keepalived+mysql雙主

keepalived+mysql雙主原理：

（1）主庫(kù)1上的keepalived啟動(dòng)之后，會(huì)檢查mysql服務(wù)是否活著，如果活著，keepalived進(jìn)入master狀態(tài)，獲得VIP；

（2）主庫(kù)2上的keepalived啟動(dòng)之后，也會(huì)檢查mysql是否活著，然后檢查keepalived組內(nèi)是否有master狀態(tài)，如果有，則主庫(kù)2上的keepalived進(jìn)入backup狀態(tài)，處于隨時(shí)接管VIP狀態(tài)；

（3）如果主庫(kù)1上的mysql掛了，keepalived進(jìn)入fault狀態(tài)，釋放VIP，主庫(kù)2上的keepalived會(huì)變成master狀態(tài)，獲得VIP；

實(shí)驗(yàn)環(huán)境：

OS：CentOS release 6.6 (Final)

數(shù)據(jù)庫(kù)：mysql 5.7.14

A: master ：192.168.91.23

B: slave ：192.168.91.22 

VIP：192.168.91.100

操作系統(tǒng)時(shí)間一致更改：

date -s "20170227 16:25"

hwclock --systohc

主從參數(shù)：

A：

server_id = 330623

gtid_mode=ON

log_slave_updates = 0

enforce_gtid_consistency = ON

auto_increment_offset =1

auto_increment_increment =2

B：

server_id = 330622

gtid_mode=ON

log_slave_updates = 0

enforce_gtid_consistency = ON

auto_increment_offset=2

auto_increment_increment=2

配置AB互為主從：

A：

創(chuàng)建復(fù)制賬戶：

create user rep@'192.168.91.%' identified by '147258';

grant replication slave on *.* to rep@'192.168.91.%';

把A做個(gè)全備，還原到B上（這里省略不寫(xiě)）

B：添加A為B的主庫(kù)：

change master to master_host='192.168.91.23', master_port=3306, master_user='rep',master_password='147258', master_auto_position=1;

start slave；

A：添加B為A的主庫(kù)：

change master to master_host='192.168.91.22', master_port=3306, master_user='rep',master_password='147258', master_auto_position=1;

start slave；

創(chuàng)建一個(gè)監(jiān)控賬戶：（后面checkMySQL.py 腳本會(huì)用到，用于檢測(cè)mysql數(shù)據(jù)庫(kù)狀態(tài)，這個(gè)用戶只要有usage權(quán)限即可）

GRANT REPLICATION CLIENT ON *.* TO 'monitor'@'%' IDENTIFIED BY 'm0n1tor';

Ａ和Ｂ都要安裝keepalived軟件：

 yum install keepalived -y

 yum install MySQL-python -y

A的keepalived配置文件：

[root@Darren1 keepalived]#cat << EOF > keepalived.conf

vrrp_script vs_mysql_23 {         #可以根據(jù)實(shí)際情況命名

    script "/etc/keepalived/checkMySQL.py -h 192.168.91.23 -P 3306"

    interval 60      #切換時(shí)間

}

vrrp_instance VI_23 {          #可以根據(jù)實(shí)際情況命名

    state BACKUP               #剛開(kāi)始時(shí)使其處于backup狀態(tài)

    nopreempt                  #設(shè)置為不搶占，m1掛了，m2接管VIP,m1重啟不會(huì)自動(dòng)搶回VIP

    interface eth0             #VIP用的網(wǎng)卡

    virtual_router_id 23       #路由id，范圍是0-255，不能和路由器高可用的id一樣，同一集群中該數(shù)值要相同

    priority 100               #優(yōu)先級(jí)，同一個(gè)vrrp_instance的MASTER優(yōu)先級(jí)必須比BACKUP高。

