手机版
你好,游客 登录 注册 搜索
背景:
阅读新闻

MySQL双主互备+Keepalived高可用架构介绍

[日期:2017-02-28] 来源:Linux社区  作者:cubix [字体: ]

一、MySQL双主互备+keepalived高可用架构介绍

MySQL主从复制架构可以在很大程度保证MySQL的高可用,在一主多从的架构中还可以利用读写分离将读操作分配到从库中,减轻主库压力。但是在这种架构中,主库出现故障时需要手动将一台从库提升为主库。在对写操作要求较高的环境中,主库故障在主从架构中会成为单点故障。因此需要主主互备架构,避免主节点故障造成写操作失效。

在双主互备的架构中,每台MySQL都充当主服务器,同时充当对方的从服务器。在任意一台服务器上的写操作都会被复制到另一台服务器上,从而保证了数据的可靠性。

在双主互备的基础上加上keepalived,在其中一台机器上绑定虚拟ip(VIP)。利用vip统一对外服务,可以避免在两个节点同时写数据造成冲突。同时当keepalived主节点发生故障时,keeplived会自动将VIP切换到备节点上,从而实现主服务器的高可用。

MySQL双主互备+keepalived高可用架构图

二、MySQL双主互备架构部署

在上一篇MySQL主从复制操作中已经完成了MySQL主从架构的部署。在此基础上,需要在原Master上开启relay-log,在原Slave服务器上开启log-bin,同时在MySQL248上指定MySQL249为自己的主服务器并开启slave即可。

1.开启原Master上的relay-log,指定不复制的库

[root@Mysql-248 mysql-5.6.30]# grep -A8 'mysqld' my.cnf
[mysqld]
port = 3306
server_id = 1
log-bin=mysql-bin
#以下部分为在原基础上新增的内容
relay-log = mysql-relay-bin
replicate-wild-ignore-table=mysql.%
replicate-wild-ignore-table=test.%
replicate-wild-ignore-table=information_schema.%

2.开启原Slave上的log-bin,指定不复制的库

[root@Mysql-249 mysql-5.6.30]# grep -A8 'mysqld' my.cnf

[mysqld]
port = 3306
server_id = 2
relay-log = mysql-relay-bin
#以下部分为在原基础上新增的内容
log-bin=mysql-bin
replicate-wild-ignore-table=mysql.%
replicate-wild-ignore-table=test.%
replicate-wild-ignore-table=information_schema.%

3.在Mysql248服务器上指定Mysql249为自己的主服务器并开启slave

####  ① 在Mysql249上查看当前master信息,并建立复制用户
mysql> show master status;
+------------------+----------+--------------+------------------+-------------------+
| File            | Position | Binlog_Do_DB | Binlog_Ignore_DB | Executed_Gtid_Set |
+------------------+----------+--------------+------------------+-------------------+
| mysql-bin.000001 |      493 |              |                  |                  |
+------------------+----------+--------------+------------------+-------------------+
1 row in set (0.00 sec)
mysql> grant replication slave on *.* to 'repl_user'@'192.168.175.%' identified by '123456';
Query OK, 0 rows affected (0.00 sec)
mysql> flush privileges;
Query OK, 0 rows affected (0.00 sec)
 
#### ② 在Mysql248上指定Mysql249为自己的主服务器,开启slave
mysql> change master to \
    -> master_host='192.168.175.249',
    -> master_user='repl_user',
    -> master_password='123456',
    -> master_log_file='mysql-bin.000001',
    -> master_log_pos=493;
Query OK, 0 rows affected, 2 warnings (0.02 sec)
mysql> start slave;
Query OK, 0 rows affected (0.03 sec)

4.检验双主互备

  ①通过分别在两台服务器上使用show slave status\G,查询主库信息以及IO进程、SQL进程工作状态。若两台服务器的查询结果都为Slave_IO_Running: Yes,Slave_SQL_Running: Yes;则表示当前双主互备状态正常。

