Keepalived實現Redis Failover自動故障切換
參考資料：
http://patrick-tang.blogspot.com/2012/06/redis-keepalived-failover-system.html
http://deidara.blog.51cto.com/400447/302402
背景介紹：
目前，Redis還沒有一個類似於MySQL Proxy或Oracle RAC的官方HA方案。
Redis作者有一個名爲Redis Sentinel的計劃(http://redis.io/topics/sentinel)，據稱將會有監控，報警和自動故障轉移三大功能，非常不錯。
但可惜的是短期內恐怕還不能開發完成。
因此，如何在出現故障時自動轉移是一個需要解決的問題。
通過對網上一些資料的搜索，有建議採用HAProxy或Keepalived來實現的，事實上如果是做Failover而非負載均衡的話，Keepalived的效率肯定是超過HAProxy的，所以我決定採用Keepalived的方案。
環境介紹：
Master: 10.6.1.143
Slave: 10.6.1.144
Virtural IP Address (VIP): 10.6.1.200
設計思路：
當 Master 與 Slave 均運作正常時, Master負責服務，Slave負責Standby；
當 Master 掛掉，Slave 正常時, Slave接管服務，同時關閉主從複製功能；
當 Master 恢復正常，則從Slave同步數據，同步數據之後關閉主從複製功能，恢復Master身份，於此同時Slave等待Master同步數據完成之後，恢復Slave身份。
然後依次循環。
需要注意的是，這樣做需要在Master與Slave上都開啓本地化策略，否則在互相自動切換的過程中，未開啓本地化的一方會將另一方的數據清空，造成數據完全丟失。
下面，是具體的實施步驟：
在Master和Slave上安裝Keepalived
$ sudo apt-get install keepalived
修改Master和Slave的/etc/hosts文件
$ sudo vim /etc/hosts
1
127.0.0.1 localhost
2
10.6.1.143 redis
3
10.6.1.144 redis-slave
默認安裝完成keepalived之後是沒有配置文件的，因此我們需要手動創建：
首先，在Master上創建如下配置文件：
$ sudo vim /etc/keepalived/keepalived.conf
01
vrrp_script chk_redis {
02
script "/etc/keepalived/scripts/redis_check.sh" ###監控腳本
03
interval 2 ###監控時間
04
}
05
vrrp_instance VI_1 {
06
state MASTER ###設置爲MASTER
07
interface eth0 ###監控網卡
08
virtual_router_id 51
09
priority 101 ###權重值
10
authentication {
11
auth_type PASS ###加密
12
auth_pass redis ###密碼
13
}
14
track_script {
15
chk_redis ###執行上面定義的chk_redis
16
}
17
virtual_ipaddress {
18
10.6.1.200 ###VIP
19
}
20
notify_master /etc/keepalived/scripts/redis_master.sh
21
notify_backup /etc/keepalived/scripts/redis_backup.sh
22
notify_fault /etc/keepalived/scripts/redis_fault.sh
23
notify_stop /etc/keepalived/scripts/redis_stop.sh
24
}
然後，在Slave上創建如下配置文件：
$ sudo vim /etc/keepalived/keepalived.conf
01
vrrp_script chk_redis {
02
script "/etc/keepalived/scripts/redis_check.sh" ###監控腳本
03
interval 2 ###監控時間
04
}
05
vrrp_instance VI_1 {
06
state BACKUP ###設置爲BACKUP
07
interface eth0 ###監控網卡
08
virtual_router_id 51
09
priority 100 ###比MASTRE權重值低
10
authentication {
11
auth_type PASS
12
auth_pass redis ###密碼與MASTRE相同
13
}
14
track_script {
15
chk_redis ###執行上面定義的chk_redis
16
}
17
virtual_ipaddress {
18
10.6.1.200 ###VIP
19
}
20
notify_master /etc/keepalived/scripts/redis_master.sh
21
notify_backup /etc/keepalived/scripts/redis_backup.sh
22
notify_fault /etc/keepalived/scripts/redis_fault.sh
23
notify_stop /etc/keepalived/scripts/redis_stop.sh
24
}
在Master和Slave上創建監控Redis的腳本
$ sudo mkdir /etc/keepalived/scripts
$ sudo vim /etc/keepalived/scripts/redis_check.sh
01
#!/bin/bash
02
03
ALIVE=`/opt/redis/bin/redis-cli PING`
04
if [ "$ALIVE" == "PONG" ]; then
05
echo $ALIVE
06
exit 0
07
else
08
echo $ALIVE
09
exit 1
10
fi
編寫以下負責運作的關鍵腳本：
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
因爲Keepalived在轉換狀態時會依照狀態來呼叫：
當進入Master狀態時會呼叫notify_master
當進入Backup狀態時會呼叫notify_backup
當發現異常情況時進入Fault狀態呼叫notify_fault
當Keepalived程序終止時則呼叫notify_stop
首先，在Redis Master上創建notity_master與notify_backup腳本：
$ sudo vim /etc/keepalived/scripts/redis_master.sh
01
#!/bin/bash
02
03
REDISCLI="/opt/redis/bin/redis-cli"
04
LOGFILE="/var/log/keepalived-redis-state.log"
05
06
echo "[master]" >> $LOGFILE
07
date >> $LOGFILE
08
