pts/10 00:00:00 ./bin/redis-server 10.93.84.53:7005 [cluster]
至此，ip=10.93.84.53机器上创建了6个实例，端口号为port=7000~7005。
Redis 3.0以上的集群方式是通过Redis安装目录下的bin/redis-trib.rb脚本搭建。
这个脚本是用Ruby编写的，尝试运行，如果打印如下，你可以跳过本文的第二部分。
idata@qa-f1502-xg01.xg01:~/yangfan/local/redis-3.2.1/bin$ ruby redis-trib.rb Usage: redis-trib <command> <options> <arguments ...> create host1:port1 ... hostN:portN --replicas <arg> check host:port info host:port fix host:port --timeout <arg> reshard host:port --from <arg> --to <arg> --slots <arg> --yes --timeout <arg> --pipeline <arg> rebalance host:port --weight <arg> --auto-weights --use-empty-masters --timeout <arg> --simulate --pipeline <arg> --threshold <arg> add-node new_host:new_port existing_host:existing_port --slave --master-id <arg> del-node host:port node_id set-timeout host:port milliseconds call host:port command arg arg .. arg import host:port --from <arg> --copy --replace help (show this help) For check, fix, reshard, del-node, set-timeout you can specify the host and port of any working node in the cluster.
如果执行失败，那么不幸的是你的机器没有Ruby运行的环境，那么你需要安装Ruby。进入第二部分。
2、安装ruby
下面的过程都是在root权限下完成的。
1）yum安装ruby和依赖的包。
# yum -y install ruby ruby-devel rubygems rpm-build
一般来说，这一步是能正常完成的。
2）使用gem这个命令来安装redis接口
gem是ruby的一个工具包
# gem install redis
安装过程出错
郁闷，看样子要升级ruby版本。
3）升级Ruby的版本
安装rvm，我不知道这是个什么东西，但是感觉像是Ruby的一个包管理器。
# curl -L get.rvm.io | bash -s stable
WTF，又出问题了
气急败坏的照着他说的做
# gpg2 --recv-keys 409B6B1796C275462A1703113804BB82D39DC0E3 gpg: keyring `/root/.gnupg/secring.gpg' created gpg: requesting key D39DC0E3 from hkp server keys.gnupg.net gpg: /root/.gnupg/trustdb.gpg: trustdb created gpg: key D39DC0E3: public key "Michal Papis (RVM signing) <mpapis@gmail.com>" imported gpg: no ultimately trusted keys found gpg: Total number processed: 1 gpg: imported: 1 (RSA: 1)
然后重新下载rvm安装，成功了。
# curl -L get.rvm.io | bash -s stable % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 24090 100 24090 0 0 18023 0 0:00:01 0:00:01 --:--:-- 129k Downloading https://github.com/rvm/rvm/archive/1.29.3.tar.gz Downloading https://github.com/rvm/rvm/releases/download/1.29.3/1.29.3.tar.gz.asc gpg: Signature made Mon 11 Sep 2017 04:59:21 AM CST using RSA key ID BF04FF17 gpg: Good signature from "Michal Papis (RVM signing) <mpapis@gmail.com>" gpg: aka "Michal Papis <michal.papis@toptal.com>" gpg: aka "[jpeg image of size 5015]" gpg: WARNING: This key is not certified with a trusted signature! gpg: There is no indication that the signature belongs to the owner. Primary key fingerprint: 409B 6B17 96C2 7546 2A17 0311 3804 BB82 D39D C0E3 Subkey fingerprint: 62C9 E5F4 DA30 0D94 AC36 166B E206 C29F BF04 FF17 GPG verified '/usr/local/rvm/archives/rvm-1.29.3.tgz' Creating group 'rvm' Installing RVM to /usr/local/rvm/ Installation of RVM in /usr/local/rvm/ is almost complete: * First you need to add all users that will be using rvm to 'rvm' group, and logout - login again, anyone using rvm will be operating with `umask u=rwx,g=rwx,o=rx`. * To start using RVM you need to run `source /etc/profile.d/rvm.sh` in all your open shell windows, in rare cases you need to reopen all shell windows.
