使用Keepalived实现MySQL双主高可用

324次阅读

没有评论

共计 10965 个字符，预计需要花费 28 分钟才能阅读完成。

MySQL 双主配置

环境准备：

OS：CentOS7
master：192.168.1.10
backup：192.168.1.20
VIP：192.168.1.30

一、安装 MySQL 数据库.

在 master 和 backup 上安装 mysql, 安装完后自动启动，mysql root 密码为 123456

二、修改 MySQL 配置文件：

1.master 端配置文件如下：

# vim /etc/my.cnf #添加
server_id = 1 #backup 上设置为 2
log-bin = /data/mysql/mysql-bin
log-bin-index=/data/mysql/my-bin.index
binlog-ignore-db = mysql,information_schema #忽略写入 binlog 日志的库
auto-increment-increment = 2 #字段变化增量值
auto-increment-offset = 1 #初始字段 ID 为 1
slave-skip-errors = all #忽略所有复制产生的错误

# systemctl restart mysqld

2. backup 端配置文件如下：

master 端和 backup 端配置只有 server_id 不一样，别的都一致.

三、创建数据同步用户并查看 log bin 日志和 pos 位置：

1.> master 上创建 mysql 同步账号并查看 log bin 日志和 pos 位置：

# mysql -uroot -p123456

mysql> GRANT REPLICATION SLAVE ON *.* TO ‘repl’@’192.168.1.%’ IDENTIFIED BY ‘repl’;

mysql> flush privileges;

master 配置如下：

# mysql -uroot -p123456

mysql> change master to
-> master_host=’192.168.1.20′, #这里填 backup 的 IP
-> master_user=’repl’,
-> master_password=’repl’,
-> master_log_file=’mysql-bin.000001′,
-> master_log_pos=618;

mysql> start slave;

2.> backup 上创建 mysql 同步账号配置如下：

# mysql -uroot -p123456

mysql> GRANT REPLICATION SLAVE ON *.* TO ‘repl’@’192.168.1.%’ IDENTIFIED BY ‘repl’;

mysql> flush privileges;

mysql> change master to
-> master_host=’192.168.1.10′, #这里填 master 的 IP
-> master_user=’repl’,
-> master_password=’repl’,
-> master_log_file=’mysql-bin.000001′,
-> master_log_pos=618;

mysql> start slave;
———————

分别查看同步状态：

master 查看：

mysql> show slave status\G;
*************************** 1. row ***************************
Slave_IO_State: Waiting for master to send event
Master_Host: 192.168.1.20
Master_User: replication
Master_Port: 3306
Connect_Retry: 60
Master_Log_File: mysql-bin.000001
Read_Master_Log_Pos: 1082
Relay_Log_File: test2-relay-bin.000002
Relay_Log_Pos: 784
Relay_Master_Log_File: mysql-bin.000001
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
———————

backup 查看：

mysql> show slave status\G;
*************************** 1. row ***************************
Slave_IO_State: Waiting for master to send event
Master_Host: 192.168.1.10
Master_User: replication
Master_Port: 3306
Connect_Retry: 60
Master_Log_File: mysql-bin.000001
Read_Master_Log_Pos: 618
Relay_Log_File: test3-relay-bin.000002
Relay_Log_Pos: 320
Relay_Master_Log_File: mysql-bin.000001
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
———————

Slave_IO 和 Slave_SQL 是 YES 说明主主同步成功。

四、MySQL 主主同步测试

master 上插入数据测试：

mysql> create database testdb;

mysql> use testdb;

mysql> create table user (number INT(10),name VARCHAR(255));

mysql> insert into user values(01,’testid’);

mysql> show tables;
+—————-+
| Tables_in_test |
+—————-+
| user |
+—————-+
———————

backup 上查看：

mysql> use testdb;

mysql> show tables;
+—————-+
| Tables_in_test |
+—————-+
| user |
+—————-+

mysql> select number,name from user;
+——–+——+
| number | name |
+——–+——+
| 1 | testid |
+——–+——+
———————

