keepalived+MySQL双主实现高可用性

使用MySQL主主复制技术+Keepalived是一种简单、便捷的解决方案,在高可用集群环境中,keepalived使用虚拟VIP,使用Keepalived自带的服务监控功能和自定义脚本来实现MySQL故障时自动切换,非常灵活。如果有一台MySQL服务器死机,或工作出现故障,keepalived将检测到,并将有故障的MySQL服务器从系统中去除。当MySQL服务器工作正常时,则自动将MySQL服务器加入到服务器集群中,无需人工干预。下面环境为debugo01、debugo02两个节点,使用CentOS 6.6系统。MySQL server的版本是Percona Server 5.6.21,已经配置好了两边的MySQL实例。

1. 配置MySQL Multi-Master Replication

编辑my.cnf文件,添加相关复制选项。

vim /etc/my.cnf
server-id=1002
log-bin=mysql-bin
relay-log=mysqld-relay-bin
binlog_format=mixed
replicate_ignore_db=information_schema
auto-increment-increment=2
auto-increment-offset=1
 
#debugo01
mysql> SHOW master STATUS \G
*************************** 1. ROW ***************************
             File: mysql-bin.000001
         POSITION: 648
     Binlog_Do_DB: 
 Binlog_Ignore_DB: 
Executed_Gtid_Set: 
1 ROW IN SET (0.01 sec)
mysql> GRANT replication slave,replication client ON *.* TO rep@'%' IDENTIFIED BY 'debugo';
mysql> FLUSH privileges;
#debugo02
mysql> GRANT replication slave,replication client ON *.* TO rep@'%' IDENTIFIED BY 'debugo';
mysql> FLUSH privileges;
mysql> CHANGE MASTER TO MASTER_HOST='debugo01', MASTER_USER='rep', MASTER_PASSWORD='debugo', MASTER_LOG_FILE='mysql-bin.000001', MASTER_LOG_POS=648;
mysql> START slave;
Query OK, 0 ROWS affected (0.01 sec)
 
#debugo01
mysql> CHANGE MASTER TO MASTER_HOST='debugo02', MASTER_USER='rep', MASTER_PASSWORD='debugo', MASTER_LOG_FILE='mysql-bin.000001', MASTER_LOG_POS=426;
mysql> START slave;

检查日志,没有错误说明配置成功。

2. 配置keepalived

CentOS 6.6的光盘中已经带了keepalived-1.2.13-4的软件包,通过yum安装即可:

yum install keepalived

完成后编辑keepalived.conf文件

vim /etc/keepalived/keepalived.conf
! Configuration File for keepalived
global_defs {
   router_id Keepalived_MySQL
}
 
vrrp_script check_run {
    script "/etc/keepalived/check_MySQL.sh"
    interval 5
}
 
vrrp_sync_group VG1{
    group {
        VI_1
    }
}
 
vrrp_instance VI_1 {
    state BACKUP
    interface eth1
    virtual_router_id 51
    priority 100
    advert_int 1
    nopreempt
    track_script {
        check_run
    }
    authentication {
        auth_type PASS
        auth_pass debugo
    }
    virtual_ipaddress {
        10.6.2.200
    }
}

配置说明(参考官方文档,Keepalived+LVS 集群的部署和keepalived.conf 文件详解)
keepalived 的全局配置文件对keepalived起效的配置,而VRRPD 是keepalived 的核心,LVS配置只是需要使用keepalived来配置和管理LVS时需要使用,当然如果是只用keepalived 配置高可用集群是,LVS模块就不需要配置。配置文件都是以块的形式组织的,每个模块的配置都需要用 { } 的范围之内
在keepalived 的主配置文件中主要分为三个文本模块分别是:
全局模块(Global Configuration),VRRPD配置模块(VRRP 实例模块),LVS模块(虚拟服务模块);

1. 全局配置(Global Configuration) 包括两个部分: 即全局定义(global definition) 和静态地址路由(Static ipaddress /routes)。 全局定义:主要是设置Keeplived 的通知机制和标识,同时也不建议用该方法进行状态监控,进行状态监控时nagios 是不二的选择。静态地址路由:该处是配置静态地址和静态路由的,换句话说不随着Vrrp instance的开关的变化而变化,当然VIP不是static IPaddress,会随着VRRPD而添加和删除,设置静态地址和静态路由的格式如:src $SRC_IP to $DST_IP dev $SRC_DEVICE 或者src $SRC_IP to $DST_IP via $GW dev $SRC_DEVICE2. VRRPD 配置包括三个部分: 分别是VRRP脚本(VRRP scripts) . VRRP同步组(syncgriuzation group) 和VRRP 实例(VRRP Instance):