  ②在Mysql248数据库上建库建表,检查Mysql249上是否同步正常;然后在Mysql249上建库建表,检查Mysql248上是否同步正常。 

三、keepalived高可用部署

以下操作在两台机器上基本相同,不同之处请留意备注标识

1.keepalived编译安装

#编译操作两台机器一样
 
[root@Mysql-248 src]# tar -zxf keepalived-1.2.12.tar.gz
[root@Mysql-248 src]# cd keepalived-1.2.12
[root@Mysql-248 keepalived-1.2.12]# ./configure  --sysconf=/etc --with-kernel-dir=/usr/src/kernels/2.6.32-573.el6.x86_64/
[root@Mysql-248 keepalived-1.2.12]# make
[root@Mysql-248 keepalived-1.2.12]# make install
[root@Mysql-248 keepalived-1.2.12]# ln -s /usr/local/sbin/keepalived  /sbin/
[root@Mysql-248 keepalived-1.2.12]# chkconfig  --add keepalived
[root@Mysql-248 keepalived-1.2.12]# chkconfig  --level 35 keepalived on

2.keepalived配置

#配置文件中需要注意主备节点的priority与nopreempt配置不同
 
[root@Mysql-248 ~]# cd /etc/keepalived/
[root@Mysql-248 keepalived]# cat keepalived.conf
! Configuration File for keepalived
global_defs {
  notification_email {
    acassen@firewall.loc
    failover@firewall.loc
    sysadmin@firewall.loc
  }
  notification_email_from Alexandre.Cassen@firewall.loc
  smtp_server 192.168.200.1
  smtp_connect_timeout 30
  router_id MySQLHA_DEVEL
}
vrrp_script check_mysqld {
    script "/etc/keepalived/mysqlcheck/check_slave.sh" #定义检测mysql复制状态的脚本
    interval 2
    }
vrrp_instance HA_1 {
    state BACKUP                #两台mysql服务器均配置为BACKUP
    interface eth0
    virtual_router_id 80
    priority 100  #备用节点设置为90
    advert_int 2
    nopreempt  #不抢占模式,只在优先级高的机器上设置即可,备用节点不加此参数
    authentication {
        auth_type PASS
        auth_pass qweasdzxc
    }
   
    track_script {
        check_mysqld #调用mysql检测脚本
    }
    virtual_ipaddress {
        192.168.175.254/24 dev eth0    #mysql的对外服务IP,即VIP
    }
}
[root@Mysql-248 keepalived]#

3.配置Mysql检测脚本

#检测脚本主备节点都一样
 
[root@Mysql-248 keepalived]# cat /etc/keepalived/mysqlcheck/check_slave.sh
#!/bin/bash
#This scripts is check for Mysql Slave status
 
Mysqlbin=/usr/local/mysql-5.6.30/bin/mysql
user=root
pw='qwe123``'
port=3306
host=127.0.0.1
sbm=120
 
#Check for $Mysqlbin
if [ ! -f $Mysqlbin ];then
        echo 'Mysqlbin not found,check the variable Mysqlbin'
        exit 99
fi
 
#Get Mysql Slave Status
IOThread=`$Mysqlbin -h $host -P $port -u$user -p$pw -e 'show slave status\G'  2>/dev/null|grep 'Slave_IO_Running:'|awk '{print $NF}'`
SQLThread=`$Mysqlbin -h $host -P $port -u$user -p$pw -e 'show slave status\G' 2>/dev/null|grep 'Slave_SQL_Running:'|awk '{print $NF}'`
SBM=`$Mysqlbin -h $host -P $port -u$user -p$pw -e 'show slave status\G' 2>/dev/null|grep 'Seconds_Behind_Master:'|awk '{print $NF}'`
 
#Check if the mysql run
if [[ -z "$IOThread" ]];then
        exit 1
fi
 
#Check if the thread run
if [[ "$IOThread" = "No" || "$SQLThread" = "No" ]];then
        exit 1
        elif [[ $SBM -ge $sbm ]];then
                exit 1
        else
                exit 0
fi
[root@Mysql-248 keepalived]#

4.启动keepalived

#分别在Mysql248、Mysql249上启动keepalived
#需要注意由于设置了nopreempt非强占模式,所以先启动的节点会成为主节点,后启动的成为备节点。
 
[root@Mysql-248 keepalived]# /etc/init.d/keepalived  start
Starting keepalived: [  OK  ]
[root@Mysql-249 keepalived]# /etc/init.d/keepalived start
Starting keepalived: [  OK  ]

