Docker swarm cluster builds highly available MySQL active and standby

docker swarm Cluster construction is highly available mysql The main equipment

Premise ： for fear of mysql A single point of failure , It is necessary to introduce mysql The main equipment . The main and standby are on different machines , And need to maintain data consistency

technology ： Used here nginx+mysql+mysql Be a highly available active standby , utilize nginx Acting as a reverse agent , When one of them mysql When the server goes down , Automatically switch to another , Note that load balancing cannot be done here , Because there are problems with high concurrency .

 

One , start-up docker swarm colony  

1,docker stack The configuration is as follows , If you can't build docker swarm Clustered , Reference article ： build docker swarm Highly available clusters and common commands

 

#mysql1 service 
  mysql-1:
    image: mysql:5.7
    ports:
      - 3310:3306
    networks:
      # Specify the network the container joins 
      - network
    environment:
      # It is best to use this to set the time zone , Other still images can also be used 
      - TZ=CST-8
      # Appoint mysql Access code for 
      - MYSQL_ROOT_PASSWORD=123456
    # You can add --default-time-zone='+8:00' Set time zone 
    volumes:
      # Local file directory , Persistence 
      - /home/AuthenticationCenter/new_mysql/mone/data:/var/lib/mysql
      # Specify on binlog Configuration file for 
      - /home/AuthenticationCenter/new_mysql/mone/conf/my.cnf:/etc/mysql/my.cnf
    command: --character-set-server=utf8 
             --collation-server=utf8_general_ci
             --sql-mode="STRICT_TRANS_TABLES,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION,NO_ZERO_DATE,NO_ZERO_IN_DATE,ERROR_FOR_DIVISION_BY_ZERO"
    deploy:
      placement:
        constraints:
          # The hostname Specify the host on which the container starts 
          - node.hostname == ecs2d8ed9c368b9

#mysql2 service 
  mysql-2:
    image: mysql:5.7
    ports:
      - 3311:3306
    networks:
      # Specify the network the container joins 
      - network
    environment:
      # It is best to use this to set the time zone , Other still images can also be used 
      - TZ=CST-8
      # Appoint mysql Access code for 
      - MYSQL_ROOT_PASSWORD=123456
    # You can add --default-time-zone='+8:00' Set time zone 
    volumes:
      # Local file directory mount , Persistence 
      - /home/AuthenticationCenter/new_mysql/mtwo/data:/var/lib/mysql
      # Specify on binlog Configuration file for 
      - /home/AuthenticationCenter/new_mysql/mtwo/conf/my.cnf:/etc/mysql/my.cnf
    command: --character-set-server=utf8 
             --collation-server=utf8_general_ci
             --sql-mode="STRICT_TRANS_TABLES,NO_AUTO_CREATE_USER,NO_ENGINE_SUBSTITUTION,NO_ZERO_DATE,NO_ZERO_IN_DATE,ERROR_FOR_DIVISION_BY_ZERO"
    deploy:
        placement:
          constraints:
            # The hostname Specify the host on which the container starts 
            - node.hostname == ecseafe0d11214a




#mysql-nginx service 
  mysql_nginx:
    image: nginx:1.19.2
    ports:
      - 33060:3306
    networks:
      - network
    volumes:
      # Local file directory 
      - /home/AuthenticationCenter/new_mysql/nginx/nginx.conf:/etc/nginx/nginx.conf
      - /home/AuthenticationCenter/new_mysql/nginx/data:/var/www/html/upload
    deploy:
      mode: replicated
      # Here open two copies , prevent nginx A single point of failure 
      replicas: 2
      placement:
        constraints: [node.role == manager]    
    depends_on:
     - mysql-1

lookup mysql-1 and mysql-2 Of hostname, Can pass manager The management node executes the following command to view ,  The results are as follows

command ：docker nodel ls 

mysql-1 Of my.cnf To configure ：

[mysqld]
server_id=1
log-bin= mysql-bin

replicate-ignore-db=mysql
replicate-ignore-db=sys
replicate-ignore-db=information_schema
replicate-ignore-db=performance_schema

read-only=0
relay_log=mysql-relay-bin
log-slave-updates=on
auto-increment-offset=1
auto-increment-increment=2

