利用Keepalived+mysql构建高可用MySQL双主自动切转

转载:http://www.it300.com/index.php/article-15266.html

关于MySQL-HA,目前有多种解决方案,比如heartbeat、drbd、mmm、共享存储,但是它们各有优缺点。heartbeat、drbd配置较为复杂,需要自己写脚本才能实现MySQL自动切换,对于不会脚本语言的人来说,这无疑是一种脑裂问题;对于mmm,生产环境中很少有人用,且mmm管理端需要单独运行一台服务器上,要是想实现高可用,就得对mmm管理端做HA,这样无疑又增加了硬件开支;对于共享存储,个人觉得MySQL数据还是放在本地较为安全,存储设备毕竟存在单点隐患。

    使用MySQL双master+keepalived是一种非常好的解决方案,在MySQL-HA环境 中,MySQL互为主从关系,这样就保证了两台MySQL数据的一致性,然后用keepalived实现虚拟IP,通过keepalived自带的服务监 控功能来实现MySQL故障时自动切换。

    下面,我把即将上线的一个生产环境中的架构与大家分享一下,看一下这个架构中,MySQL-HA是如何实现的,环境拓扑如下:

1 MySQL-VIP:192.168.1.200 
2 MySQL-master1:192.168.1.201 
3 MySQL-master2:192.168.1.202 
4   
5 OS版本:CentOS 5.4 
6 MySQL版本:5.0.89 
7 Keepalived版本:1.1.20

    一、MySQL master-master配置

1、修改MySQL配置文件

两台MySQL均如要开启binlog日志功能,开启方法:在MySQL配置文件[MySQLd]段中加上log-bin=MySQL-bin选项

两台MySQL的server-ID不能一样,默认情况下两台MySQL的serverID都是1,需将其中一台修改为2即可

Master1配置:

01 #vim /etc/my.cnf
02  
03 log-bin=mysql-bin   //开启binlog日志功能
04 log =/usr/local/mysql/var/mysql.log    //会打印mysql的所以sql语句
05  
06 server-id= 1  //
07  
08 binlog-do-db = test    //需要同步的库名称
09 auto-increment-increment= 2
10 auto-increment-offset= 2<span style="font-family:Simsun;"><span style="white-space:normal;"> </span></span>

Master2配置:

01 #vim /etc/my.cnf
02  
03 log-bin=mysql-bin    //开启binlog日志功能
04 log =/usr/local/mysql/var/mysql.log  //会打印mysql的所以sql语句
05  
06 server-id= 2
07  
08 binlog-do-db = test   //需要同步的库名称
09 auto-increment-increment= 2
10 auto-increment-offset= 2

2、将192.168.1.201设为192.168.1.202的主服务器

在192.168.1.201上新建授权用户

01 MySQL> grant replication slave on *.* to 'replication'@'%' identified by'replication'
02 Query OK, 0 rows affected (0.00 sec) 
03   
04 MySQL> show master status; 
05 +------------------+----------+--------------+------------------+ 
06 | File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | 
07 +------------------+----------+--------------+------------------+ 
08 | MySQL-bin.000003 |      374 |              |                  |  
09 +------------------+----------+--------------+------------------+ 
10 1 row in set (0.00 sec)

