【Keepalived+MySQL】MySQL双主互备+高可用

一、基本信息说明

【DB1】

IP: 192.168.102.144

hostname: LVS-Real1

【DB2】

IP: 192.168.102.145

hostname: LVS-Real2

【VIP】

IP:  192.168.102.146

二、MySQL配置主主互备

 1.配置DB1和DB2的/etc/my.cnf

【DB1】

[root@LVS-Real1 ~]# more /etc/my.cnf

[client]
port =
socket = /tmp/mysql.sock  

[mysqld]
user=mysql
port =
server_id = 1                        #需保证唯一性
socket=/tmp/mysql.sock
basedir =/usr/local/mysql
datadir =/usr/local/mysql/data
pid-file=/usr/local/mysql/data/mysqld.pid
log-error=/usr/local/mysql/log/mysql-error.log

log-bin=mysql-bin                     #开启二进制日志
relay-log=mysql-relay-bin

replicate-wild-ignore-table=mysql.%   #忽略复制mysql数据库下的所有对象，以下依次类推
replicate-wild-ignore-table=test.%
replicate-wild-ignore-table=information_schema.%

【DB2】

[root@LVS-Real2 ~]# more /etc/my.cnf
[client]
port =
socket = /tmp/mysql.sock  

[mysqld]
user=mysql
port =
server_id = 2                       #需保证唯一性
socket=/tmp/mysql.sock
basedir =/usr/local/mysql
datadir =/usr/local/mysql/data
pid-file=/usr/local/mysql/data/mysqld.pid
log-error=/usr/local/mysql/log/mysql-error.log

log-bin=mysql-bin                  #开启二进制日志
relay-log=mysql-relay-bin
replicate-wild-ignore-table=mysql.%
replicate-wild-ignore-table=test.%
replicate-wild-ignore-table=information_schema.%

 2.手动同步数据库

如果DB1上有数据，在执行主主互备之前，需要将DB1和DB2上两个数据库保持同步，首先在DB1上执行备份，执行如下语句：

mysql>flush tables with read lock;

在关闭上述终端的情况下，新开启一个终端打包数据库。

3.创建复制用户并授权

• 首先在【DB1】上的MySQL库中创建复制用户

mysql>grant replication slave on *.* to 'repl_user'@'192.168.102.145' identified by 'repl_passwd';

• 在【DB1】上执行如下语句，并记下File和Position的值

mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000002 |     1004 |              |                  |                   |
+------------------+----------+--------------+------------------+-------------------+
 row in set (0.02 sec)

• 然后在【DB2】上将DB1设为自己的主服务器，如下：

change master to \
master_host='192.168.102.144',           #DB1的IP地址
master_user='repl_user',
master_password='repl_passwd',
master_log_file='mysql-bin.000002',      #DB1上查询出的File值
master_log_pos=;                     #DB1上查询出的Position值

• 在【DB2】上启动slave服务，并查询slave的运行状态

mysql>start slave;

查看Slave的运行状态，这里需要关注Slave_IO_Running和Slave_SQL_Running.这两个就是在Slave节点上运行的主从复制线程，正常情况下两个值都应该为Yes.