接着，source环境，让rvm可用。
# source /usr/local/rvm/scripts/rvm
查看Ruby可用版本
# rvm list known # MRI Rubies [ruby-]1.8.6[-p420] [ruby-]1.8.7[-head] # security released on head [ruby-]1.9.1[-p431] [ruby-]1.9.2[-p330] [ruby-]1.9.3[-p551] [ruby-]2.0.0[-p648] [ruby-]2.1[.10] [ruby-]2.2[.7] [ruby-]2.3[.4] [ruby-]2.4[.1]
可以看到最新的版本是2.4.1，那么我们装最新的吧。
# rvm install 2.4.1 Searching for binary rubies, this might take some time. No binary rubies available for: centos/6/x86_64/ruby-2.4.1. Continuing with compilation. Please read 'rvm help mount' to get more information on binary rubies. Checking requirements for centos. Installing requirements for centos. Installing required packages: libffi-devel, libyaml-devel...... Requirements installation successful. Installing Ruby from source to: /usr/local/rvm/rubies/ruby-2.4.1, this may take a while depending on your cpu(s)... ruby-2.4.1 - #downloading ruby-2.4.1, this may take a while depending on your connection... curl: (35) SSL connect error There was an error(35). Checking fallback: https://ftp.ruby-lang.org/pub/ruby/2.4/ruby-2.4.1.tar.bz2 % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 11.9M 100 11.9M 0 0 1753k 0 0:00:07 0:00:07 --:--:-- 2919k ruby-2.4.1 - #extracting ruby-2.4.1 to /usr/local/rvm/src/ruby-2.4.1.... ruby-2.4.1 - #applying patch /usr/local/rvm/patches/ruby/2.4.1/random_c_using_NR_prefix.patch. ruby-2.4.1 - #configuring.................................................................. ruby-2.4.1 - #post-configuration.. ruby-2.4.1 - #compiling.............................................................................................. ruby-2.4.1 - #installing......................... ruby-2.4.1 - #making binaries executable.. ruby-2.4.1 - #downloading rubygems-2.6.14 % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 751k 100 751k 0 0 443k 0 0:00:01 0:00:01 --:--:-- 628k No checksum for downloaded archive, recording checksum in user configuration. ruby-2.4.1 - #extracting rubygems-2.6.14.... ruby-2.4.1 - #removing old rubygems......... ruby-2.4.1 - #installing rubygems-2.6.14........................... ruby-2.4.1 - #gemset created /usr/local/rvm/gems/ruby-2.4.1@global ruby-2.4.1 - #importing gemset /usr/local/rvm/gemsets/global.gems............................................... ruby-2.4.1 - #generating global wrappers........ ruby-2.4.1 - #gemset created /usr/local/rvm/gems/ruby-2.4.1 ruby-2.4.1 - #importing gemsetfile /usr/local/rvm/gemsets/default.gems evaluated to empty gem list ruby-2.4.1 - #generating default wrappers........ ruby-2.4.1 - #adjusting #shebangs for (gem irb erb ri rdoc testrb rake). Install of ruby-2.4.1 - #complete Ruby was built without documentation, to build it run: rvm docs generate-ri
至此，我们升级了Ruby的版本。
4）安装gem redis接口，成功！
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 0 seconds 1 gem installed
5）安装rubygems，成功！
# yum install -y rubygems Loaded plugins: fastestmirror, security Setting up Install Process Loading mirror speeds from cached hostfile base | 3.7 kB 00:00 didi_jenkins_enable | 1.5 kB 00:00 didi_op_toa_enable | 1.5 kB 00:00 didi_txjenkins_enable | 1.5 kB 00:00 didi_update | 1.5 kB 00:00 epel | 4.3 kB 00:00 extras | 3.4 kB 00:00 tmprepo | 1.5 kB 00:00 updates | 3.4 kB 00:00 Package rubygems-1.3.7-5.el6.noarch already installed and latest version Nothing to do
至此，我们的Ruby和运行redis-trib.rb需要的环境安装完成了。