在192.168.1.202上将192.168.1.201设为自己的主服务器

01 MySQL> change master to master_host='192.168.1.201',master_user='replication',master_password='replication',master_log_file='MySQL-bin.000003',master_log_pos=374; 
02 Query OK, 0 rows affected (0.05 sec) 
03   
04 MySQL> start slave; 
05 Query OK, 0 rows affected (0.00 sec) 
06   
07 MySQL> show slave statusG 
08 *************************** 1. row *************************** 
09              Slave_IO_State: Waiting for master to send event 
10                 Master_Host: 192.168.1.201 
11                 Master_User: replication 
12                 Master_Port: 3306 
13               Connect_Retry: 60 
14             Master_Log_File: MySQL-bin.000003 
15         Read_Master_Log_Pos: 374 
16              Relay_Log_File: MySQL-master2-relay-bin.000002 
17               Relay_Log_Pos: 235 
18       Relay_Master_Log_File: MySQL-bin.000003 
19            Slave_IO_Running: Yes 
20           Slave_SQL_Running: Yes 
21             Replicate_Do_DB:  
22         Replicate_Ignore_DB:  
23          Replicate_Do_Table:  
24      Replicate_Ignore_Table:  
25     Replicate_Wild_Do_Table:  
26 Replicate_Wild_Ignore_Table:  
27                  Last_Errno: 0 
28                  Last_Error:  
29                Skip_Counter: 0 
30         Exec_Master_Log_Pos: 374 
31             Relay_Log_Space: 235 
32             Until_Condition: None 
33              Until_Log_File:  
34               Until_Log_Pos: 0 
35          Master_SSL_Allowed: No 
36          Master_SSL_CA_File:  
37          Master_SSL_CA_Path:  
38             Master_SSL_Cert:  
39           Master_SSL_Cipher:  
40              Master_SSL_Key:  
41       Seconds_Behind_Master: 0 
42 1 row in set (0.00 sec)

3、将192.168.1.202设为192.168.1.201的主服务器

在192.168.1.202上新建授权用户

01 MySQL> grant replication slave on *.* to 'replication'@'%' identified by'replication'
02 Query OK, 0 rows affected (0.00 sec) 
03   
04 MySQL> show master status; 
05 +------------------+----------+--------------+------------------+ 
06 | File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | 
07 +------------------+----------+--------------+------------------+ 
08 | MySQL-bin.000003 |      374 |              |                  |  
09 +------------------+----------+--------------+------------------+ 
10 1 row in set (0.00 sec)

在192.168.1.201上,将192.168.1.202设为自己的主服务器

01 MySQL> change master to master_host='192.168.1.202',master_user='replication',master_password='replication',master_log_file='MySQL-bin.000003',master_log_pos=374; 
02 Query OK, 0 rows affected (0.05 sec) 
03   
04 MySQL> start slave; 
05 Query OK, 0 rows affected (0.00 sec) 
06   
07 MySQL> show slave statusG 
08 *************************** 1. row *************************** 
09              Slave_IO_State: Waiting for master to send event 
10                 Master_Host: 192.168.1.202 
11                 Master_User: replication 
12                 Master_Port: 3306 
13               Connect_Retry: 60 
14             Master_Log_File: MySQL-bin.000003 
15         Read_Master_Log_Pos: 374 
16              Relay_Log_File: MySQL-master1-relay-bin.000002 
17               Relay_Log_Pos: 235 
18       Relay_Master_Log_File: MySQL-bin.000003 
19            Slave_IO_Running: Yes 
20           Slave_SQL_Running: Yes 
21             Replicate_Do_DB:  
22         Replicate_Ignore_DB:  
23          Replicate_Do_Table:  
24      Replicate_Ignore_Table:  
25     Replicate_Wild_Do_Table:  
26 Replicate_Wild_Ignore_Table:  
27                  Last_Errno: 0 
28                  Last_Error:  
29                Skip_Counter: 0 
30         Exec_Master_Log_Pos: 374 
31             Relay_Log_Space: 235 
32             Until_Condition: None 
33              Until_Log_File:  
34               Until_Log_Pos: 0 
35          Master_SSL_Allowed: No
36          Master_SSL_CA_File:  
37          Master_SSL_CA_Path:  
38             Master_SSL_Cert:  
39           Master_SSL_Cipher:  
40              Master_SSL_Key:  
41       Seconds_Behind_Master: 0 
42 1 row in set (0.00 sec)

