CentOS中搭建Redis伪分布式集群
解压redis
先到官网https://redis.io/下载redis安装包，然后在CentOS操作系统中解压该安装包：
tar -zxvf redis-3.2.9.tar.gz
编译redis
cd redis-3.2.9
make && make install
此时如果报如下错误：
bash: make: command not found
说明系统没有安装make，所有要先安装make，即：
yum -y install gcc automake autoconf libtool make
如果此时再报如下错误：
Error Downloading Packages:
glibc-headers-2.12-1.107.el6.x86_64: failure: Packages/glibc-headers-2.12-1.107.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
1:perl-Module-Pluggable-3.90-129.el6.x86_64: failure: Packages/perl-Module-Pluggable-3.90-129.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
mpfr-2.4.1-6.el6.x86_64: failure: Packages/mpfr-2.4.1-6.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
1:make-3.81-20.el6.x86_64: failure: Packages/make-3.81-20.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
cloog-ppl-0.15.7-1.2.el6.x86_64: failure: Packages/cloog-ppl-0.15.7-1.2.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
gcc-4.4.7-3.el6.x86_64: failure: Packages/gcc-4.4.7-3.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
libgomp-4.4.7-3.el6.x86_64: failure: Packages/libgomp-4.4.7-3.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
kernel-headers-2.6.32-358.el6.x86_64: failure: Packages/kernel-headers-2.6.32-358.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
autoconf-2.63-5.1.el6.noarch: failure: Packages/autoconf-2.63-5.1.el6.noarch.rpm from c6-media: [Errno 256] No more mirrors to try.
automake-1.11.1-4.el6.noarch: failure: Packages/automake-1.11.1-4.el6.noarch.rpm from c6-media: [Errno 256] No more mirrors to try.
4:perl-libs-5.10.1-129.el6.x86_64: failure: Packages/perl-libs-5.10.1-129.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
3:perl-version-0.77-129.el6.x86_64: failure: Packages/perl-version-0.77-129.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
4:perl-5.10.1-129.el6.x86_64: failure: Packages/perl-5.10.1-129.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
1:perl-Pod-Simple-3.13-129.el6.x86_64: failure: Packages/perl-Pod-Simple-3.13-129.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
glibc-devel-2.12-1.107.el6.x86_64: failure: Packages/glibc-devel-2.12-1.107.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
1:perl-Pod-Escapes-1.04-129.el6.x86_64: failure: Packages/perl-Pod-Escapes-1.04-129.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
libtool-2.2.6-15.5.el6.x86_64: failure: Packages/libtool-2.2.6-15.5.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
ppl-0.10.2-11.el6.x86_64: failure: Packages/ppl-0.10.2-11.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
cpp-4.4.7-3.el6.x86_64: failure: Packages/cpp-4.4.7-3.el6.x86_64.rpm from c6-media: [Errno 256] No more mirrors to try.
说明需要配置本地yum源或者让机器可以上网即可，如果需要配置本地yum源，可以参考另一篇文章《CentOS配置本地yum源》
安装好make后，在redis-3.2.9目录下，再执行
make && make install
出现如下信息，说明执行成功了：
Hint: It's a good idea to run 'make test'
make[1]: Leaving directory `/opt/redis-3.2.9/src'
cd src && make install
make[1]: Entering directory `/opt/redis-3.2.9/src'
Hint: It's a good idea to run 'make test'
INSTALL install
INSTALL install
INSTALL install
INSTALL install
INSTALL install
make[1]: Leaving directory `/opt/redis-3.2.9/src'
此时redis-3.2.9目录下，多了一个src文件夹，里面内容如下：
此时编译redis完成。
创建redis节点
由于要让redis集群正常运行，就至少需要3个主节点，所以在此就在一台主机上创建6个redis节点来演示集群配置，实际生产环境中需要每个节点一台主机。同时对应的redis节点的ip和端口对应关系如下：
192.168.242.10 6380
192.168.242.10 6381
192.168.242.10 6382
192.168.242.10 6383
192.168.242.10 6384
192.168.242.10 6385
在redis-3.2.9目录下，新建一个redis_cluster目录，用于存放各位节点的配置信息：
mkdir redis_cluster
同时在redis_cluster目录下分别新建6380、6381、6382、6383、6384、6385文件夹：
mkdir 6380
mkdir 6381
mkdir 6382
mkdir 6383
mkdir 6384
mkdir 6385
在6380文件夹下，新建redis.conf文件：
touch redis.conf
编辑如下内容：
daemonize yes
pidfile redis_6380.pid
logfile redis_6380.log
appendonly yes
bind 192.168.242.10
port 6380
cluster-enabled yes
cluster-config-file nodes-6380.conf
cluster-node-timeout 15000
cluster-slave-validity-factor 10
cluster-migration-barrier 1
cluster-require-full-coverage yes
集群相关的配置说明：
cluster-enabled 是否开启集群配置
cluster-config-file 集群节点集群信息配置文件，每个节点都有一个，由redis生成和更新，配置时避免名称冲突
cluster-node-timeout 集群节点互连超时的阀值，单位毫秒
cluster-slave-validity-factor 进行故障转移时，salve会 申请成为master。有时slave会和master失联很久导致数据较旧，这样的slave不应该成为master。这个配置用来判断slave是否和master失联时间过长。
同样在其它文件夹下，也新建redis.conf文件，里面的配置信息相应的修改即可。
启动各个节点：
首先进入redis-3.2.9目录下的，src目录，然后再执行启动各个节点：
./redis-server ../redis_cluster/6380/redis.conf
./redis-server ../redis_cluster/6381/redis.conf
./redis-server ../redis_cluster/6382/redis.conf
./redis-server ../redis_cluster/6383/redis.conf
./redis-server ../redis_cluster/6384/redis.conf
./redis-server ../redis_cluster/6385/redis.conf
检查各个节点是否正常启动了：
如果出现如下信息，说明6个节点正常启动了：
创建redis集群
redis官方提供了redis-trib.rb这个工具来创建redis集群，同时这个工具就在redis-3.2.9目录下的src目录中，执行如下命令：
./redis-trib.rb create --replicas 1 192.168.242.10:6380 192.168.242.10:6381 192.168.242.10:6382 192.168.242.10:6383 192.168.242.10:6384 192.168.242.10:6385
此时如果报如下错误，说明没有安装ruby:
/usr/bin/env: ruby: No such file or directory
所以先安装ruby：
yum install ruby
安装完成ruby后，再执行上面的创建集群的命令，此时如果再报如下错误，因为缺少redis和ruby的接口，使用gem安装：
./redis-trib.rb:24:in `require': no such file to load -- rubygems (LoadError)
from ./redis-trib.rb:24
由于安装redis和ruby的接口需要用gem，所以先装gem：
yum install rubygems
然而此时如果报如下错误：
Error Downloading Packages:
rubygems-1.3.7-1.el6.noarch: failure: Packages/rubygems-1.3.7-1.el6.noarch.rpm from c6-media: [Errno 256] No more mirrors to try.