3、Redis集群搭建
有了Ruby执行环境，可以开始将之前的6个实例组建成集群了。
命令方式：
ruby ./bin/redis-trib.rb create --replicas 1 10.93.84.53:7000 10.93.84.53:7001 10.93.84.53:7002 10.93.84.53:7003 10.93.84.53:7004 10.93.84.53:7005
--replicas 1表示为集群的master节点创建1个副本。那么6个实例里，有三个master，有三个是slave。
后面跟上6个实例就好了，形式就是ip:port
执行情况：
# ruby ./bin/redis-trib.rb create --replicas 1 10.93.84.53:7000 10.93.84.53:7001 10.93.84.53:7002 10.93.84.53:7003 10.93.84.53:7004 10.93.84.53:7005 >>> Creating cluster >>> Performing hash slots allocation on 6 nodes... Using 3 masters: 10.93.84.53:7000 10.93.84.53:7001 10.93.84.53:7002 Adding replica 10.93.84.53:7003 to 10.93.84.53:7000 Adding replica 10.93.84.53:7004 to 10.93.84.53:7001 Adding replica 10.93.84.53:7005 to 10.93.84.53:7002 M: 6346ae8c7af7949658619fcf4021cc7aca454819 10.93.84.53:7000 slots:0-5460 (5461 slots) master M: 5ac973bceab0d486c497345fe884ff54d1bb225a 10.93.84.53:7001 slots:5461-10922 (5462 slots) master M: cc46a4a1c0ec3f621b6b5405c6c10b7cffe73932 10.93.84.53:7002 slots:10923-16383 (5461 slots) master S: 92f62ec93a0550d962f81213ca7e9b3c9c996afd 10.93.84.53:7003 replicates 6346ae8c7af7949658619fcf4021cc7aca454819 S: 942c9f97dc68198c39f425d13df0d8e3c40c5a58 10.93.84.53:7004 replicates 5ac973bceab0d486c497345fe884ff54d1bb225a S: a92a81532b63652bbd862be6f19a9bd8832e5e05 10.93.84.53:7005 replicates cc46a4a1c0ec3f621b6b5405c6c10b7cffe73932 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 10.93.84.53:7000) M: 6346ae8c7af7949658619fcf4021cc7aca454819 10.93.84.53:7000 slots:0-5460 (5461 slots) master 1 additional replica(s) S: a92a81532b63652bbd862be6f19a9bd8832e5e05 10.93.84.53:7005 slots: (0 slots) slave replicates cc46a4a1c0ec3f621b6b5405c6c10b7cffe73932 M: 5ac973bceab0d486c497345fe884ff54d1bb225a 10.93.84.53:7001 slots:5461-10922 (5462 slots) master 1 additional replica(s) S: 942c9f97dc68198c39f425d13df0d8e3c40c5a58 10.93.84.53:7004 slots: (0 slots) slave replicates 5ac973bceab0d486c497345fe884ff54d1bb225a S: 92f62ec93a0550d962f81213ca7e9b3c9c996afd 10.93.84.53:7003 slots: (0 slots) slave replicates 6346ae8c7af7949658619fcf4021cc7aca454819 M: cc46a4a1c0ec3f621b6b5405c6c10b7cffe73932 10.93.84.53:7002 slots:10923-16383 (5461 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.
可以看到16384个slot都已经创建完成，并且建立了3个master和对应的replica:
Using 3 masters: 10.93.84.53:7000 10.93.84.53:7001 10.93.84.53:7002 Adding replica 10.93.84.53:7003 to 10.93.84.53:7000 Adding replica 10.93.84.53:7004 to 10.93.84.53:7001 Adding replica 10.93.84.53:7005 to 10.93.84.53:7002
。。。
[OK] All 16384 slots covered.
4、验证集群状态
登录集群客户端，-c标识以集群方式登录
./bin/redis-cli -h 10.93.84.53 -p 7000 -c
查看集群状态