可以看到已经成功同步过去，同样在 backup 插入到 user 表数据，一样同步过去，双主配置没有问题。

五、配置 keepalived 实现双机热备

1.master 安装 keepalived 并配置：

# yum install -y keepalived

# vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {
notification_email {
admin@test.com
}
notification_email_from admin@test.com
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id MYSQL_HA
}

vrrp_instance VI_1 {
state BACKUP
interface eth0 #根据实际网络接口进行更改
virtual_router_id 51
priority 100 #优先级，master 设置为 100
advert_int 1
nopreempt #不主动抢占资源，只在 master 上设置
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.30
}
}

virtual_server 192.168.1.30 3306 {
delay_loop 2
#lb_algo rr
#lb_kind NAT
persistence_timeout 50
protocol TCP

real_server 192.168.1.10 3306 {#检测本地 mysql
weight 3
notify_down /tmp/mysql.sh #当 mysql 服务 down 时，执行此脚本，杀死 keepalived 实现切换
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}

backup 安装 keepalived 并配置：

# yum install -y keepalived

# vim /etc/keepalived/keepalived.conf

! Configuration File for keepalived

global_defs {
notification_email {
admin@test.com
}
notification_email_from admin@test.com
smtp_server 127.0.0.1
smtp_connect_timeout 30
router_id MYSQL_HA
}

vrrp_instance VI_1 {
state BACKUP
interface eth0 #根据实际网络接口进行更改
virtual_router_id 51
priority 90 #优先级，backup 设置为 90
advert_int 1
#nopreempt #主动抢占资源
authentication {
auth_type PASS
auth_pass 1111
}
virtual_ipaddress {
192.168.1.30
}
}

virtual_server 192.168.1.30 3306 {
delay_loop 2
#lb_algo rr
#lb_kind NAT
persistence_timeout 50
protocol TCP

real_server 192.168.1.20 3306 {#检测本地 mysql
weight 3
notify_down /tmp/mysql.sh #当 mysql 服务 down 时，执行此脚本，杀死 keepalived 实现切换
TCP_CHECK {
connect_timeout 3
nb_get_retry 3
delay_before_retry 3
}
}
}

master 和 backup 上编辑 mysql.sh

# vim /tmp/mysql.sh

#!/bin/bash
pkill keepalived

# chmod +x !$
# systemctl start keepalived

两台 mysql 服务器授权允许 root 远程登录：

# mysql -uroot -p123456789

mysql> grant all on *.* to ‘root’@’192.168.1.%’ identified by ‘123456’;

mysql> flush privileges;

测试高可用
通过 mysql 客户端通过 VIP 连接，看是否连接成功。
这里我用同网段的另一台机器，连接测试：

# mysql -h192.168.1.30 -uroot -p123456

Type ‘help;’ or ‘\h’ for help. Type ‘\c’ to clear the current input statement.

可以看到，连接成功，且查询数据没有问题，停止 master 上 mysql 服务，是否能正常切换到 backup 上，可以使用 ip addr 命令来查看 VIP 在哪台服务器上。

master 上查看是否有 VIP, 可以看到 VIP 在 master 上

# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
link/ether 00:0c:29:cf:ab:c4 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.10/24 brd 192.168.1.255 scope global noprefixroute eth0
valid_lft forever preferred_lft forever
inet 192.168.1.30/32 scope global eth0
valid_lft forever preferred_lft forever
inet6 fe80::fe8e:3c2f:4d32:e9fd/64 scope link noprefixroute
valid_lft forever preferred_lft forever
———————

停掉 master 上 mysql 服务：

# systemctl stop mysqld

# ps axu |grep keepalived
root 11074 0.0 0.0 112708 988 pts/1 S+ 15:28 0:00 grep –color=autokeepalived

# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
link/ether 00:0c:29:cf:ab:c4 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.10/24 brd 192.168.1.255 scope global noprefixroute eth0
valid_lft forever preferred_lft forever
inet6 fe80::fe8e:3c2f:4d32:e9fd/64 scope link noprefixroute
valid_lft forever preferred_lft forever
———————

