MySQL keepalived 双主.md

MySQL keepalived 双主搭建

环境说明

系统	IP	主机名	mysql	keepalived	VIP
CentOS 6.8	192.168.197.61	C6-node1	5.6.36	1.2.13	192.168.197.88
CentOS 6.8	192.168.197.62	C6-node2	5.6.36	1.2.13	192.168.197.88

安装MySQL双主

C6-node1

MySQL的安装及初始化这里不做详细介绍，下面贴出node1的my.cnf的配置文件：

[mysqld]
datadir=/data/mysql
port=3306
socket=/tmp/mysql.sock
pid-file=/data/mysql/mysqld.pid
log_error=error.log
user=mysql
skip-name-resolve
log-bin=mysql-bin
log-bin-index=mysql-bin.index
server-id=61
character_set_server=utf8
log-slave-updates=1
relay-log=mysql-relay-bin
relay-log-index=mysql-relay-bin.index
binlog_format=mixed
auto-increment-increment=2
auto-increment-offset=1
gtid_mode=on
enforce_gtid_consistency=1
skip_slave_start=1

[mysql]
prompt=(\\u@\\h) [\\d]>\\_

[client]
user=root
password=

创建复制帐号：

(rootc6-node1) [(none)]> GRANT REPLICATION SLAVE ON *.* TO 'mysync'@'192.168.197.62' IDENTIFIED BY 'redhat';
Query OK, 0 rows affected (0.05 sec)

C6-node2

[mysqld]
datadir=/data/mysql
port=3306
socket=/tmp/mysql.sock
pid-file=/data/mysql/mysqld.pid
log_error=error.log
user=mysql
skip-name-resolve
log-bin=mysql-bin
log-bin-index=mysql-bin.index
server-id=62
character_set_server=utf8
log-slave-updates=1
relay-log=mysql-relay-bin
relay-log-index=mysql-relay-bin.index
binlog_format=mixed
auto-increment-increment=2
auto-increment-offset=2
gtid_mode=on
enforce_gtid_consistency=1
skip_slave_start=1

[mysql]
prompt=(\u@\h) [\d]>\_

[client]
user=root
password=

创建复制帐号：

(root@c6-node2) [(none)]> GRANT REPLICATION SLAVE ON *.* TO 'mysync'@'192.168.197.61' IDENTIFIED BY 'redhat';
Query OK, 0 rows affected (0.00 sec)

双主配置

注意：由于这里配置的是双主模型所以两个数据库的数据一定要保持一致！如果要是不一致请以一个最新说句为主，注意连接的先后顺序！
node1配置:

(root@C6-node1) [(none)]> CHANGE MASTER TO MASTER_HOST='192.168.197.62',MASTER_USER='mysync',MASTER_PASSWORD='redhat',MASTER_PORT=3306,MASTER_AUTO_POSITION=1;

(root@C6-node1) [(none)]> START SLAVE;
Query OK, 0 rows affected (0.01 sec)

node2配置：

(root@C6-node2) [(none)]> CHANGE MASTER TO MASTER_HOST='192.168.197.61',MASTER_USER='mysync',MASTER_PASSWORD='redhat',MASTER_PORT=3306,MASTER_AUTO_POSITION=1;
Query OK, 0 rows affected, 2 warnings (0.03 sec)

(root@C6-node2) [(none)]> START SLAVE;
Query OK, 0 rows affected (0.03 sec)

keepalived 配置

node1安装配置

安装keepalived这里是用的是yum install -y keepalived进行安装。keepalived的配置文件如下：

global_defs {
   router_id KEEPALIVED_MYSQL
}

vrrp_instance VI_1 {
    state BACKUP
    interface eth0
    virtual_router_id 61
    priority 140
    nopreempt
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.197.88
    }
}
virtual_server 192.168.197.61 3306 {
     delay_loop 2
     protocol TCP
     real_server 192.168.197.61 3306 {
     notify_down "/etc/keepalived/check_mysql.sh"
     TCP_CHECK {
     connect_timeout 10
     nb_get_retry 3
     delay_before_retry 3
     connect_port 3306
     }
}
}

• virtual_router_id 取值在0-255之间，用来区分多个instance的VRRP组播。注意： 同一网段中virtual_router_id的值不能重复，否则会出错。
• priority 用来选举master的，要成为master，那么这个选项的值最好高于其他机器50个点，该项取值范围是1-255（在此范围之外会被识别成默认值100）。
• nopreempt：不主动抢占资源，只在master这台优先级高的设置，backup不设置。
• advert_int：发VRRP包的时间间隔，即多久进行一次master选举（可以认为是健康查检时间间隔）。
• delay_loop：延迟轮询时间（单位秒）。
• connect_timeout,nb_get_retry,delay_before_retry分别表示超时时长、重试次数，下次重试的时间延迟。

check_mysql 检测脚本如下：

#!/bin/bash
MYSQL=/usr/local/mysql/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 -eq 0 ] && [ $j -eq 0 ]
        then
                exit 0
        else
                if [ $i -eq 1 ] && [ $j -eq 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

