mmm架构的来源
- 众所周知,MySQL自身提供了主从复制,然后可以很轻松实现master-master双向复制,同时再为其中一个Master节点搭建一个Slave库。这样就实现了MySQL-MMM架构的基础:master1和master2之间双向复制,同时Master1和Slave1之间是主从复制。
- 这样整个体系中存在两个Master,正常情况下只有一个master对外提供写服务。如果对外提供服务的master意外宕机了,这是MySQL本身并不具备failover切换的能力,尽管集群中仍然有一个正常的master节点,但应用仍不可用。mysql-mmm就是为了解决这个问题诞生的。
mmm架构的原理
- MySQL-MMM是Master-Master Replication Manager for MySQL(mysql主主复制管理器)的简称,是Google的开源项目(Perl脚本),主要用来监控mysql主主复制并做失败转移
- 其原理是将真实数据库节点的IP(RIP)映射为虚拟IP(VIP)集,在这个虚拟的IP集中,有一个专用于write的IP,多个用于read的IP,这个用于Write的VIP映射着数据库集群中的两台master的真实IP(RIP),以此来实现Failover的切换,其他read的VIP可以用来均衡读(balance)。
mmm机构优缺点
优点
- 使用Perl脚本语言开发及完全开源
- 提供了读写VIP(虚拟IP),使服务器角色的变更对前端应用透明
- MMM提供了从服务器的延迟监控
- MMM提供了主数据库故障转移后从服务器对新主的重新同步功能
- 很容易对发生故障的主数据库重新上线
缺点
- 发布时间比较早不支持MySQL新的复制功能(基于GTID的复制)
- 没有读负载均衡的功能
- 在进行主从切换时,容易造成数据丢失
- MMM监控服务存在单点故障
mmm架构原理图
实验部署
环境部署
master1IP地址:192.168.144.167
master2IP地址:192.168.144.151
slave1IP地址:192.168.144.168
slave2IP地址:192.168.144.145
monitorIP地址:192.168.144.164
在master1\master2\slave1\slave2服务器中安装MYSQL数据库
[root@master1 ~]# wget -O /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repo //获取源地址
[root@master1 ~]# yum -y install epel-release //安装epel源
[root@master1 ~]# yum clean all && yum makecache //yum缓存清空
[root@master1 ~]# yum -y install mariadb-server mariadb //安装mariadb数据库
[root@master1 ~]# systemctl stop firewalld.service //关闭防火墙
[root@master1 ~]# setenforce 0
[root@master1 ~]# systemctl start mariadb.service //开启数据库
修改master1数据库配置文件
[root@master1 ~]# vim /etc/my.cnf //进入编辑配置文件
[mysqld]
log_error=/var/lib/mysql/mysql.err //错误日志文件
log=/var/lib/mysql/mysql_log.log //主从日志存放位置
log_slow_queries=/var/lib/mysql_slow_queris.log //man日志
binlog-ignore-db=mysql,information_schema //二进制文件
character_set_server=utf8 //字符集
log_bin=mysql_bin //二进制日志文件
server_id=11 //服务id(不能相同)
log_slave_updates=true //允许从服务器更新
sync_binlog=1 //同步日志
auto_increment_increment=2 //自增列
auto_increment_offset=1 //起始点
[root@master1 ~]# systemctl restart mariadb.service //重启数据库
使用scp复制数据库配置文件到其他MySQL服务,并在其他MySQL服务器中修改server_id
[root@master1 ~]# scp /etc/my.cnf root@192.168.144.151:/etc/ (server-id=22)
[root@master1 ~]# scp /etc/my.cnf root@192.168.144.168:/etc/ (server-id=33)
[root@master1 ~]# scp /etc/my.cnf root@192.168.144.145:/etc/ (server-id=44)
[root@master1 ~]# netstat -anpt | grep 3306 //查看端口3306
tcp 0 0 0.0.0.0:3306 0.0.0.0:* LISTEN 4235/mysqld
配置主主复制
master1
[root@master1 ~]# mysql ##进入数据库
MariaDB [(none)]> show master status; ##查看主服务器的状态信息
+------------------+----------+--------------+--------------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB |
+------------------+----------+--------------+--------------------------+
| mysql_bin.000001 | 245 | | mysql,information_schema |
+------------------+----------+--------------+--------------------------+
1 row in set (0.00 sec)
MariaDB [(none)]> grant replication slave on *.* to 'replication'@'192.168.144.%' identified by '123456'; //授权给144段的网段复制的权限用户名replication密码123456
Query OK, 0 rows affected (0.00 sec)
master2
MariaDB [(none)]> change master to master_host='192.168.144.167',master_user='replication',master_password='123456',master_log_file='mysql_bin.000001',master_log_pos=245; //在master2上同步master1服务器
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> grant replication slave on *.* to 'replication'@'192.168.144.%' identified by '123456'; //在master2上授权复制权限
