通过Keepalived实现Redis Failover自动故障切换功能
目前，Redis还没有一个类似于MySQL Proxy或Oracle RAC的官方HA方案。 Redis作者有一个名为Redis Sentinel的计划(http://redis.io/topics/sentinel)，据称将会有监控，报警和自动故障转移三大功能，非常不错。 但可惜的是短期内恐怕还不能开发完成。
因此，如何在出现故障时自动转移是一个需要解决的问题。
通过对网上一些资料的搜索，有建议采用HAProxy或Keepalived来实现的，事实上如果是做Failover而非负载均衡的话，Keepalived的效率肯定是超过HAProxy的，所以我决定采用Keepalived的方案。
环境介绍： Master: 10.6.1.143 Slave: 10.6.1.144 Virtural IP Address (VIP): 10.6.1.200
设计思路： 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，同时关闭主从复制功能； 当 Master 恢复正常，则从Slave同步数据，同步数据之后关闭主从复制功能，恢复Master身份，于此同时Slave等待Master同步数据完成之后，恢复Slave身份。 然后依次循环。
需要注意的是，这样做需要在Master与Slave上都开启本地化策略，否则在互相自动切换的过程中，未开启本地化的一方会将另一方的数据清空，造成数据完全丢失。
下面，是具体的实施步骤：
在Master和Slave上安装Keepalived $ sudo apt-get install keepalived
修改Master和Slave的/etc/hosts文件 $ sudo vim /etc/hosts
127.0.0.1 localhost 10.6.1.143 redis 10.6.1.144 redis-slave 默认安装完成keepalived之后是没有配置文件的，因此我们需要手动创建：
首先，在Master上创建如下配置文件： $ sudo vim /etc/keepalived/keepalived.conf
vrrp_script chk_redis {
script "/etc/keepalived/scripts/redis_check.sh" ###监控脚本
interval 2 ###监控时间
}
vrrp_instance VI_1 {
state MASTER ###设置为MASTER
interface eth0 ###监控网卡
virtual_router_id 51
priority 101 ###权重值
authentication {
auth_type PASS ###加密
auth_pass redis ###密码
}
track_script {
chk_redis ###执行上面定义的chk_redis
}
virtual_ipaddress {
10.6.1.200 ###VIP
}
notify_master /etc/keepalived/scripts/redis_master.sh
notify_backup /etc/keepalived/scripts/redis_backup.sh
notify_fault /etc/keepalived/scripts/redis_fault.sh
notify_stop /etc/keepalived/scripts/redis_stop.sh
}
然后，在Slave上创建如下配置文件：
$ sudo vim /etc/keepalived/keepalived.conf
vrrp_script chk_redis {
script "/etc/keepalived/scripts/redis_check.sh" ###监控脚本
interval 2 ###监控时间
}
vrrp_instance VI_1 {
state BACKUP ###设置为BACKUP
interface eth0 ###监控网卡
virtual_router_id 51
priority 100 ###比MASTRE权重值低
authentication {
auth_type PASS
auth_pass redis ###密码与MASTRE相同
}
track_script {
chk_redis ###执行上面定义的chk_redis
}
virtual_ipaddress {
10.6.1.200 ###VIP
}
notify_master /etc/keepalived/scripts/redis_master.sh
notify_backup /etc/keepalived/scripts/redis_backup.sh
notify_fault /etc/keepalived/scripts/redis_fault.sh
notify_stop /etc/keepalived/scripts/redis_stop.sh
}
在Master和Slave上创建监控Redis的脚本 $ sudo mkdir /etc/keepalived/scripts $ sudo vim /etc/keepalived/scripts/redis_check.sh
#!/bin/bash ALIVE=`/opt/redis/bin/redis-cli PING` if [ "$ALIVE" == "PONG" ]; then echo $ALIVE exit 0 else echo $ALIVE exit 1 fi
编写以下负责运作的关键脚本： notify_master /etc/keepalived/scripts/redis_master.sh notify_backup /etc/keepalived/scripts/redis_backup.sh notify_fault /etc/keepalived/scripts/redis_fault.sh notify_stop /etc/keepalived/scripts/redis_stop.sh
因为Keepalived在转换状态时会依照状态来呼叫： 当进入Master状态时会呼叫notify_master 当进入Backup状态时会呼叫notify_backup 当发现异常情况时进入Fault状态呼叫notify_fault 当Keepalived程序终止时则呼叫notify_stop
首先，在Redis Master上创建notity_master与notify_backup脚本： $ sudo vim /etc/keepalived/scripts/redis_master.sh
#!/bin/bash REDISCLI="/opt/redis/bin/redis-cli" LOGFILE="/var/log/keepalived-redis-state.log" echo "[master]" >> $LOGFILE date >> $LOGFILE echo "Being master...." >> $LOGFILE 2>&1 echo "Run SLAVEOF cmd ..." >> $LOGFILE $REDISCLI SLAVEOF 10.6.1.144 6379 >> $LOGFILE 2>&1 sleep 10 #延迟10秒以后待数据同步完成后再取消同步状态 echo "Run SLAVEOF NO ONE cmd ..." >> $LOGFILE $REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1
