Mysql Log based master-slave replication

First installation docker Don't say the , Here we're going to 3 Servers

from docker Remote warehouse pull Mysql:5.7 Mirror source

Here we need to mount the volume and port mapping in the container , Then we can modify the corresponding configuration , Make master-slave replication
docker run --name master -p 3306:3306 -p 33060:33060 -v /my/custom/:/etc/mysql/ -e MYSQL_ROOT_PASSWORD=123456 -d mysql:5.7
We do not have a customized configuration file when mounting , Set the configuration file directory /my/custom/ It's empty , This will cause after mounting , The files and folders that should be in the container are emptied
terms of settlement : Is to open another one mysql Containers , adopt docker cp The order will /etc/mysql/ Copy the files and folders under the directory to the local

The soft connection here is in an invalid state , We don't need to care about it here , It's still normal in the container , The function of this file in the container is to 【mysqld】 Such configuration fields are split into multiple files , The path of the file is saved in mysqld in
The above steps should be carried out on all three hosts

Next, we will start to arrange master-slave replication

1. modify mysqld.cnf Configuration fields in
   
   stay 【mysqld】 Add the following fields below
   
   These two fields are open binlog journal , Set up services id
   Be careful ： The slave can not start the log , The host must be turned on , The log of the slave is provided to , service id It cannot be repeated in a cluster , Nor can it be 0,0 Indicates that master-slave replication is not enabled
2. Create users to copy
   Each copy uses MySQL User name and password connect to source , Therefore, there must be a user account on the source where the replica can be used for connection .
   Create users first . Then grant this account the privileges required for replication , Use GRANT. If you create an account for replication purposes only , This account only needs REPLICATION SLAVE jurisdiction . for example , To set up a new user , The user can connect to any host in the domain for replication , Publish the following statement on the source ：

   mysql> CREATE USER 'repl'@'%.example.com' IDENTIFIED BY 'password';
   mysql> GRANT REPLICATION SLAVE ON *.* TO 'repl'@'%.example.com';
   

3. Get the binary log coordinates of the replication source
   There are two ways 1. Restart container 2.Flush Close the table to obtain a read lock
   Don't say the first , Look directly at the second type and enter the following command

   mysql> FLUSH TABLES WITH READ LOCK;
   

   Then enter the following command ：
   SHOW MASTER STATUS
   This step is to get the host , At present, we have reached the number of entries in that log , The slave can start synchronization at this position .
4. In fact, there should be another step here is the transfer of data , But we use new hosts and slaves , Don't do this step
5. The last step is to turn on the slave
   First release the read lock on the host UNLOCK TABLES;
   Then modify it on the slave 【mysqld】 Set up server-id
   Then configure the necessary information for connecting to the host

   mysql> CHANGE MASTER TO
       ->     MASTER_HOST='source_host_name',
       ->     MASTER_USER='replication_user_name',
       ->     MASTER_PASSWORD='replication_password',
       ->     MASTER_LOG_FILE='recorded_log_file_name',
       ->     MASTER_LOG_POS=recorded_log_position;
   

   Modify according to the field , The first is the main sentence ip, The second is the user name created on the host , The third is the user's password , The fourth is the serial number of the current log SHOW MASTER STATUS Found in , the last one recorded_log_position Is the number of lines in the log .
   Finally, start the slave thread on the two slave nodes START SLAVE;
   Finally, the master-slave establishment is completed