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> show master status; //查看master2的服务器的状态信息
+------------------+----------+--------------+--------------------------+
| File | Position | Binlog_Do_DB | Binlog_Ignore_DB |
+------------------+----------+--------------+--------------------------+
| mysql_bin.000001 | 410 | | mysql,information_schema |
+------------------+----------+--------------+--------------------------+
1 row in set (0.00 sec)
MariaDB [(none)]> flush privileges; //刷新权限
Query OK, 0 rows affected (0.00 sec)
master1
MariaDB [(none)]> change master to master_host='192.168.144.151',master_user='replication',master_password='123456',master_log_file='mysql_bin.000001',master_log_pos=410; //master1同步master2服务器数据库
Query OK, 0 rows affected (0.03 sec)
MariaDB [(none)]> flush privileges;
Query OK, 0 rows affected (0.00 sec)
master1,master2上开启同步
MariaDB [(none)]> start slave; //开启同步
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> show slave status\G; //查看同步状态信息
...
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
...
slave1,slave2做主从同步
MariaDB [(none)]> change master to master_host='192.168.144.167',master_user='replication',master_password='123456',master_log_file='mysql_bin.000001',master_log_pos=245; //在从服务器上同步master1主服务器
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> flush privileges; //刷新权限
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> start slave; //开启同步
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> show slave status\G; //查看同步的状态信息
...
Slave_IO_Running: Yes
Slave_SQL_Running: Yes
...
测试主主,主从的同步状态
主服务器master1
MariaDB [(none)]> create database school; //创建数据库
Query OK, 1 row affected (0.00 sec)
MariaDB [(none)]> show databases; //查看数据库
+--------------------+
| Database |
+--------------------+
| information_schema |
| mysql |
| performance_schema |
| school |
| test |
+--------------------+
5 rows in set (0.00 sec)
slave服务器上查看数据库
MariaDB [(none)]> show databases; //实现主主,主从同步
+--------------------+
| Database |
+--------------------+
| information_schema |
| mysql |
| performance_schema |
| school |
| test |
+--------------------+
5 rows in set (0.00 sec)
monitor服务器上配置epel-release源清空缓存,然后安装MMM
wget -O /etc/yum.repos.d/CentOS-Base.repo http://mirrors.aliyun.com/repo/Centos-7.repo
yum -y install epel-release
yum clean all && yum makecache
yum -y install mysql-mmm*
所有MySQL服务器上安装mmm
yum -y install mysql-mmm*
所有服务器上对mmm进行配置
[root@master1 ~]# vim /etc/mysql-mmm/mmm_common.conf //所有主机上都要配置,直接复制多份
……
<host default>
cluster_interface ens33 //修改网卡
…
replication_user replication //修改用户名
replication_password 123456 //密码
agent_user mmm_agent
agent_password 123456 //密码
<host db1>
ip 192.168.144.167 //master1地址
mode master
peer db2
</host>
<host db2>
ip 192.168.144.151 //master2地址
mode master
peer db1
</host>
<host db3>
ip 192.168.144.168 //slave1地址
mode slave
</host>
<host db4>
ip 192.168.144.145 //slave2地址
mode slave
</host>
<role writer>
hosts db1, db2 ##写服务器虚拟ip
ips 192.168.144.250
mode exclusive
</role>
<role reader>
hosts db3, db4 ##读服务器虚拟ip
ips 192.168.144.251, 192.168.144.252
mode balanced
</role>
##复制到其他的服务器中
[root@master1 ~]# scp /etc/mysql-mmm/mmm_common.conf root@192.168.144.151:/etc/mysql-mmm/