    advert_int 5

    authentication {

        auth_type PASS   #認(rèn)證加密

        auth_pass 1111   # 認(rèn)證密碼，但密碼不要超過(guò) 8 位

    }

    track_script {

       vs_mysql_23     #調(diào)用這個(gè)腳本，返回0就持有VIP，返回1就釋放VIP

    }

    virtual_ipaddress {

        192.168.11.100     #VIP地址

    }

}

EOF

B配置文件：

[root@Darren2 keepalived]# cat << EOF > keepalived.conf

vrrp_script vs_mysql_22 {

    script "/etc/keepalived/checkMySQL.py -h 192.168.91.22 -P 3306"    #此處和A不同，其他都相同

    interval 60

}

vrrp_instance VI_22 {

    state BACKUP

    nopreempt

    interface eth0

    virtual_router_id 23

    priority 90

    advert_int 5

    authentication {

        auth_type PASS

        auth_pass 1111

    }

    track_script {

       vs_mysql_22

    }

    virtual_ipaddress {

        192.168.91.100

    }

}

EOF

checkMySQL.py腳本作用（這里省略不寫(xiě)）：

腳本的作用是判斷mysql進(jìn)程是否存在，如果存在返回0，如果不存在返回1；

A和B啟用keepalived

/etc/init.d/keepalived start   （開(kāi)始開(kāi)的時(shí)候，A和B誰(shuí)先啟動(dòng)，VIP就先在誰(shuí)上）

 chkconfig –level 2345 keepalived on

keepalived啟動(dòng)過(guò)程：

此時(shí)A開(kāi)啟keepalived服務(wù)：

[root@Darren1 ~]# /etc/init.d/keepalived start

[root@Darren1 ~]# tail -f /var/log/messages       

May  9 14:41:05 Darren1 Keepalived[28172]: Starting Keepalived v1.2.13 (03/19,2015)

May  9 14:41:05 Darren1 Keepalived[28173]: Starting Healthcheck child process, pid=28175

May  9 14:41:05 Darren1 Keepalived[28173]: Starting VRRP child process, pid=28176

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Netlink reflector reports IP 192.168.91.23 added

May  9 14:41:05 Darren1 Keepalived_healthcheckers[28175]: Netlink reflector reports IP 192.168.91.23 added

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Netlink reflector reports IP fe80::20c:29ff:fe56:5380 added

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Registering Kernel netlink reflector

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Registering Kernel netlink command channel

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Registering gratuitous ARP shared channel

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Opening file '/etc/keepalived/keepalived.conf'.

May  9 14:41:05 Darren1 Keepalived_healthcheckers[28175]: Netlink reflector reports IP fe80::20c:29ff:fe56:5380 added

May  9 14:41:05 Darren1 Keepalived_healthcheckers[28175]: Registering Kernel netlink reflector

May  9 14:41:05 Darren1 Keepalived_healthcheckers[28175]: Registering Kernel netlink command channel

May  9 14:41:05 Darren1 Keepalived_healthcheckers[28175]: Opening file '/etc/keepalived/keepalived.conf'.

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Configuration is using : 62873 Bytes

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: Using LinkWatch kernel netlink reflector...

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) Entering BACKUP STATE

May  9 14:41:05 Darren1 Keepalived_healthcheckers[28175]: Configuration is using : 5173 Bytes

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]

May  9 14:41:05 Darren1 Keepalived_healthcheckers[28175]: Using LinkWatch kernel netlink reflector...

May  9 14:41:05 Darren1 Keepalived_vrrp[28176]: VRRP_Script(vs_mysql_23) succeeded

May  9 14:41:21 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) Transition to MASTER STATE

May  9 14:41:26 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) Entering MASTER STATE

May  9 14:41:26 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) setting protocol VIPs.

May  9 14:41:26 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) Sending gratuitous ARPs on eth0 for 192.168.91.100

May  9 14:41:26 Darren1 Keepalived_healthcheckers[28175]: Netlink reflector reports IP 192.168.91.100 added

May  9 14:41:31 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) Sending gratuitous ARPs on eth0 for 192.168.91.100

總結(jié)啟動(dòng)過(guò)程：

（1）啟動(dòng)keepalived三個(gè)進(jìn)程，分別是主進(jìn)程，健康檢查子進(jìn)程，VRRP協(xié)議子進(jìn)程；

（2）啟動(dòng)結(jié)束后，VRRP_Instance開(kāi)始進(jìn)入backup狀態(tài)；

（3）進(jìn)入backup成功后，VRRP_Instance轉(zhuǎn)變狀態(tài)為master，然后進(jìn)入master狀態(tài)；

（4）獲取VIP，并且用ARP廣播告訴其他服務(wù)器；

keepalived切換過(guò)程：

停止A的mysql服務(wù)：

[root@Darren1 ~]# /etc/init.d/mysqld stop

Shutting down MySQL............ SUCCESS!

此時(shí)A的變化： 

[root@Darren1 ~]# tail -f /var/log/messages

May  9 14:43:25 Darren1 Keepalived_vrrp[28176]: VRRP_Script(vs_mysql_23) failed

May  9 14:43:26 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) Entering FAULT STATE

May  9 14:43:26 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) removing protocol VIPs.