!includedir /etc/mysql/conf.d/
!includedir /etc/mysql/mysql.conf.d/

mysql-2 Of my.cnf To configure ：

[mysqld]
server_id=2
log-bin= mysql-bin

replicate-ignore-db=mysql
replicate-ignore-db=sys
replicate-ignore-db=information_schema
replicate-ignore-db=performance_schema

read-only=0
relay_log=mysql-relay-bin
log-slave-updates=on
auto-increment-offset=2
auto-increment-increment=2

!includedir /etc/mysql/conf.d/
!includedir /etc/mysql/mysql.conf.d/

 

utilize nginx Do the reverse proxy of the fourth layer , You can also use Haproxy

nginx The configuration is as follows  

user www-data;
worker_processes auto;
pid /run/nginx.pid;

events {
    use   epoll;
    worker_connections  1024;
}


stream {
    upstream mysql {
        #backup As standby mysql, When mysql-1 Automatic switching after failure mysql-2, Achieve the active and standby effect ,
        server mysql-1:3306 max_fails=3 fail_timeout=30s;
        server mysql-2:3306      backup;
    }

    server {
        listen    3306;
        proxy_connect_timeout 3000s;
        proxy_timeout 6000s;
        proxy_pass mysql;
    }
}

Start by command docker swarm colony ： docker stack deploy -c docker-stack-test.yml mysqltest

Two , Turn on mysql Master master synchronization ,

1,   Execute the following command on the corresponding host , Get into mysql-1 and mysql-2 Containers

see docker The container of id:  docker ps -f "name=mysql-1"     perhaps   docker ps -f "name=mysql-2"

Into the container ：docker exec -it ( Containers ID) /bin/bash            ( Containers ID Is the position of the red box in the figure below )

 

 

2,   Respectively into the mysql-1 and mysql-2 After the container, execute the following command to enter mysql service     ( password ：123456) 

Get into mysql service ：mysql -u root -p 

Input password ：123456

3,mysql-1 and mysql-2 Execute the following commands respectively , Allow others slave Of io thread Monitor your own binlog file , The principle is as follows

     GRANT REPLICATION SLAVE ON *.* to 'slave'@'%' identified by '123456';

4, Execute the following command to record mysql-1 synchronous binlog File status

   show master status;

Here's the picture ： Record mysql-bin.000003 and 3825

 

5,mysql-2 Execute the following command , Turn on IO Thread listening mysql-1 Of binlog file , Yellow message by mysql-1 Recorded information

explain ：master_host： Here you can specify the container name , Because in the same docker swarm colony ,  

           master_password: visit mysql-1 perhaps mysql-2 Password

command ：change master to master_host='mysql-1',master_user='slave',master_password='123456',master_log_file='mysql-bin.000003',master_log_pos=3825,master_port=3306;

6,mysql-2  Start synchronization in

start slave ;

 7, Record mysql-2 synchronous binlog File status

show master status;

Here's the picture ： Record mysql-bin.000004 and 3161 

8, stay mysql-1 Execute the following command in , Yellow message by mysql-2 Recorded information

change master to master_host='mysql-2',master_user='slave',master_password='123456',master_log_file='mysql-bin.000004',master_log_pos=3161,master_port=3306;

 9,mysql-1 Turn on synchronization

start slave ;

10, Respectively in mysql-1 and mysql-2 Execute the following command in to check whether it is successfully opened .

show slave status\G;

  Here's the picture ： If both are yes , It means success .

doubt ：

1, By entering the corresponding MySQL Container configuration MySQL Master slave configuration , Do you need to reconfigure after the next restart ,    answer ： No need to reconfigure , Because the configuration file has been persisted to the host , Restart does not affect .
2, Join one MySQL It's down. , When this one MySQL Will the previously unsynchronized data be automatically synchronized after restart     take ： It will be , Because the corresponding master Of binlog Synchronized location ,positon The location of .

 

 

3、 ... and ,mysql Data recovery

If someone MySQL The server data is lost , Data recovery needs to be done again , And it needs to be reconfigured MySQL Lord , Let's talk about this in the next issue .......