四、检验Mysql双主互备+keepalived高可用架构

1.启动后检查message日志信息

#在248上查看message日志,发现启动时节点进入MASTER状态,并且绑定了192.168.175.254的VIP
[root@Mysql-248 ~]# tail -f /var/log/messages
Sep  1 08:16:41 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Transition to MASTER STATE
Sep  1 08:16:43 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Entering MASTER STATE
Sep  1 08:16:43 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) setting protocol VIPs.
Sep  1 08:16:43 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Sending gratuitous ARPs on eth0 for 192.168.175.254
Sep  1 08:16:43 Mysql-248 Keepalived_healthcheckers[71142]: Netlink reflector reports IP 192.168.175.254 added
Sep  1 08:16:48 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Sending gratuitous ARPs on eth0 for 192.168.175.254
 
#在249上查看message日志,发现启动时节点进入BACKUP状态
[root@Mysql-249 ~]# tail -f /var/log/messages
Sep  1 08:19:15 Mysql-249 Keepalived_vrrp[54929]: Truncating auth_pass to 8 characters
Sep  1 08:19:15 Mysql-249 Keepalived_vrrp[54929]: Configuration is using : 65520 Bytes
Sep  1 08:19:15 Mysql-249 Keepalived_healthcheckers[54928]: Configuration is using : 7520 Bytes
Sep  1 08:19:15 Mysql-249 Keepalived_vrrp[54929]: Using LinkWatch kernel netlink reflector...
Sep  1 08:19:15 Mysql-249 Keepalived_vrrp[54929]: VRRP_Instance(HA_1) Entering BACKUP STATE
Sep  1 08:19:15 Mysql-249 Keepalived_healthcheckers[54928]: Using LinkWatch kernel netlink reflector...
Sep  1 08:19:15 Mysql-249 Keepalived_vrrp[54929]: VRRP sockpool: [ifindex(2), proto(112), unicast(0), fd(10,11)]

若两台服务器启动后都进入MASTER状态,都绑定了VIP,则需要检查防火墙是否放行了VRRP。

防火墙放行VRRP后,会有一台机器进入BACKUP状态。

#分别在两台服务器上放行vrrp协议(CentOS5/6配置方法):
[root@Mysql-248 keepalived]#iptables -I INPUT -p vrrp -d 224.0.0.18/32 -j ACCEPT
#保存放行规则。若是通过开机脚本控制防火墙规则,请注意修改对应脚本。
[root@Mysql-248 keepalived]#/etc/init.d/iptables

2.检查两台服务器的IP,并在客户机通过VIP连接数据库

#在248上查看,发现此时vip在此机器
[root@Mysql-248 keepalived]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    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 qlen 1000
    link/ether 00:0c:29:47:a4:ea brd ff:ff:ff:ff:ff:ff
    inet 192.168.175.248/24 brd 192.168.175.255 scope global eth0
    inet 192.168.175.254/24 scope global secondary eth0
    inet6 fe80::20c:29ff:fe47:a4ea/64 scope link
      valid_lft forever preferred_lft forever
       
       
#在249上查看,并无VIP
[root@Mysql-249 keepalived]# ip addr
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN
    link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
    inet 127.0.0.1/8 scope host lo
    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 qlen 1000
    link/ether 00:0c:29:e9:01:e0 brd ff:ff:ff:ff:ff:ff
    inet 192.168.175.249/24 brd 192.168.175.255 scope global eth0
    inet6 fe80::20c:29ff:fee9:1e0/64 scope link
      valid_lft forever preferred_lft forever

客户端通过VIP连接数据库

3.在客户机通过VIP连接Mysql,同时测试不间断ping vip254,并停止Mysql248上的Mysql服务器。停止后新建一个数据库测试主节点故障后的写操作是否正常。

[root@Mysql-248 keepalived]# /etc/init.d/mysqld stop
Shutting down MySQL............ SUCCESS!
[root@Mysql-248 keepalived]#

不间断ping测试故障切换时间

故障切换后测试写入数据

  在模拟故障的过程中,VIP切换时网络会出现一下中断,几次测试发现VIP切换的时间在几百毫秒延迟到一个丢包之间,切换时间非常短。但是测试写入数据提示连接中断,重试后发现可以正常写入数据。所以此架构需要程序端有数据库断开时重连的机制。在此情况下,主节点故障对业务的影响时间就降低到了秒级,无需人为干预才能恢复业务。

    一台节点故障后,仅剩下一台节点能够正常工作,只需要配合监控发现故障后,及时修复故障节点,就可以将恢复后的节点重新成为keepalived的备用节点,以保证架构的持续高可用。