2. VRRP scripts: 脚本会安周期性执行,每秒( 或者按照自己设定的时间间隔),被监控的实例会记录退出的代码,这些脚本被至少有被监控的实例的权重为o的实例才能执行,因此任意脚本可能的声明都不会被系统所记录,如果没有特殊的设置,权重等于2,就意味着那些被监控的实例的优先级会被增加2,反过来,如果失败,则那些实例的优先级会被减去2

3. syncgriuzation group: 如果不使用 VRRP Sync Groups 如果keepalived 主机有两个网段,每个网段开启一个VRRP 实例,如果对外的网段出现问题,VRRPD认为自己仍然认为健康,因此 Master和Backup 相互切换,从而导致服务不能正常使用,同时高可用集群也不能正常运行,Sync group 就是为了解决该问题
注意:
router_id : 标识当前节点.两个节点即可以相同,也可以不同。
virtual_router_id : 这个标识是同一个vrrp实例使用唯一的标识。即同一个vrrp_stance,MASTER和BACKUP的virtual_router_id是一致的,同时在整个vrrp内是唯一的。
而priority 一般MASTER可以设置大一些。
vrrp_script中,MySQL的检测脚本如下,当检测到MySQL失败后,virtual_ipaddress会飘逸到BACKUP上。

vim /etc/keepalived/check_MySQL.sh
#!/bin/bash
MYSQL=/usr/bin/mysql
MYSQL_HOST=localhost
MYSQL_USER=root
CHECK_COUNT=5
 
counter=1
while true
do
        mysql -h $MYSQL_HOST -u $MYSQL_USER -e "show status;"  >/dev/null 2>&1
        i=$?
        ps aux | grep mysqld | grep -v grep > /dev/null 2>&1
        j=$?
        if [ $i = 0 ] && [ $j = 0 ]
        then
                exit 0
        else
                if [ $i = 1 ] && [ $j = 0 ]
                then
                        exit 0
                else
                        if [ $counter -gt $CHECK_COUNT ]
                        then
                                break
                        fi
                let counter++
                continue
                fi
        fi
        done
/etc/init.d/keepalived stop
exit 1

3. 测试

启动keepalived

service keepalived start

此时,在/var/log/messages中可以看到定期check_run脚本执行成功:

.....
Jan 21 19:58:35 debugo02 Keepalived_vrrp[22065]: VRRP_Script(check_run) succeeded
......

下面关闭debugo01(MASTER)的MySQL

service mysql stop

#debugo01 keepalived log
an 21 20:00:23 debugo01 Keepalived[25202]: Stopping Keepalived v1.2.13 (10/15,2014)
Jan 21 20:00:23 debugo01 Keepalived_vrrp[25204]: VRRP_Instance(VI_1) sending 0 priority
Jan 21 20:00:23 debugo01 Keepalived_vrrp[25204]: VRRP_Instance(VI_1) removing protocol VIPs.

#debugo02 keepalived log
Jan 21 19:58:35 debugo02 Keepalived_vrrp[22065]: VRRP_Script(check_run) succeeded
Jan 21 20:00:24 debugo02 Keepalived_vrrp[22065]: VRRP_Instance(VI_1) Transition to MASTER STATE
Jan 21 20:00:24 debugo02 Keepalived_vrrp[22065]: VRRP_Group(VG1{) Syncing instances to MASTER state
Jan 21 20:00:25 debugo02 Keepalived_vrrp[22065]: VRRP_Instance(VI_1) Entering MASTER STATE
Jan 21 20:00:25 debugo02 Keepalived_vrrp[22065]: VRRP_Instance(VI_1) setting protocol VIPs.
Jan 21 20:00:25 debugo02 Keepalived_vrrp[22065]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth1 for 10.6.2.200
Jan 21 20:00:25 debugo02 Keepalived_healthcheckers[22064]: Netlink reflector reports IP 10.6.2.200 added
Jan 21 20:00:30 debugo02 Keepalived_vrrp[22065]: VRRP_Instance(VI_1) Sending gratuitous ARPs on eth1 for 10.6.2.20

检查IP地址:

#debugo02
ip addr show
......
    inet 10.6.2.200/32 scope global eth1
......

还没有评论,快来抢沙发!

发表评论

  • 😉
  • 😐
  • 😡
  • 😈
  • 🙂
  • 😯
  • 🙁
  • 🙄
  • 😛
  • 😳
  • 😮
  • emoji-mrgree
  • 😆
  • 💡
  • 😀
  • 👿
  • 😥
  • 😎
  • ➡
  • 😕
  • ❓
  • ❗
  • 67 queries in 0.424 seconds