mysql> show slave status\G;
*************************** . row ***************************
               Slave_IO_State: Waiting for master to send event
                  Master_Host: 192.168.102.144
                  Master_User: repl_user
                  Master_Port:
                Connect_Retry:
              Master_Log_File: mysql-bin.
          Read_Master_Log_Pos:
               Relay_Log_File: mysql-relay-bin.
                Relay_Log_Pos:
        Relay_Master_Log_File: mysql-bin.
             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes
              Replicate_Do_DB:
          Replicate_Ignore_DB:
           Replicate_Do_Table:
       Replicate_Ignore_Table:
      Replicate_Wild_Do_Table:
  Replicate_Wild_Ignore_Table: mysql.%,test.%,information_schema.%
                   Last_Errno:
                   Last_Error:
                 Skip_Counter:
          Exec_Master_Log_Pos:
              Relay_Log_Space:
              Until_Condition: None
               Until_Log_File:
                Until_Log_Pos:
           Master_SSL_Allowed: No
           Master_SSL_CA_File:
           Master_SSL_CA_Path:
              Master_SSL_Cert:
            Master_SSL_Cipher:
               Master_SSL_Key:
        Seconds_Behind_Master:
Master_SSL_Verify_Server_Cert: No
                Last_IO_Errno:
                Last_IO_Error:
               Last_SQL_Errno:
               Last_SQL_Error:
  Replicate_Ignore_Server_Ids:
             Master_Server_Id:
                  Master_UUID: 64e9b20f-2eee-11e8-ab62-000c29889112
             Master_Info_File: /usr/local/mysql/data/master.info
                    SQL_Delay:
          SQL_Remaining_Delay: NULL
      Slave_SQL_Running_State: Slave has read all relay log; waiting for more updates
           Master_Retry_Count:
                  Master_Bind:
      Last_IO_Error_Timestamp:
     Last_SQL_Error_Timestamp:
               Master_SSL_Crl:
           Master_SSL_Crlpath:
           Retrieved_Gtid_Set:
            Executed_Gtid_Set:
                Auto_Position:
         Replicate_Rewrite_DB:
                 Channel_Name:
           Master_TLS_Version:
 row in set (0.00 sec)

• 接下来开始配置从DB2到DB1的MySQL主从复制，这个配置过程和上面一样。
• 首先在【DB2】上的MySQL库中创建复制用户，并查看数据库状态，记下File和Position值。

grant replication slave on *.* to 'repl_user'@'192.168.102.144' identified by 'repl_passwd';

mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File             | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000002 |     1004 |              |                  |                   |
+------------------+----------+--------------+------------------+-------------------+
 row in set (0.02 sec)

• 然后在【DB1】上将DB2设为自己的主服务器

change master to \
master_host='192.168.102.145',
master_user='repl_user',
master_password='repl_passwd',
master_log_file='mysql-bin.000002',
master_log_pos=;

• 在【DB1】上启动slave服务

mysql>start slave;

在【DB1】查看Slave的运行状态，这里需要关注Slave_IO_Running和Slave_SQL_Running.这两个就是在Slave节点上运行的主从复制线程，正常情况下两个值都应该为Yes.

mysql> show slave status\G;
*************************** . row ***************************
               Slave_IO_State: Waiting for master to send event
                  Master_Host: 192.168.102.145
                  Master_User: repl_user
                  Master_Port:
                Connect_Retry:
              Master_Log_File: mysql-bin.
          Read_Master_Log_Pos:
               Relay_Log_File: mysql-relay-bin.
                Relay_Log_Pos:
        Relay_Master_Log_File: mysql-bin.
             Slave_IO_Running: Yes
            Slave_SQL_Running: Yes
              Replicate_Do_DB:
          Replicate_Ignore_DB:
           Replicate_Do_Table:
       Replicate_Ignore_Table:
      Replicate_Wild_Do_Table:
  Replicate_Wild_Ignore_Table: mysql.%,test.%,information_schema.%
                   Last_Errno:
                   Last_Error:
                 Skip_Counter:
          Exec_Master_Log_Pos:
              Relay_Log_Space:
              Until_Condition: None
               Until_Log_File:
                Until_Log_Pos:
           Master_SSL_Allowed: No
           Master_SSL_CA_File:
           Master_SSL_CA_Path:
              Master_SSL_Cert:
            Master_SSL_Cipher:
               Master_SSL_Key:
        Seconds_Behind_Master:
Master_SSL_Verify_Server_Cert: No
                Last_IO_Errno:
                Last_IO_Error:
               Last_SQL_Errno:
               Last_SQL_Error:
  Replicate_Ignore_Server_Ids:
             Master_Server_Id:
                  Master_UUID: a35a032d-2ef8-11e8-bd3c-000c2910f959
             Master_Info_File: /usr/local/mysql/data/master.info
                    SQL_Delay:
          SQL_Remaining_Delay: NULL
      Slave_SQL_Running_State: Slave has read all relay log; waiting for more updates
           Master_Retry_Count:
                  Master_Bind:
      Last_IO_Error_Timestamp:
     Last_SQL_Error_Timestamp:
               Master_SSL_Crl:
           Master_SSL_Crlpath:
           Retrieved_Gtid_Set:
            Executed_Gtid_Set:
                Auto_Position:
         Replicate_Rewrite_DB:
                 Channel_Name:
           Master_TLS_Version:
 row in set (0.00 sec)

