keepalived实现redis主从高可用

meikkiie

方案说明
两台机器（称为A和B），以统一的VIP对外提供服务
1.正常情况下，A和B都启动，B会把A的数据同步过来（B is slave of A）
2.当A挂了后，VIP漂移到B；B的keepalived 通知redis 执行：slaveof no one，由B提供服务
3.当A起来后，VIP不切换，仍在B上面；而A的keepalived 通知redis 执行slaveof B，开始把B的数据同步过来（A is slave of B）
4.依此循环
也就是说：
当两台机器都正常时，一个为master，一个为slave；
当master挂掉时，slave升级为master；
当原master再次起来时，它不会抢占，而是作为slave；依此循环

机器
10.75.201.67:/home/redis
10.75.201.66: /home/redis
初次启动时，10.75.201.67为master，10.75.201.66为slave
配置

//A机器
vrrp_script chk_nutcraker {
script "/home/redis/redis/bin/redis_check.sh"
interval 2
}
vrrp_instance VI_2 {
state BACKUP        #both BACKUP
interface eth1
virtual_router_id 12
priority 101    #101 on master, 100 on backup
nopreempt       #both nopreempt
track_script {
chk_nutcraker
}
virtual_ipaddress {
10.75.201.3
}
notify /home/redis/redis/bin/redis_notify.sh
}


//B机器
vrrp_script chk_nutcraker {
script "/home/redis/redis/bin/redis_check.sh"
interval 2
}
vrrp_instance VI_2 {
state BACKUP        #both BACKUP
interface eth1
virtual_router_id 12
priority 100    #101 on master, 100 on backup
nopreempt       #both nopreempt
track_script {
chk_nutcraker
}
virtual_ipaddress {
10.75.201.3
}
notify /home/redis/redis/bin/redis_notify.sh
}


检测redis是否正常的脚本：
[iyunv@redhat1 ~]# cat /home/redis/redis/bin/redis_check.sh

#!/bin/bash
ALIVE=`/home/redis/redis/bin/redis-cli PING`
if [ "$ALIVE" == "PONG" ]; then
echo $ALIVE
exit 0
else
echo $ALIVE
exit 1
fi


redis根据keepalived状态转移而更新自己的状态：进入master时slaveof no one，进入backup时，slaveof另一台机器：
[iyunv@redhat1 ~]# cat /home/redis/redis/bin/redis_notify.sh

#!/bin/bash
REMOTE=10.75.201.67  #在10.75.201.67机器上则为10.75.201.66
PORT=6379
REDIS_HOME=/home/redis/redis
REDISCLI="$REDIS_HOME/bin/redis-cli"
LOGFILE="$REDIS_HOME/logs/redis_notify.log"
DATE_TIME=`date "+%Y-%m-%d %H:%M:%S"`
TYPE=$1
NAME=$2
STATE=$3
case $STATE in
"MASTER")
echo "$DATE_TIME:became master:" >> $LOGFILE
$REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1
exit 0
;;
"BACKUP")  
echo "$DATE_TIME:became slave:" >> $LOGFILE
$REDISCLI SLAVEOF $REMOTE $PORT >> $LOGFILE  2>&1
exit 0
;;
"FAULT")  
echo "$DATE_TIME:[fault]" >> $LOGFILE
exit 0
;;
*)        echo "unknown state"
exit 1
;;
Esac