echo "Being master...." >> $LOGFILE 2>&1
09
10
echo "Run SLAVEOF cmd ..." >> $LOGFILE
11
$REDISCLI SLAVEOF 10.6.1.144 6379 >> $LOGFILE 2>&1
12
sleep 10 #延遲10秒以後待數據同步完成後再取消同步狀態
13
14
echo "Run SLAVEOF NO ONE cmd ..." >> $LOGFILE
15
$REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1
$ sudo vim /etc/keepalived/scripts/redis_backup.sh
01
#!/bin/bash
02
03
REDISCLI="/opt/redis/bin/redis-cli"
04
LOGFILE="/var/log/keepalived-redis-state.log"
05
06
echo "[backup]" >> $LOGFILE
07
date >> $LOGFILE
08
echo "Being slave...." >> $LOGFILE 2>&1
09
10
sleep 15 #延遲15秒待數據被對方同步完成之後再切換主從角色
11
echo "Run SLAVEOF cmd ..." >> $LOGFILE
12
$REDISCLI SLAVEOF 10.6.1.144 6379 >> $LOGFILE 2>&1
接着，在Redis Slave上創建notity_master與notify_backup腳本：
$ sudo vim /etc/keepalived/scripts/redis_master.sh
01
#!/bin/bash
02
03
REDISCLI="/opt/redis/bin/redis-cli"
04
LOGFILE="/var/log/keepalived-redis-state.log"
05
06
echo "[master]" >> $LOGFILE
07
date >> $LOGFILE
08
echo "Being master...." >> $LOGFILE 2>&1
09
10
echo "Run SLAVEOF cmd ..." >> $LOGFILE
11
$REDISCLI SLAVEOF 10.6.1.143 6379 >> $LOGFILE 2>&1
12
sleep 10 #延遲10秒以後待數據同步完成後再取消同步狀態
13
14
echo "Run SLAVEOF NO ONE cmd ..." >> $LOGFILE
15
$REDISCLI SLAVEOF NO ONE >> $LOGFILE 2>&1
$ sudo vim /etc/keepalived/scripts/redis_backup.sh
01
#!/bin/bash
02
03
REDISCLI="/opt/redis/bin/redis-cli"
04
LOGFILE="/var/log/keepalived-redis-state.log"
05
06
echo "[backup]" >> $LOGFILE
07
date >> $LOGFILE
08
echo "Being slave...." >> $LOGFILE 2>&1
09
10
sleep 15 #延遲15秒待數據被對方同步完成之後再切換主從角色
11
echo "Run SLAVEOF cmd ..." >> $LOGFILE
12
$REDISCLI SLAVEOF 10.6.1.143 6379 >> $LOGFILE 2>&1
然後在Master與Slave創建如下相同的腳本：
$ sudo vim /etc/keepalived/scripts/redis_fault.sh
1
#!/bin/bash
2
3
LOGFILE=/var/log/keepalived-redis-state.log
4
5
echo "[fault]" >> $LOGFILE
6
date >> $LOGFILE
$ sudo vim /etc/keepalived/scripts/redis_stop.sh
1
#!/bin/bash
2
3
LOGFILE=/var/log/keepalived-redis-state.log
4
5
echo "[stop]" >> $LOGFILE
6
date >> $LOGFILE
給腳本都加上可執行權限：
$ sudo chmod +x /etc/keepalived/scripts/*.sh
腳本創建完成以後，我們開始按照如下流程進行測試：
1.啓動Master上的Redis
$ sudo /etc/init.d/redis start
2.啓動Slave上的Redis
$ sudo /etc/init.d/redis start
3.啓動Master上的Keepalived
$ sudo /etc/init.d/keepalived start
4.啓動Slave上的Keepalived
$ sudo /etc/init.d/keepalived start
5.嘗試通過VIP連接Redis:
$ redis-cli -h 10.6.1.200 INFO
連接成功，Slave也連接上來了。
role:master
slave0:10.6.1.144,6379,online
6.嘗試插入一些數據：
$ redis-cli -h 10.6.1.200 SET Hello Redis
OK
從VIP讀取數據
$ redis-cli -h 10.6.1.200 GET Hello
"Redis"
從Master讀取數據
$ redis-cli -h 10.6.1.143 GET Hello
"Redis"
從Slave讀取數據
$ redis-cli -h 10.6.1.144 GET Hello
"Redis"
下面，模擬故障產生：
將Master上的Redis進程殺死：
$ sudo killall -9 redis-server
查看Master上的Keepalived日誌
$ tailf /var/log/keepalived-redis-state.log
[fault]
Thu Sep 27 08:29:01 CST 2012
同時Slave上的日誌顯示：
$ 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
然後我們可以發現，Slave已經接管服務，並且擔任Master的角色了。
$ redis-cli -h 10.6.1.200 INFO
$ redis-cli -h 10.6.1.144 INFO
role:master
然後我們恢復Master的Redis進程
$ sudo /etc/init.d/redis start
查看Master上的Keepalived日誌
$ 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
同時Slave上的日誌顯示：
$ tailf /var/log/keepalived-redis-state.log
[backup]
Fri Sep 28 14:16:37 CST 2012
Being slave....
Run SLAVEOF cmd ...
OK
可以發現目前的Master已經再次恢復了Master的角色，故障切換以及自動恢復都成功了。