• 至此，主主复制配置完毕；

三、Keepalived的安装与配置

1.下载Keepalived

http://www.keepalived.org/download.html

2.安装Keepalived

#.安装依赖包
yum -y install gcc openssl-devel  libnfnetlink libnfnetlink-devel

#.开始安装
tar -xvf keepalived-1.2..tar.gz
cd keepalived-1.2.
./configure --prefix=/usr/local/keepalived
make
make install

3.复制文件到相应目录

cp /usr/local/keepalived/sbin/keepalived               /usr/sbin/ 

cp /usr/local/keepalived/etc/rc.d/init.d/keepalived   /etc/init.d/  

cp /usr/local/keepalived/etc/sysconfig/keepalived     /etc/sysconfig/

mkdir -p /etc/keepalived
cp -r /usr/local/keepalived/etc/keepalived/keepalived.conf  /etc/keepalived

4.配置keepalived.conf

 【DB1】

[root@LVS-Real1 keepalived]# more /etc/keepalived/keepalived.conf
global_defs {
     notification_email {
         guanyy0911@.com
     }
     notification_email_from guanyy0911@.com
     smtp_server 127.0.0.1
     smtp_connect_timeout
     router_id MySQL-ha
}    

vrrp_instance VI_1 {
     state BACKUP
     interface eth0
     virtual_router_id
     priority 100       #设置优先级
     advert_int
     nopreempt          #设置不抢占，当因为故障切换到DB2后，如果DB1恢复，则不再切回DB1，直到DB2出现故障才切换回DB1
     authentication {
         auth_type PASS
         auth_pass
     }
     virtual_ipaddress {
         192.168.102.146    #设置虚拟IP，即VIP
     }
}    

virtual_server 192.168.102.146  {
     delay_loop
     lb_algo wrr
     lb_kind DR
     persistence_timeout
     protocol TCP
     real_server 192.168.102.144  {
         weight
         notify_down /etc/keepalived/mysql.sh
         TCP_CHECK {
             connect_timeout
             nb_get_retry
             delay_before_retry
             connect_port
         }
     }
}  

mysql.sh脚本的内容如下：

[root@LVS-Real1 keepalived]# more /etc/keepalived/mysql.sh
#!/bin/bash
pkill keepalived

说明：该脚本主要用来当MySQL服务关闭时杀掉keepalived进程，进而达到切换的目的。

【DB2】

[root@LVS-Real2 keepalived]# more /etc/keepalived/keepalived.conf
global_defs {
     notification_email {
         guanyy0911@.com
     }
     notification_email_from guanyy0911@.com
     smtp_server 127.0.0.1
     smtp_connect_timeout
     router_id MySQL-ha
}    

vrrp_instance VI_1 {
     state BACKUP
     interface eth0
     virtual_router_id
     priority 90        #设置优先级，要比DB1低
     advert_int
     authentication {
         auth_type PASS
         auth_pass
     }
     virtual_ipaddress {
         192.168.102.146    #设置虚拟IP，即VIP
     }
}    

virtual_server 192.168.102.146  {
     delay_loop
     lb_algo wrr
     lb_kind DR
     persistence_timeout
     protocol TCP
     real_server 192.168.102.145  {
         weight
         notify_down /etc/keepalived/mysql.sh
         TCP_CHECK {
             connect_timeout
             nb_get_retry
             delay_before_retry
             connect_port
         }
     }
}   

mysql.sh脚本的内容如下：

[root@LVS-Real1 keepalived]# more /etc/keepalived/mysql.sh
#!/bin/bash
pkill keepalived