说明：
1.check_mysql.sh 脚本的需要赋予执行权限。
2.check_mysql脚本没有使用密码登录获取mysql的状态，这里建议创建一个普通用户来检测mysql 的状态。
3.我这边没有使用keepalived中轮询算法的配置，在网上文章说明需要安装ipvsadm。

node2配置

global_defs {
   router_id KEEPALIVED_MYSQL
}
vrrp_instance VI_1 {
    state BACKUP
    interface eth0
    virtual_router_id 61
    priority 130
    advert_int 1
    authentication {
        auth_type PASS
        auth_pass 1111
    }
    virtual_ipaddress {
        192.168.197.88
    }
}

virtual_server 192.168.197.62 3306 {
     delay_loop 2
     protocol TCP
     real_server 192.168.197.62 3306 {
     notify_down "/etc/keepalived/check_mysql.sh"
     TCP_CHECK {
     connect_timeout 10
     nb_get_retry 3
     delay_before_retry 3
     connect_port 3306
     }
}
}

check_mysql.sh脚本和node1的相同。

启动并测试

1.启动时先保证MySQL双主模式配置是好的，即Slave_IO_Running: Yes和Slave_SQL_Running: Yes，这两个要是Yes状态。
2.keepalived启动时先启动priority 低的节点，然后在启动priority高的节点，避免出现抢占VIP。
3.注意，我在mysql双主模式中设置了skip_slave_start，即启动mysql不会自动执行start slave;而是需要手动执行！

测试

# hostname -I
192.168.197.61 192.168.197.88
# ip addr show  eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:9c:7a:30 brd ff:ff:ff:ff:ff:ff
    inet 192.168.197.61/24 brd 192.168.197.255 scope global eth0
    inet 192.168.197.88/32 scope global eth0
    inet6 fe80::20c:29ff:fe9c:7a30/64 scope link
       valid_lft forever preferred_lft forever
# mysql -u root -h 192.168.197.88 -p
Enter password:

(root@192.168.197.88) [(none)]> show variables like 'server_id';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| server_id     | 61    |
+---------------+-------+
1 row in set (0.00 sec)

(root@192.168.197.88) [(none)]> select * from test.DOTA;
+----+--------------+------------+
| Id | Name         | Short_name |
+----+--------------+------------+
|  2 | 敌法师       | DF         |
|  4 | 隐形刺客     | SA         |
|  5 | 沙王         | SK         |
|  7 | 骷髅王       | SNK        |
|  8 | 混沌骑士     | CK         |
|  9 | 冰魂         | AA         |
| 10 | 斧王         | FV         |
| 11 | 司夜刺客     | NA         |
| 12 | 全能骑士     | OK         |
+----+--------------+------------+
9 rows in set (0.00 sec)

停止node1上的mysql并测试：

# hostname -I
192.168.197.62 192.168.197.88
# ip addr show eth0
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether 00:0c:29:41:57:51 brd ff:ff:ff:ff:ff:ff
    inet 192.168.197.62/24 brd 192.168.197.255 scope global eth0
    inet 192.168.197.88/32 scope global eth0
    inet6 fe80::20c:29ff:fe41:5751/64 scope link
       valid_lft forever preferred_lft forever
# mysql -uroot -h192.168.197.88 -p
Enter password:
(root@192.168.197.88) [(none)]> show variables like 'server_id';
+---------------+-------+
| Variable_name | Value |
+---------------+-------+
| server_id     | 62    |
+---------------+-------+
1 row in set (0.00 sec)

(root@192.168.197.88) [(none)]> select * from test.DOTA;
+----+--------------+------------+
| Id | Name         | Short_name |
+----+--------------+------------+
|  2 | 敌法师       | DF         |
|  4 | 隐形刺客     | SA         |
|  5 | 沙王         | SK         |
|  7 | 骷髅王       | SNK        |
|  8 | 混沌骑士     | CK         |
|  9 | 冰魂         | AA         |
| 10 | 斧王         | FV         |
| 11 | 司夜刺客     | NA         |
| 12 | 全能骑士     | OK         |
+----+--------------+------------+
9 rows in set (0.00 sec)

说明：
1.以上配置的不足之处是node1在重新启动mysql和 keepalived时会抢占VIP，这个原因个人认为是我的检测脚本把keepalived 给关闭了而在重新启动时由于priority的关系还是会抢占。
2.网上参考的是主上配置nopreempt参数，但是测试的效果还是不行。如有懂得的还请指教。