4、MySQL同步测试

如上述均正确配置,现在任何一台MySQL上更新数据都会同步到另一台MySQL,MySQL同步在此不再演示

二、keepalived安装及配置

1、192.168.1.201服务器上keepalived安装及配置

安装keepalived

02 #tar xvzf keepalived-1.2.2.tar.gz
03 #cd keepalived-1.2.2
04 #./configure 
05 #make && make install
06 #cp /usr/local/etc/rc.d/init.d/keepalived /etc/init.d/
07 cp /usr/local/etc/sysconfig/keepalived /etc/sysconfig/
08 chmod +x /etc/init.d/keepalived
09 chkconfig --add keepalived
10 chkconfig keepalived on
11 mkdir /etc/keepalived
12 ln -s /usr/local/sbin/keepalived /usr/sbin/

配置keepalived

我们自己在新建一个配置文件,默认情况下keepalived启动时会去/etc/keepalived目录下找配置文件

01 #vi /etc/keepalived/keepalived.conf 
02 global_defs { 
03      notification_email { 
04          haohailuo@163.com
05      
06      notification_email_from haohailuo@163.com 
07      smtp_server 127.0.0.1 
08      smtp_connect_timeout 30 
09      router_id MySQL-ha 
10
11   
12 vrrp_instance VI_1 { 
13      state BACKUP   #两台配置此处均是BACKUP 
14      interface eth0 
15      virtual_router_id 51 
16      priority 100   #优先级,另一台改为90 
17      advert_int 1 
18      nopreempt  #不抢占,只在优先级高的机器上设置即可,优先级低的机器不设置 
19      authentication { 
20          auth_type PASS 
21          auth_pass 1111 
22      
23      virtual_ipaddress { 
24          192.168.1.200 
25      
26
27   
28 virtual_server 192.168.1.200 3306 { 
29      delay_loop 2   #每个2秒检查一次real_server状态 
30      lb_algo wrr   #LVS算法 
31      lb_kind DR    #LVS模式 
32      persistence_timeout 60   #会话保持时间 
33      protocol TCP 
34      real_server 192.168.1.201 3306 { 
35          weight 3 
36          notify_down /usr/local/MySQL/bin/MySQL.sh  #检测到服务down后执行的脚本 
37          TCP_CHECK { 
38              connect_timeout 10    #连接超时时间 
39              nb_get_retry 3       #重连次数 
40              delay_before_retry 3   #重连间隔时间 
41              connect_port 3306   #健康检查端口
42          }
43      
44 }
编写检测服务down后所要执行的脚本
1 #vi /usr/local/MySQL/bin/MySQL.sh 
2 #!/bin/sh 
3 pkill keepalived 
4 #chmod +x /usr/local/MySQL/bin/MySQL.sh

注:此脚本是上面配置文件notify_down选项所用到的,keepalived使用notify_down选项来检查real_server 的服务状态,当发现real_server服务故障时,便触发此脚本;我们可以看到,脚本就一个命令,通过pkill keepalived强制杀死keepalived进程,从而实现了MySQL故障自动转移。另外,我们不用担心两个MySQL会同时提供数据更新操作, 因为每台MySQL上的keepalived的配置里面只有本机MySQL的IP+VIP,而不是两台MySQL的IP+VIP

