MySQL high availability architecture MMM architecture

MMM (Multi-Master Replication Manager, multi-master replication management architecture)

Main role

Monitor and manage MySQL The master-master replication topology is used, and when the current master server fails, master-slave switching and failover between the master and master-slave servers are performed.

Related recommendations: [mysql tutorial]

Main functions:

● Monitor the health status of MySQL master-slave replication (active Master-master replication in active mode (master-master), master-master replication in active-passive mode (master-master-standby))

● Perform failover when the master database goes down and automatically configure other slave DBs to new Replication of the master DB

● Provides a virtual IP that can be mastered, written (read), and can automatically migrate the virtual IP when there is a problem with the master and slave servers

Architecture:

Resources:

## Deployment:

配置主主复制及主从同步集群
安装主从节点所需要的支持包（perl）
安装及配置 MMM 工具集
允许 MMM 监控服务
测试配置

Demonstration:

Steps:

1、配置主(100)<->主（101）[参考: [Tony 老师搭建 MySQL 主从服务器](https://learnku.com/articles/31832)]
2、配置主(100)<->从(102)   [参考: [Tony 老师搭建 MySQL 主从服务器](https://learnku.com/articles/31832)]
3、 安装
     rpm -vih <http://dl.fedoraproject.org/pub/epel/7/x86_64/Packages/e/epel-release-7-11.noarch.rpm>
     rpm -Uvh <http://rpms.remirepo.net/enterprise/remi-release-7.rpm>
4、 yum clean all && yum makecache
5、更改yum源配置
    yum search mmm  # 查询mmm支持包
    yum -y install mysql-mmm-agent.noarch #每个服务器都需要安装mmm代理
    yum -y install mysql-mmm*  #监控服务器安装监控服务
    systemctl status mysql-mmm-agent
    systemctl start mysql-mmm-agent
    systemctl stop mysql-mmm-agent
    systemctl restart mysql-mmm-agent
6、建立账号
    CREATE USER repl@&#39;192.168.71.%&#39; identified by &#39;123456&#39;； //创建，建议从服务器IP段 ①
    CREATE USER mmm_monitor@&#39;192.168.71.%&#39; identified by &#39;123456&#39;; //创建，建议从服务器IP段 ①
    CREATE USER mmm_agent@&#39;192.168.71.%&#39; identified by &#39;123456&#39;; //创建，建议从服务器IP段 ①
7、账号授权
    GRANT replication client on *.* to &#39;mmm_monitor&#39;@&#39;192.168.71.%&#39; identified by &#39;123456&#39;;   #用于MMM监控服务器使用，在master上建立
    GRANT super,replication client ,process on *.* to &#39;mmm_agent&#39;@&#39;192.168.71.%&#39; identified by &#39;123456&#39;;    #MMM代理服务，改变故障转移和主从切换
    GRANT REPLICATION SLAVE ON *.* TO repl@&#39;192.168.71.%&#39;; //授权 复制账号
8、配置MMM（数据库节点配置,三台均配置）
    cd /etc/mysq-mmm/
    vim mmm_common.conf
    配置< host default >：
        网口：cluster_interface 
        pid和bin路径：默认
        replication_user：repl
        replication_password：123456
        agent_user：mmm_agent
        agent_password：123456
    配置< host db1 >:
        ip ：192.168.71.244
        mode ：master
        peer：db2
    配置 < host db2 > :
        ip：192.168.71.223
        mode：master
        peer：db1
    配置< host db3 >:
        ip：192.168.71.220
        mode：slave
    配置< role writer >
        hosts : db1,db2
        ips：192.168.71.90
        mode：exclusive( 唯一的)
    配置 < role reader >
        hosts：db1,db2,db3
        ips：192.168.71.91,192.168.71.92,192.168.71.93
        mode：balanced（平衡的）
    复制配置文件到其他服务器
        scp mmm_common.conf root@192.168.71.223:/etc/mysql-mmm/
        scp mmm_common.conf root@192.168.71.220:/etc/mysql-mmm/
    vim mysql-agent.conf (三台服务器对应设置为：db1,db2,db3)
9、监控节点配置（192.168.71.220，只配置监控节点）
    vim mmm_mon.conf  #监控数据文件
    配置< host default >
        monitor_user ：mmm_monitor
        monitor_password：123456
    配置monitor
        ping_ips：192.168.71.244，192.168.71.223，192.168.71.220
10、启动MMM服务
    /etc/init.d/mysql-mmm-agent start（三台服务器均启动代理服务）
    /etc/init.d/mysql-mmm-monitor start （监控服务器启动监控服务）
11、查看监控信息
    mmm_control  
    mmm_control show #查看集群状态
    mmm_control checks all #检查集群状态
    mmm_control ping #检查与监控服务器连接状态
    mmm_control set_online/offline #设置服务器上线/下线
    mmm_control help #查看帮助
12、查看服务器虚拟IP状态
13、测试
    关闭db1的数据库:/etc/init.d/mysqld stop
    查看监控服务器上集群状态:mmm_control show
    查看DB3的主从状态

**通过mmm_control checks all 检查服务器状态**
Q1：
    UNKNOWN: Error occurred: install_driver(mysql) failed: Can&#39;t load &#39;/usr/lib64/perl5/vendor_perl/auto/DBD/mysql/mysql.so&#39; for module DBD::mysql: libmysqlclient.so.18: cannot open shared object file: No such file or directory at /usr/lib64/perl5/DynaLoader.pm line 190, <STDIN> line 1.
A1：
    perl-DBD-MySQL安装出问题
        S1:rpm -qa |grep -i DBD  #查看DBD安装
        S2:rpm -e --nodeps perl-DBD-MySQL-4.023-6.el7.x86_64 #卸载DBD
        S3：rpm -qa |grep -i DBD #查看卸载状态
        S4:wget http://dev.mysql.com/get/mysql-community-release-el7-5.noarch.rpm #下载MySQLserver,可根据系统选择具体版本
        S5:rpm -ivh mysql-community-release-el7-5.noarch.rpm #rpm MySQLserver
        S6:yum install mysql-community-server #yum安装
        S7:yum install perl-DBD-MySQL    #重新安装
        S8:systemctl restart mysql-mmm-monitor #重启
        S9:mmm_control checks all  + mmm_control show ->正常
Q2：
    当M1宕机时，MMM架构自动切换至M2，插入数据，S未更新，M1重启时，S仍未更新
A2:
    插入S一条空事务，记录与M2时相同即可。
Q3：
    handler error HA_ERR_KEY_NOT_FOUND; the event&#39;s master log mysql-bin.000002, end_log_pos 673 #跳过插入 | 更新 | 删除
A3：
    stop slave;
    insert into user(`username`) values(&#39;1&#39;);
    set global sql_slave_skip_counter=1;
    start slave;
Q4：
    The slave I/O thread stops because master and slave have equal MySQL server ids #切换主从发现，server_id相同
A4：
    show variables like &#39;**server_id**&#39;;
    vim /etc/my.cnf    修改server_id 值
**配置多线程**
stop slave #在从上，停止链路复制
set global slave_parallel_type = &#39;logical_clock&#39;;  #设置逻辑时钟的方式
set global slave_parallel_workers = 4;      #设置并发线程数
start slave；

The above is the detailed content of MySQL high availability architecture MMM architecture. For more information, please follow other related articles on the PHP Chinese website!