可以看到，keepalived 在 mysql 服务停掉之后也被停掉，VIP 不在 master 上。

backup 上查看是否有 VIP，可以看到 VIP 在 backup 上。

# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
link/ether 00:0c:29:33:80:d5 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.20/24 brd 192.168.1.255 scope global noprefixroute eth0
valid_lft forever preferred_lft forever
inet 192.168.1.30/32 scope global eth0
valid_lft forever preferred_lft forever
inet6 fe80::4b20:2e16:a957:f9a1/64 scope link noprefixroute
valid_lft forever preferred_lft forever
———————

查看 /var/log/messages 日志，可以看到主备切换过程：

Apr 8 15:27:16 hosts systemd: Stopping MySQL Server…
Apr 8 15:27:16 hosts Keepalived_healthcheckers[11048]: TCP connection to [192.168.1.10]:3306 failed.
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: TCP connection to [192.168.1.10]:3306 failed.
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: Check on service [192.168.1.10]:3306 failed after 1 retry.
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: Removing service [192.168.1.10]:3306 from VS [192.168.1.30]:3306
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: IPVS (cmd 1160, errno 2): No such destination
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: Executing [/tmp/mysql.sh] for service [192.168.1.10]:3306 in VS [192.168.1.30]:3306
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: Lost quorum 1-0=1 > 0 for VS [192.168.1.30]:3306
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: Remote SMTP server [127.0.0.1]:25 connected.
Apr 8 15:27:19 hosts Keepalived_vrrp[11049]: VRRP_Instance(VI_1) sent 0 priority
Apr 8 15:27:19 hosts Keepalived_vrrp[11049]: VRRP_Instance(VI_1) removing protocol VIPs.
Apr 8 15:27:19 hosts Keepalived[11047]: Stopping
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: IPVS (cmd 1156, errno 2): No such file or directory
Apr 8 15:27:19 hosts Keepalived_healthcheckers[11048]: Stopped
Apr 8 15:27:20 hosts Keepalived_vrrp[11049]: Stopped
Apr 8 15:27:20 hosts Keepalived[11047]: Stopped Keepalived v1.3.5 (03/19,2017), git commit v1.3.5-6-g6fa32f2
Apr 8 15:27:27 hosts systemd: Stopped MySQL Server.
———————

恢复 master 服务器故障，看是否主动抢占资源，成为活动服务器。

master 上启动 mysql 服务和 keepalived 服务：

# systemctl start mysqld

# systemctl start keepalived

# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 scope host lo
valid_lft forever preferred_lft forever
inet6 ::1/128 scope host
valid_lft forever preferred_lft forever
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP group default qlen 1000
link/ether 00:0c:29:cf:ab:c4 brd ff:ff:ff:ff:ff:ff
inet 192.168.1.10/24 brd 192.168.1.255 scope global noprefixroute eth0
valid_lft forever preferred_lft forever
inet6 fe80::fe8e:3c2f:4d32:e9fd/64 scope link noprefixroute
valid_lft forever preferred_lft forever
———————

可以看到，即使 master 故障恢复，也没有抢占资源，VIP 仍然在 backup 上，这是因为之前已经配置了 master 为非抢占模式（nopreempt）。

不过需要注意的是：

nopreempt 这个参数只能用于 state 为 BACKUP 的情况，所以在配置的时候要把 master 和 backup 的 state 都设置成 BACKUP，这样才会实现 keepalived 的非抢占模式！

也就是说：

* 当 state 状态一个为 MASTER，一个为 BACKUP 的时候，加不加 nopreempt 这个参数都是一样的效果。即都是根据 priority 优先级来决定谁抢占 vip 资源的，是抢占模式！

* 当 state 状态都设置成 BACKUP，如果不配置 nopreempt 参数，那么也是看 priority 优先级决定谁抢占 vip 资源，即也是抢占模式。

* 当 state 状态都设置成 BACKUP，如果配置 nopreempt 参数，那么就不会去考虑 priority 优先级了，是非抢占模式！即只有 vip 当前所在机器发生故障，另一台机器才能接管 vip。
即使优先级高的那一台机器恢复正常后也不会主动抢回 vip，只能等到对方发生故障，才会将 vip 切回来。

：

正文完

星哥玩云-微信公众号