$ sudo vim /etc/keepalived/scripts/redis_backup.sh
#!/bin/bash REDISCLI="/opt/redis/bin/redis-cli" LOGFILE="/var/log/keepalived-redis-state.log" echo "[backup]" >> $LOGFILE date >> $LOGFILE echo "Being slave...." >> $LOGFILE 2>&1 sleep 15 #延迟15秒待数据被对方同步完成之后再切换主从角色 echo "Run SLAVEOF cmd ..." >> $LOGFILE $REDISCLI SLAVEOF 10.6.1.144 6379 >> $LOGFILE 2>&1
接着，在Redis Slave上创建notity_master与notify_backup脚本：
$ sudo vim /etc/keepalived/scripts/redis_master.sh
#!/bin/bash REDISCLI="/opt/redis/bin/redis-cli" LOGFILE="/var/log/keepalived-redis-state.log" echo "[master]" >> $LOGFILE date >> $LOGFILE echo "Being master...." >> $LOGFILE 2>&1 echo "Run SLAVEOF cmd ..." >> $LOGFILE $REDISCLI SLAVEOF 10.6.1.143 6379 >> $LOGFILE 2>&1 sleep 10 #延迟10秒以后待数据同步完成后再取消同步状态 echo "Run SLAVEOF NO ONE cmd ..." >> $LOGFILE $REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1 $ sudo vim /etc/keepalived/scripts/redis_backup.sh#!/bin/bash REDISCLI="/opt/redis/bin/redis-cli" LOGFILE="/var/log/keepalived-redis-state.log" echo "[backup]" >> $LOGFILE date >> $LOGFILE echo "Being slave...." >> $LOGFILE 2>&1 sleep 15 #延迟15秒待数据被对方同步完成之后再切换主从角色 echo "Run SLAVEOF cmd ..." >> $LOGFILE $REDISCLI SLAVEOF 10.6.1.143 6379 >> $LOGFILE 2>&1然后在Master与Slave创建如下相同的脚本： $ sudo vim /etc/keepalived/scripts/redis_fault.sh
#!/bin/bash LOGFILE=/var/log/keepalived-redis-state.log echo "[fault]" >> $LOGFILE date >> $LOGFILE$ sudo vim /etc/keepalived/scripts/redis_stop.sh
#!/bin/bash LOGFILE=/var/log/keepalived-redis-state.log echo "[stop]" >> $LOGFILE date >> $LOGFILE给脚本都加上可执行权限： $ sudo chmod +x /etc/keepalived/scripts/*.sh
脚本创建完成以后，我们开始按照如下流程进行测试： 1.启动Master上的Redis $ sudo /etc/init.d/redis start
2.启动Slave上的Redis $ sudo /etc/init.d/redis start
3.启动Master上的Keepalived $ sudo /etc/init.d/keepalived start
4.启动Slave上的Keepalived $ sudo /etc/init.d/keepalived start
5.尝试通过VIP连接Redis: $ redis-cli -h 10.6.1.200 INFO
连接成功，Slave也连接上来了。 role:master slave0:10.6.1.144,6379,online
6.尝试插入一些数据： $ redis-cli -h 10.6.1.200 SET Hello Redis OK
从VIP读取数据 $ redis-cli -h 10.6.1.200 GET Hello "Redis"
从Master读取数据 $ redis-cli -h 10.6.1.143 GET Hello "Redis"
从Slave读取数据 $ redis-cli -h 10.6.1.144 GET Hello "Redis"
下面，模拟故障产生： 将Master上的Redis进程杀死： $ sudo killall -9 redis-server
查看Master上的Keepalived日志 $ tailf /var/log/keepalived-redis-state.log [fault] Thu Sep 27 08:29:01 CST 2012
同时Slave上的日志显示： $ tailf /var/log/keepalived-redis-state.log [master] Fri Sep 28 14:14:09 CST 2012 Being master.... Run SLAVEOF cmd ... OK Run SLAVEOF NO ONE cmd ... OK
然后我们可以发现，Slave已经接管服务，并且担任Master的角色了。 $ redis-cli -h 10.6.1.200 INFO $ redis-cli -h 10.6.1.144 INFO role:master
然后我们恢复Master的Redis进程 $ sudo /etc/init.d/redis start
查看Master上的Keepalived日志 $ tailf /var/log/keepalived-redis-state.log [master] Thu Sep 27 08:31:33 CST 2012 Being master.... Run SLAVEOF cmd ... OK Run SLAVEOF NO ONE cmd ... OK
同时Slave上的日志显示： $ tailf /var/log/keepalived-redis-state.log [backup] Fri Sep 28 14:16:37 CST 2012 Being slave.... Run SLAVEOF cmd ... OK
可以发现目前的Master已经再次恢复了Master的角色，故障切换以及自动恢复都成功了。