说明：该脚本主要用来当MySQL服务关闭时杀掉keepalived进程，进而达到切换的目的。

5.启动keepalived

service keepalived start

5.检查并测试VIP是否可用

如果在/var/log/messages文件中有如下信息，说明VIP已经可用。

Mar  :: LVS-Real1 Keepalived[]: Stopping Keepalived v1.2.18 (/,)
Mar  :: LVS-Real1 Keepalived[]: Starting Keepalived v1.2.18 (/,)
Mar  :: LVS-Real1 Keepalived[]: Starting Healthcheck child process, pid=
Mar  :: LVS-Real1 Keepalived[]: Starting VRRP child process, pid=
Mar  :: LVS-Real1 Keepalived_vrrp[]: Netlink reflector reports IP 192.168.102.144 added
Mar  :: LVS-Real1 Keepalived_vrrp[]: Netlink reflector reports IP fe80::20c:29ff:fe88: added
Mar  :: LVS-Real1 Keepalived_vrrp[]: Registering Kernel netlink reflector
Mar  :: LVS-Real1 Keepalived_vrrp[]: Registering Kernel netlink command channel
Mar  :: LVS-Real1 Keepalived_healthcheckers[]: Netlink reflector reports IP 192.168.102.144 added
Mar  :: LVS-Real1 Keepalived_vrrp[]: Registering gratuitous ARP shared channel
Mar  :: LVS-Real1 Keepalived_vrrp[]: Opening file '/etc/keepalived/keepalived.conf'.
Mar  :: LVS-Real1 Keepalived_healthcheckers[]: Netlink reflector reports IP fe80::20c:29ff:fe88: added
Mar  :: LVS-Real1 Keepalived_vrrp[]: Configuration is using :  Bytes
Mar  :: LVS-Real1 Keepalived_healthcheckers[]: Registering Kernel netlink reflector
Mar  :: LVS-Real1 Keepalived_healthcheckers[]: Registering Kernel netlink command channel
Mar  :: LVS-Real1 Keepalived_healthcheckers[]: Opening file '/etc/keepalived/keepalived.conf'.
Mar  :: LVS-Real1 Keepalived_vrrp[]: Using LinkWatch kernel netlink reflector...
Mar  :: LVS-Real1 Keepalived_healthcheckers[]: Configuration is using :  Bytes
Mar  :: LVS-Real1 Keepalived_vrrp[]: VRRP_Instance(VI_1) Entering BACKUP STATE
Mar  :: LVS-Real1 Keepalived_vrrp[]: VRRP sockpool: [ifindex(), proto(), unicast(), fd(,)]
Mar  :: LVS-Real1 Keepalived_healthcheckers[]: Using LinkWatch kernel netlink reflector...

此时从外部的客户端可以ping通该VIP.

C:\Users\Sakura>ping 192.168.102.146

正在 Ping 192.168.102.146 具有  字节的数据:
来自 192.168.102.146 的回复: 字节= 时间<1ms TTL=
来自 192.168.102.146 的回复: 字节= 时间<1ms TTL=
来自 192.168.102.146 的回复: 字节= 时间<1ms TTL=
来自 192.168.102.146 的回复: 字节= 时间<1ms TTL=

192.168.102.146 的 Ping 统计信息:
    数据包: 已发送 = ，已接收 = ，丢失 =  (% 丢失)，
往返行程的估计时间(以毫秒为单位):
    最短 = 0ms，最长 = 0ms，平均 = 0ms

6. 同时启动DB1和DB2上的mysql和keepalived服务

service mysql.server start
service keepalived start

7.从第三方客户端通过VIP来登录数据库。看是否可以登录。

我们通过navicat进行登录，发现是可以通过VIP登录。之后查询当前使用是哪个数据库。如下图查询到，使用的是DB1.

  7.我们停掉DB1数据库，看是否会切换到DB2上。

     通过实验发现，发现已经切换到DB2上了。

8. 我们再次启动DB1上的Mysql服务和keepalived服务（已经通过脚本实现MySQL服务关闭的同时，脚本会杀掉keepalived进程）

通过实验发现，由于我们设置的是不抢占，在DB1启动后，并没有切换回DB2. 达到预期的目的。

9.这次我们停掉DB2上的MySQL服务，看是否会切换回DB1.

通过实验发现，在停掉DB2上的MySQL服务后，已经自动切换回DB1上。达到预期目的。

特别提示：当MySQL被关闭时，其所在的主机的keepalived也同时被关闭。但在重新启动MySQL服务时，keepalived不会自动启动，需要手动启动。

10.至此，整个配置过程完毕！