root@192.168.144.151's password:
mmm_common.conf 100% 836 267.1KB/s 00:00
[root@master1 ~]# scp /etc/mysql-mmm/mmm_common.conf root@192.168.144.168:/etc/mysql-mmm/
root@192.168.144.168's password:
mmm_common.conf 100% 836 863.2KB/s 00:00
[root@master1 ~]# scp /etc/mysql-mmm/mmm_common.conf root@192.168.144.145:/etc/mysql-mmm/
root@192.168.144.145's password:
mmm_common.conf 100% 836 904.7KB/s 00:00
[root@master1 ~]# scp /etc/mysql-mmm/mmm_common.conf root@192.168.144.164:/etc/mysql-mmm/
在monitor服务器上配置
[root@monitor ~]# vim /etc/mysql-mmm/mmm_mon.conf
<host default>
monitor_user mmm_monitor
monitor_password 123456 ##修改monitor的密码
</host>
在所有数据库上为mmm_agent代理授权,为mmm_monitor授权监控
MariaDB [(none)]> grant super, replication client, process on *.* to 'mmm_agent'@'192.168.144.%' identified by '123456'; //授权代理
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> grant replication client on *.* to 'mmm_monitor'@'192.168.144.%' identified by '123456'; //授权监控
Query OK, 0 rows affected (0.00 sec)
MariaDB [(none)]> flush privileges; //刷新权限
Query OK, 0 rows affected (0.00 sec)
修改所有数据库的mmm_agent.conf
[root@master1 ~]# vim /etc/mysql-mmm/mmm_agent.conf
this db1 //根据规划进行逐一调整
[root@master2 ~]# vim /etc/mysql-mmm/mmm_agent.conf
this db2 //根据规划进行逐一调整
[root@slave1 ~]# vim /etc/mysql-mmm/mmm_agent.conf
this db3 //根据规划进行逐一调整
[root@slave2 ~]# vim /etc/mysql-mmm/mmm_agent.conf
this db4 //根据规划进行逐一调整
##所有数据库开启
[root@master1 ~]systemctl start mysql-mmm-agent.service //开启代理服务
[root@master1 ~]systemctl enable mysql-mmm-agent.service //加入开机自启动
在monitor上配置
[root@monitor ~]# vim /etc/mysql-mmm/mmm_mon.conf
<monitor>
ip 127.0.0.1
pid_path /run/mysql-mmm-monitor.pid
bin_path /usr/libexec/mysql-mmm
status_path /var/lib/mysql-mmm/mmm_mond.status
ping_ips 192.168.144.167,192.168.144.151,192.168.144.168,192.168.144.145 //所有数据库服务器地址
auto_set_online 10 //自动在线时间
[root@monitor ~]# systemctl stop firewalld.service //关闭防火墙
[root@monitor ~]# setenforce 0
[root@monitor ~]# systemctl start mysql-mmm-monitor.service /开启监控服务
[root@monitor ~]# mmm_control show //查看主从的飘逸地址
db1(192.168.144.167) master/ONLINE. Roles: writer(192.168.144.250)
db2(192.168.144.151) master/ONLINE. Roles:
db3(192.168.144.168) slave/ONLINE. Roles: reader(192.168.144.252)
db4(192.168.144.145) slave/ONLINE. Roles: reader(192.168.144.251)
测试漂移地址
[root@master1 ~]# systemctl stop mariadb.service //模拟停止master1服务器
monitor服务器上查看
[root@monitor ~]# mmm_control show
db1(192.168.144.167) master/HARD_OFFLINE. Roles:
db2(192.168.144.151) master/ONLINE. Roles: writer(192.168.144.250)
db3(192.168.144.168) slave/ONLINE. Roles: reader(192.168.144.252)
db4(192.168.144.145) slave/ONLINE. Roles: reader(192.168.144.251)
(重启master1数据库服务,虚拟地址不会被抢占到master1)
mmm_control checks all //需要各种OK
mmm_control move_role writer db1 //可以切换虚拟地址
在monitor上安装MySQL作为测试机用虚拟ip登录数据库
[root@monitor ~]# yum install mysql -y
//master1服务器上授权monitor地址访问//
MariaDB [(none)]> grant all on *.* to 'testdba'@'192.168.144.164' identified by '123456';
##授权monitor地址访问
Query OK, 0 rows affected (0.01 sec)
MariaDB [(none)]> flush privileges; ##刷新权限
Query OK, 0 rows affected (0.00 sec)
[root@monitor ~]# mysql -utestdba -p -h 192.168.144.250 ##使用虚拟地址即可登录数据库
Enter password:
MariaDB [(none)]>
来源:51CTO
作者:SiceLc
链接:https://blog.51cto.com/14473285/2458541