启动keepalived

1 #/usr/local/keepalived/sbin/keepalived –D 
2 #ps -aux | grep keepalived

测试

找一台局域网PC,然后去ping  MySQL的VIP,这时候MySQL的VIP是可以ping的通的

停止MySQL服务,看keepalived健康检查程序是否会触发我们编写的脚本

2、192.168.1.202上keepalived安装及配置

安装keepalived,安装方法参照192.168.1.201的安装方法

配置keepalived

这台配置和上面基本一样,但有三个地方不同:优先级为90、无抢占设置、real_server为本机IP

01 #vi /etc/keepalived/keepalived.conf
02 global_defs { 
03      notification_email { 
04          haohailuo@163.com 
05      
06      notification_email_from haohailuo@163.com 
07      smtp_server 127.0.0.1 
08      smtp_connect_timeout 30 
09      router_id MySQL-ha 
10
11   
12 vrrp_instance VI_1 { 
13      state BACKUP 
14      interface eth0 
15      virtual_router_id 51 
16      priority 90 
17      advert_int 1 
18      authentication { 
19          auth_type PASS 
20          auth_pass 1111 
21      
22      virtual_ipaddress { 
23          192.168.1.200 
24      
25
26   
27 virtual_server 192.168.1.200 3306 { 
28      delay_loop 2 
29      lb_algo wrr 
30      lb_kind DR 
31      persistence_timeout 60 
32      protocol TCP 
33      real_server 192.168.1.202 3306 { 
34          weight 3 
35          notify_down /usr/local/MySQL/bin/MySQL.sh 
36          TCP_CHECK { 
37              connect_timeout 10 
38              nb_get_retry 3 
39              delay_before_retry 3 
40              connect_port 3306 
41          }
42      }
43 }

编写检测服务down后所要执行的脚本

1 #vi /usr/local/MySQL/bin/MySQL.sh 
2 #!/bin/sh 
3 pkill keepalived 
4 #chmod +x /usr/local/MySQL/bin/MySQL.sh

启动keepalived

1 #/usr/local/keepalived/sbin/keepalived –D 
2 #ps -aux | grep keepalived

测试

停止MySQL服务,看keepalived健康检查程序是否会触发我们编写的脚本

三、测试

MySQL远程登录测试

我们找一台安装有MySQL客户端的windows,然后登录VIP,看是否能登录,在登录之两台MySQL服务器都要授权允许从远程登录

1 MySQL> grant all privileges on *.* to 'root'@'%' identified by '123456'
2 Query OK, 0 rows affected (0.00 sec) 
3   
4 MySQL> flush privileges; 
5 Query OK, 0 rows affected (0.00 sec)
使用客户端登录VIP测试
1 C:MySQLin>MySQL.exe -uroot -p123456 -h192.168.1.200 -P3306 
2 Welcome to the MySQL monitor.  Commands end with ; or g. 
3 Your MySQL connection id is 224 
4 Server version: 5.0.89-log Source distribution 
5   
6 Type 'help;' or 'h' for help. Type 'c' to clear the current input statement. 
7   
8 MySQL>

keepalived故障转移测试

※在windows客户端一直去ping  VIP,然后关闭192.168.1.201上的keepalived,正常情况下VIP就会切换到192.168.1.202上面去

※开启192.168.1.201上的keepalived,关闭192.168.1.202上的keepalived,看是否能自动切换,正常情况下VIP又会属于192.168.1.201

注:keepalived切换速度还是非常块的,整个切换过程只需1-3秒

● MySQL故障转移测试

※在192.168.1.201上关闭MySQL服务,看VIP是否会切换到192.168.1.202上

※开启192.168.1.201上的MySQL和keepalived,然后关闭192.168.1.202上的MySQL,看VIP是否会切换到192.168.1.201上

下面是用windows客户端连接的MySQL的VIP,在切换时我执行了一个MySQL查询命令,从执行show databases到显示出结果时间为3-5秒(大家可以看到上面有个错误提示,不过不用担心,因为我们的keepalived切换大概为3秒左右,这3 秒左右VIP是谁都不属于的)

01 MySQL> show databases; 
02 ERROR 2006 (HY000): MySQL server has gone away 
03 No connection. Trying to reconnect... 
04 Connection id:    592 
05 Current database: *** NONE *** 
06   
07 +--------------------+ 
08 | Database           | 
09 +--------------------+ 
10 | information_schema | 
11 | MySQL              | 
12 | test               | 
13 +--------------------+ 
14 3 rows in set (9.01 sec)
后话:世间万事万物,都不具备绝对的完美,就像上面的MySQL-HA一样,keepalived只能做到对3306的健康检查,但是做不到比如像 MySQL复制中的slave-SQL、slave-IO进程的检查。所以要想做到一些细致的健康检查,还得需要借助额外的监控工具,比如nagios, 然后用nagios实现短信、邮件报警,从而能够有效地解决问题。
原文地址:https://www.cnblogs.com/lijiasnong/p/5362600.html