说明在镜像文件中没有rubygems的安装包，此时只能先将rubygems的安装下载下来，然后在安装。
这里将rubygems-1.3.7-5.el6.noarch.rpm文件上传到CentOS中，然后执行安装命令：
rpm -ivh rubygems-1.3.7-5.el6.noarch.rpm
此时会报如下错误：
warning: rubygems-1.3.7-5.el6.noarch.rpm: Header V3 RSA/SHA1 Signature, key ID c105b9de: NOKEY
error: Failed dependencies:
ruby-rdoc is needed by rubygems-1.3.7-5.el6.noarch
说明需要先将rubygems依赖的软件先安装好，即将ruby-irb和ruby-rdoc先安装：
yum install ruby-irb ruby-rdoc
然后再执行上面安装rubygems-1.3.7-5.el6.noarch.rpm的命令，出现如下信息，说明安装成功了：
再来安装redis和ruby的接口：
gem install redis
出现如下信息，说明安装成功：
此时再来执行创建redis集群的命令：
./redis-trib.rb create --replicas 1 192.168.242.10:6380 192.168.242.10:6381 192.168.242.10:6382 192.168.242.10:6383 192.168.242.10:6384 192.168.242.10:6385
出现如下信息，说明redis集群启动成功：
>>> Creating cluster
>>> Performing hash slots allocation on 6 nodes...
Using 3 masters:
192.168.242.10:6380
192.168.242.10:6381
192.168.242.10:6382
Adding replica 192.168.242.10:6383 to 192.168.242.10:6380
Adding replica 192.168.242.10:6384 to 192.168.242.10:6381
Adding replica 192.168.242.10:6385 to 192.168.242.10:6382
M: cd9a1cfb495fd728ebf86fe3c2ae484d33a55efa 192.168.242.10:6380
slots:0-5460 (5461 slots) master
M: 86ff5dd6c91f2b1f723d20e0611b6b000b5564b9 192.168.242.10:6381
slots:5461-10922 (5462 slots) master
M: c7124e5837d89a108d4fedce46d18d190d831355 192.168.242.10:6382
slots:10923-16383 (5461 slots) master
S: 1eae195a6d3628b7ac4c2152518cd0afd2dcc4a2 192.168.242.10:6383
replicates cd9a1cfb495fd728ebf86fe3c2ae484d33a55efa
S: 3dd781d5161ac8367cbf16b97414f7ecbb271fd1 192.168.242.10:6384
replicates 86ff5dd6c91f2b1f723d20e0611b6b000b5564b9
S: a942e13b48d63e54f22425a4694fb6ba71b7be89 192.168.242.10:6385
replicates c7124e5837d89a108d4fedce46d18d190d831355
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 192.168.242.10:6380)
M: cd9a1cfb495fd728ebf86fe3c2ae484d33a55efa 192.168.242.10:6380
slots:0-5460 (5461 slots) master
1 additional replica(s)
S: a942e13b48d63e54f22425a4694fb6ba71b7be89 192.168.242.10:6385
slots: (0 slots) slave
replicates c7124e5837d89a108d4fedce46d18d190d831355
S: 1eae195a6d3628b7ac4c2152518cd0afd2dcc4a2 192.168.242.10:6383
slots: (0 slots) slave
replicates cd9a1cfb495fd728ebf86fe3c2ae484d33a55efa
M: 86ff5dd6c91f2b1f723d20e0611b6b000b5564b9 192.168.242.10:6381
slots:5461-10922 (5462 slots) master
1 additional replica(s)
M: c7124e5837d89a108d4fedce46d18d190d831355 192.168.242.10:6382
slots:10923-16383 (5461 slots) master
1 additional replica(s)
S: 3dd781d5161ac8367cbf16b97414f7ecbb271fd1 192.168.242.10:6384
slots: (0 slots) slave
replicates 86ff5dd6c91f2b1f723d20e0611b6b000b5564b9
[OK] All nodes agree about slots configuration.
>>> Check for open slots...
>>> Check slots coverage...
[OK] All 16384 slots covered.
验证redis集群
直接通过redis-cli工具来连接集群，该redis-cli工具也是在src目录下的。
./redis-cli -h 192.168.242.10 -c -p 6380
注意：
需要加参数 -c 可连接到集群，-p的端口只要是集群中某个节点的端口就行。
连接成功后，测试往redis集群中存入数据和取数据：
CentOS中搭建Redis伪分布式集群