10.93.84.53:7000> 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:8 cluster_my_epoch:8 cluster_stats_messages_sent:215 cluster_stats_messages_received:215 10.93.84.53:7000> cluster nodes 942c9f97dc68198c39f425d13df0d8e3c40c5a58 10.93.84.53:7004 slave 5ac973bceab0d486c497345fe884ff54d1bb225a 0 1507806791940 5 connected 5ac973bceab0d486c497345fe884ff54d1bb225a 10.93.84.53:7001 master - 0 1507806788937 2 connected 5461-10922 a92a81532b63652bbd862be6f19a9bd8832e5e05 10.93.84.53:7005 slave cc46a4a1c0ec3f621b6b5405c6c10b7cffe73932 0 1507806790939 6 connected cc46a4a1c0ec3f621b6b5405c6c10b7cffe73932 10.93.84.53:7002 master - 0 1507806789937 3 connected 10923-16383 6346ae8c7af7949658619fcf4021cc7aca454819 10.93.84.53:7000 myself,slave 92f62ec93a0550d962f81213ca7e9b3c9c996afd 0 0 1 connected 92f62ec93a0550d962f81213ca7e9b3c9c996afd 10.93.84.53:7003 master - 0 1507806792941 8 connected 0-5460
一些原理
redis cluster在设计的时候，就考虑到了去中心化，去中间件，也就是说，集群中的每个节点都是平等的关系，都是对等的，每个节点都保存各自的数据和整个集群的状态。每个节点都和其他所有节点连接，而且这些连接保持活跃，这样就保证了我们只需要连接集群中的任意一个节点，就可以获取到其他节点的数据。
Redis集群没有并使用传统的一致性哈希来分配数据，而是采用另外一种叫做哈希槽（hash slot）的方式来分配的，一致性哈希对向集群中新增和删除实例的支持很好，但是哈希槽对向集群新增实例或者删除实例的话，需要额外的操作，需要手动的将slot重新平均的分配到新集群的实例中。
redis cluster 默认分配了 16384 个slot，当我们set一个key时，会用CRC16算法来取模得到所属的slot，然后将这个key分到哈希槽区间的节点上，具体算法就是：CRC16(key)%16384。
Redis 集群会把数据存在一个master节点，然后在这个master和其对应的salve之间进行数据同步。当读取数据时，也根据一致性哈希算法到对应的master节点获取数据。只有当一个master 挂掉之后，才会启动一个对应的salve节点，充当master。
需要注意的是：必须要3个或以上的主节点，否则在创建集群时会失败，并且当存活的主节点数小于总节点数的一半时，整个集群就无法提供服务了。
5、python集群客户端
以Python Redis Cluster集群的使用方式为例，简单说明一下如何使用，让大家更直观的了解一下Redis集群。
Redis集群方式与单机方式在python客户端实现上是有很大不同的。
- 包不同，单机依赖包redis-py，集群依赖包redis-py-cluster
- 对同时操作多个keys的命令(mset, mget, sinter, ...)，redis-py-cluster重写StriceRedis(单机Redis操作类)了其方法，而这些方法丧失了命令的原子性。
- Pipelines在集群中的表现也不同了。在单机版中，pipeline是批量batch提交redis批量执行的，但是在集群版中，是one by one提交redis执行完成直接返回，在客户端重新组合成一个列表返回。所以集群中只是看起来像是批量执行。
- 一些命令是会Fanout(扇形)发送到集群中各个shard中执行并汇总的，如ping，keys等。一些命令是不提倡使用(blocked)，如watch，unwatch等。
我只是简单翻译了官网的一些内容，相信信息大家可以参考：
http://redis-py-cluster.readthedocs.io/en/master/commands.html
http://redis-py-cluster.readthedocs.io/en/master/limitations-and-differences.html
1）安装redis-py-cluster
简单的通过pip安装redis-py-cluster包。如果安装失败，可以自助下载安装。
# pip install redis-py-cluster Collecting redis-py-cluster Downloading redis_py_cluster-1.3.4-py2.py3-none-any.whl Requirement already satisfied: redis>=2.10.2 in /home/idata/pythonEnv/idataPlatEnv/lib/python2.7/site-packages/redis-2.10.5-py2.7.egg (from redis-py-cluster) Installing collected packages: redis-py-cluster Successfully installed redis-py-cluster-1.3.4
2）一个简单的demo
封装了RedisCluster操作类，实现了一些方法，其实就是做了一层封装。
封装的意义是：我喜欢对这些封装增加一些装饰器，控制异常和重试等逻辑。
# -*- coding:utf-8 -*- from rediscluster import StrictRedisCluster redis_nodes = [ {'host': '10.93.84.53', 'port': 7000}, {'host': '10.93.84.53', 'port': 7001}, {'host': '10.93.84.53', 'port': 7002}, {'host': '10.93.84.53', 'port': 7003}, {'host': '10.93.84.53', 'port': 7004}, {'host': '10.93.84.53', 'port': 7005}, ] class RedisCluster(object): def __init__(self, redis_nodes): self.cluster = StrictRedisCluster(startup_nodes=redis_nodes) # 无差别的方法 def set(self, name, value, ex=None, px=None, nx=False, xx=False): return self.cluster.set(name, value, ex, px, nx, xx) # 无差别的方法 def get(self, name): return self.cluster.get(name) # 扇形发送的命令 def cluster_info(self): return self.cluster.cluster_info() # 重写StrictRedis的方法 def mset(self, *args, **kwargs): return self.cluster.mset(args, kwargs) # 重写StrictRedis的方法 def mget(self, keys, *args): return self.cluster.mget(keys, args) cluster = RedisCluster(redis_nodes) cluster.cluster_info()