4.故障切换过程

#在故障节点248上查看日志,发现VRRP_Script检测到Mysql状态faild,节点进入FAULT状态并移除vip
[root@Mysql-248 ~]# tail -f /var/log/messages
Sep  1 08:32:51 Mysql-248 Keepalived_vrrp[71143]: VRRP_Script(check_mysqld) failed
Sep  1 08:32:52 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Entering FAULT STATE
Sep  1 08:32:52 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) removing protocol VIPs.
Sep  1 08:32:52 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Now in FAULT state
Sep  1 08:32:52 Mysql-248 Keepalived_healthcheckers[71142]: Netlink reflector reports IP 192.168.175.254 removed
 
#249服务器进入MASTER状态,并绑定VIP
[root@Mysql-249 ~]# tail -f /var/log/messages
Sep  1 08:32:54 Mysql-249 Keepalived_vrrp[54929]: VRRP_Instance(HA_1) Transition to MASTER STATE
Sep  1 08:32:56 Mysql-249 Keepalived_vrrp[54929]: VRRP_Instance(HA_1) Entering MASTER STATE
Sep  1 08:32:56 Mysql-249 Keepalived_vrrp[54929]: VRRP_Instance(HA_1) setting protocol VIPs.
Sep  1 08:32:56 Mysql-249 Keepalived_vrrp[54929]: VRRP_Instance(HA_1) Sending gratuitous ARPs on eth0 for 192.168.175.254
Sep  1 08:32:56 Mysql-249 Keepalived_healthcheckers[54928]: Netlink reflector reports IP 192.168.175.254 added
Sep  1 08:33:01 Mysql-249 Keepalived_vrrp[54929]: VRRP_Instance(HA_1) Sending gratuitous ARPs on eth0 for 192.168.175.254

5.故障节点恢复

#将测试时关闭的数据库打开,查看日志发现节点进入BACKUP状态
#查看数据库发现故障时候创建的数据库已经同步过来了
#此时又恢复Mysql双主复制的高可用状态了
[root@Mysql-248 keepalived]# /etc/init.d/mysqld start
Starting MySQL.. SUCCESS!
[root@Mysql-248 keepalived]# tail -5 /var/log/messages
Sep  1 08:32:52 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) removing protocol VIPs.
Sep  1 08:32:52 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Now in FAULT state
Sep  1 08:32:52 Mysql-248 Keepalived_healthcheckers[71142]: Netlink reflector reports IP 192.168.175.254 removed
Sep  1 08:51:42 Mysql-248 Keepalived_vrrp[71143]: VRRP_Script(check_mysqld) succeeded
Sep  1 08:51:44 Mysql-248 Keepalived_vrrp[71143]: VRRP_Instance(HA_1) Entering BACKUP STATE
[root@Mysql-248 keepalived]#
[root@Mysql-248 mysql-5.6.30]# mysql -uroot -p'qwe123``' -e 'show databases;'
Warning: Using a password on the command line interface can be insecure.
+--------------------+
| Database          |
+--------------------+
| information_schema |
| cubix              |
| db1                |
| mysql              |
| performance_schema |
+--------------------+

构建高可用集群Keepalived+Haproxy负载均衡  http://www.linuxidc.com/Linux/2016-12/138917.htm

Keepalived安装与配置 http://www.linuxidc.com/Linux/2017-02/140421.htm

《Keepalived权威指南》下载见 http://www.linuxidc.com/Linux/2012-05/60951.htm

Nginx+Keepalived实现站点高可用  http://www.linuxidc.com/Linux/2016-12/137883.htm

Nginx+Keepalived实现站点高可用(负载均衡)  http://www.linuxidc.com/Linux/2016-12/138221.htm

构建高可用集群Keepalived+Haproxy负载均衡 http://www.linuxidc.com/Linux/2016-12/138917.htm

CentOS6.5下 Keepalived高可用服务单实例配置 http://www.linuxidc.com/Linux/2016-12/138110.htm

Keepalived+Nginx实现高可用(HA)  http://www.linuxidc.com/Linux/2017-02/140422.htm

本文永久更新链接地址http://www.linuxidc.com/Linux/2017-02/141173.htm

linux
本文评论   查看全部评论 (0)
表情: 表情 姓名: 字数

       

评论声明
  • 尊重网上道德,遵守中华人民共和国的各项有关法律法规
  • 承担一切因您的行为而直接或间接导致的民事或刑事法律责任
  • 本站管理人员有权保留或删除其管辖留言中的任意内容
  • 本站有权在网站内转载或引用您的评论
  • 参与本评论即表明您已经阅读并接受上述条款