基于Keepalived的Redis主从高可用集群搭建

目标：

• 当 Master 与 Slave 均运作正常时, Master负责服务，Slave负责Standby；
• 当 Master 挂掉，Slave 正常时, Slave接管服务升级Master，有写权限，同时关闭主从复制功能；
• 当 Master 恢复正常, Master降级为Slave同步数据，开启主从复制。

安装机器：node00/01 192.168.5.1/192.168.5.2
vip: 192.168.5.3

1、安装&配置Redis

安装redis需要gcc编译环境，此处不赘述。以下操作在node00/01上进行
先将redis-5.0.9.tar.gz传送到服务器

$ tar -zxvf redis-5.0.9.tar.gz -C /usr/local/
$ cd redis-5.0.9
$ make && make install
# 开始配置Redis
$ mkdir /etc/redis
$ cp /usr/local/redis-5.0.9/redis.conf /etc/redis
$ vi /etc/redis/redis.conf
# 修改如下配置
bind 0.0.0.0   # 修改支持网络访问
protected-mode no   # 同样一个安全配置，默认yes是不能网络访问的
daemonize yes   # 后台运行redis
appendonly yes  # 持久化配置(开启AOF)
# 指定Master IP，同步（仅从配置）
$ cat >> /etc/redis/redis.conf << EOF
slaveof 192.168.5.1 6379  
EOF
# 配置redis.service开机自启动
$ vi /lib/systemd/system/redis.service
[Unit]
Description=Redis
After=network.target

[Service]
ExecStart=/usr/local/bin/redis-server /etc/redis/redis.conf --daemonize no
ExecStop=/usr/local/bin/redis-cli -h 127.0.0.1 -p 6379 shutdown

[Install]
WantedBy=multi-user.target

$ systemctl daemon-reload
$ systemctl enable redis && systemctl start redis

2、安装keepalived

以下操作在node0/1上进行

$ yum install keepalived -y
$ rm -fr /etc/keepalived/keepalived.conf 
$ vi /etc/keepalived/keepalived.conf 
! Configuration File for keepalived
vrrp_script chk_redis {
    script "/etc/keepalived/redis_check.sh"    ###监控脚本
    interval 2                         ###监控时间
    weight -20                         ##如果条件成立则权重减20
    }

vrrp_instance VI_1 {
    state BACKUP                ###都设置为BACKUP,非抢占式
    nopreempt                   # 不抢占MASTER
    interface ens32                ###监控网卡
    virtual_router_id 117
    priority 100            ###权重值,node1上设为90
    authentication {
        auth_type PASS        ###加密
        auth_pass 123        ###密码
    }
    #设置额外的监控，里面的任意一个网卡出现问题，都会进入FAULT状态
    track_interface {
       ens32
    }
    track_script {
         chk_redis            ###执行上面定义的chk_redis
    }
    virtual_ipaddress {
       192.168.5.3        ######VIP
    }
        notify_master /etc/keepalived/redis_master.sh
        notify_backup /etc/keepalived/redis_backup.sh
        notify_fault  /etc/keepalived/redis_fault.sh
        notify_stop  /etc/keepalived/redis_stop.sh
}

$ systemctl enable keepalived && systemctl restart keepalived

以下为4个检测脚本

#  定时检测脚本
$ vi redis_check.sh
#!/bin/bash

ALIVE=`/usr/local/bin/redis-cli PING`
# 当Redis服务宕机时立即停止keepalived
if [ "$ALIVE" != "PONG" ]; then
   systemctl stop keepalived
   exit 1
fi

# VIP不在本机,要保证为slave
COUNT=`ip a | grep 5.3 |wc -l`
if [ $COUNT -eq 0 ]; then   
   COUNT=`/usr/local/bin/redis-cli info replication | grep role:slave |wc -l`
   if [ $COUNT -eq 0 ]; then   
      /etc/keepalived/redis_slave.sh     
   fi
   exit 0
fi

# VIP主机上Redis角色不是Master，则要置主
COUNT=`/usr/local/bin/redis-cli info replication | grep role:master |wc -l`
if [ $COUNT -eq 0 ]; then   
   /etc/keepalived/redis_master.sh
   exit 0
fi


# 被选入Master时，执行的脚本
$ vi /etc/keepalived/redis_master.sh  
#!/bin/bash

REDISCLI="/usr/local/bin/redis-cli"
REDISCONF="/etc/redis/redis.conf"

LOGFILE="/var/log/keepalived-redis-state.log"
echo "[master]" >> $LOGFILE
date >> $LOGFILE
echo "Being master...." >> $LOGFILE 2>&1

echo "Run SLAVEOF NO ONE cmd ..." >> $LOGFILE
$REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1

# 检测服务不正常时，执行fault脚本
$ vi /etc/keepalived/redis_fault.sh
#!/bin/bash 

LOGFILE=/var/log/keepalived-redis-state.log
echo "[fault]" >> $LOGFILE
date >> $LOGFILE

# keepalive终止时写入日志
$ vi /etc/keepalived/redis_stop.sh 
#!/bin/bash 

LOGFILE=/var/log/keepalived-redis-state.log
echo "[stop]" >> $LOGFILE
date >> $LOGFILE

# 进入slave时执行的脚本
$ vi /etc/keepalived/redis_backup.sh 
#!/bin/bash 
 
REDISCLI="/usr/local/bin/redis-cli"
REDISCONF="/etc/redis/redis.conf"
LOGFILE="/var/log/keepalived-redis-state.log"
 
echo "[backup]" >> $LOGFILE
date >> $LOGFILE
echo "Being slave...." >> $LOGFILE 2>&1
echo "Run SLAVEOF cmd ..." >> $LOGFILE
# IP为指定另一个redis的IP,在node1上，这个值为192.168.5.2
$REDISCLI SLAVEOF 192.168.5.2 6379 >> $LOGFILE  2>&1  
$ chmod +x /etc/keepalived/*.sh

说明：在网上大多数的教程中，对redis_check.sh定时检测脚本处理得不够，可能会导致出现如下场景：Keepalived的VIP这台机上，Redis的角色是“从”，这样，客户端通过vip连接上Redis之后，读没有问题，但却无法写、删除了（出错提示：You can't write against a read only slave)。
因此，我们在检测脚本中实现了：
1）将Keepalived的vip与Redis的“主”角色绑定，
2）当Redis服务宕机时立即随之停止keepalived服务，即将此台机器从集群中剔除，这样，另外一台机器就必然自动置主。
3）当然，当这台机的Redis服务重新启动后，要手工启动keepalived服务，将它重新加回到集群。