May  9 14:43:26 Darren1 Keepalived_vrrp[28176]: VRRP_Instance(VI_23) Now in FAULT state

May  9 14:43:26 Darren1 Keepalived_healthcheckers[28175]: Netlink reflector reports IP 192.168.91.100 removed

總結(jié)A的變化：VRRP_Instance進(jìn)入fault狀態(tài)，釋放VIP；

此時(shí)B的變化： 

[root@Darren2 ~]# tail -f /var/log/messages       

May  9 14:43:26 Darren2 Keepalived_vrrp[35138]: VRRP_Instance(VI_22) Transition to MASTER STATE

May  9 14:43:31 Darren2 Keepalived_vrrp[35138]: VRRP_Instance(VI_22) Entering MASTER STATE

May  9 14:43:31 Darren2 Keepalived_vrrp[35138]: VRRP_Instance(VI_22) setting protocol VIPs.

May  9 14:43:31 Darren2 Keepalived_vrrp[35138]: VRRP_Instance(VI_22) Sending gratuitous ARPs on eth0 for 192.168.91.100

May  9 14:43:31 Darren2 Keepalived_healthcheckers[35137]: Netlink reflector reports IP 192.168.91.100 added

May  9 14:43:36 Darren2 Keepalived_vrrp[35138]: VRRP_Instance(VI_22) Sending gratuitous ARPs on eth0 for 192.168.91.100

總結(jié)B的變化：VRRP_Instance進(jìn)入master狀態(tài)，獲得VIP，ARP廣播通知；

使用VIP登陸數(shù)據(jù)庫(kù)：

創(chuàng)建登陸用戶：

create user 'keepalived'@'%' identified by '147258';

grant all on *.* to 'keepalived'@'%';

此時(shí)VIP在B上：

[root@Darren2 ~]# ip addr |grep 192

    inet 192.168.91.22/24 brd 192.168.91.255 scope global eth0

    inet 192.168.91.100/32 scope global eth0

在A上用'keepalived'@'%'賬戶登陸，是可以登陸成功的，證明此VIP是有效的：

[root@Darren1 ~]# mysql -ukeepalived -p147258 -h292.168.91.100

keepalived@192.168.91.100 [(none)]>select user(),current_user();

+--------------------------+----------------+

| user()                   | current_user() |

+--------------------------+----------------+

| keepalived@192.168.91.23 | keepalived@%   |

+--------------------------+----------------+

總結(jié)：

幾種VIP切換情況：

（1）m1主機(jī)宕機(jī)，VIP會(huì)切換到m2;

（2）m1上的mysql掛了，VIP會(huì)切換到m2;

（3）m1上的keepalived服務(wù)掛了，又分為兩種情況：

/etc/init.d/keepalived stop：正常切換

kill -9 keepalived_pid：因?yàn)閗eepalived是直接退出，m1和m2都有VIP,但是連接時(shí)候只有一個(gè)是生效的；

（4）腦裂的情況，m1和m2都各自認(rèn)為自己是master狀態(tài)，搶占VIP，VIP一會(huì)在m1上，一會(huì)在m2上；

在同一個(gè)交換機(jī)下不存在腦裂情況，這個(gè)在比較復(fù)雜的網(wǎng)絡(luò)環(huán)境中會(huì)發(fā)生。

可以在腳本中防范：ping一下網(wǎng)關(guān)，如果連網(wǎng)關(guān)都ping不通，vrrp_script腳本就放回1，keepalived進(jìn)入fault狀態(tài)；

m1掛了，m2接管VIP，m1修復(fù)好了之后怎么辦？

（1）如果是GTID復(fù)制直接把m1change master to m2上，如果是傳統(tǒng)復(fù)制，需要找到binlog位置；

（2）等待m1同步完成；

（3）啟動(dòng)keepalived；

keepalived+mysql雙主缺點(diǎn)和對(duì)應(yīng)方法：

（1）數(shù)據(jù)庫(kù)一致性難保障：

可以使用增強(qiáng)半同步，把主庫(kù)等待從庫(kù)回應(yīng)的參數(shù)rpl_semi_sync_master_timeout 調(diào)大點(diǎn)，

出現(xiàn)master的日志沒(méi)能實(shí)時(shí)的傳到slave上，需要手工把binlog截取出來(lái)補(bǔ)到從庫(kù)上；如果系統(tǒng)不存在了，可以通過(guò)binlog server 補(bǔ)日志；

（2）需要手動(dòng)把出現(xiàn)故障的主庫(kù)加入到原來(lái)的結(jié)構(gòu)中；