Principle and implementation of MySQL master-slave replication

What is master-slave replication ？

Master slave replication refers to the transfer of the master database DDL and DML Operations are passed to the slave database through binary logs , These logs are then re executed on the slave database , So that the data of the slave database and the master database are consistent .

The principle of master-slave replication

• Mysql When the transaction is committed, the master database records the data changes as events in the binary log Binlog in ;
• The main library pushes binary log files Binlog Relay logs from events in to slave libraries Relay log in , After that, the slave database redoes the data change operation according to the relay log , The consistency of the master and slave databases is achieved by replication ;
• Mysql Through three threads to complete the master-slave database data replication , among Binlog Dump Threads run on the main library ,I/O Threads and SQL The thread runs on the slave library ;
• When starting replication from the slave Library , First create I/O Threads connect to the main library , The master library is then created Binlog Dump Thread reads database events and sends them to I/O Threads ,I/O The thread gets the event and updates it to the relay log of the slave Library Relay log In the middle , And then from the library sql Thread read relay log Relay log Update database events and apply , As shown in the figure below ：
  

Build the master instance

• function mysql Main instance

docker run -p 3307:3306 --name mysql-master \
-v /mydata/mysql-master/log:/var/log/mysql \
-v /mydata/mysql-master/data:/var/lib/mysql \
-v /mydata/mysql-master/conf:/etc/mysql \
-e MYSQL_ROOT_PASSWORD=root  \
-d mysql:5.7

• stay mysql Configuration folder /mydata/mysql-master/conf Create a configuration file in my.cnf：

touch my.cnf

• Modify the configuration file my.cnf, The configuration information is as follows

[mysqld]
##  Set up server_id, You need to be unique in the same LAN 
server_id=101 
##  Specify the name of the database that does not need to be synchronized 
binlog-ignore-db=mysql  
##  Turn on binary log function 
log-bin=wdj-mysql-bin  
##  Set the memory size of binary log （ Business ）
binlog_cache_size=1M  
##  Set the binary log format to use （mixed,statement,row）
binlog_format=mixed  
##  Binary log expiration cleanup time . The default value is 0, Does not automatically clean up .
expire_logs_days=7  
##  Skip all errors encountered in master-slave replication or errors of the specified type , avoid slave End copy interrupt .
##  Such as ：1062 An error is a duplicate of some primary keys ,1032 The error is because the master-slave database data is inconsistent 
slave_skip_errors=1062

• Restart the instance after modifying the configuration

docker restart mysql-master

• Get into mysql-master In the container

docker exec -it mysql-master /bin/bash

• Use... In containers mysql To connect to the client

mysql -uroot -proot

• Create data synchronization users

CREATE USER 'slave'@'%' IDENTIFIED BY '123456';
GRANT REPLICATION SLAVE, REPLICATION CLIENT ON *.* TO 'slave'@'%';

Build from instance

• function mysql From the example

docker run -p 3308:3306 --name mysql-slave \
-v /mydata/mysql-slave/log:/var/log/mysql \
-v /mydata/mysql-slave/data:/var/lib/mysql \
-v /mydata/mysql-slave/conf:/etc/mysql \
-e MYSQL_ROOT_PASSWORD=root  \
-d mysql:5.7

• stay mysql Configuration folder /mydata/mysql-slave/conf Create a configuration file in my.cnf

touch my.cnf

• Modify the configuration file my.cnf

[mysqld]
#server_id, You need to be unique in the same LAN 
server_id=102
##  Specify the name of the database that does not need to be synchronized 
binlog-ignore-db=mysql  
##  Turn on binary log function , in preparation for Slave As an instance of another database Master When using 
log-bin=wdj-mysql-slave1-bin  
##  Set the memory size of binary log （ Business ）
binlog_cache_size=1M  
##  Set the binary log format to use （mixed,statement,row）
binlog_format=mixed  
##  Binary log expiration cleanup time . The default value is 0, Does not automatically clean up .
expire_logs_days=7  
##  Skip all errors encountered in master-slave replication or errors of the specified type , avoid slave End copy interrupt .
##  Such as ：1062 An error is a duplicate of some primary keys ,1032 The error is because the master-slave database data is inconsistent 
slave_skip_errors=1062  
## relay_log Configure relay logs 
relay_log=wdj-mysql-relay-bin  
## log_slave_updates Express slave Write the copy event to your own binary log 
log_slave_updates=1  
## slave Set to read only （ have super Except for users with permission ）
read_only=1

• Restart the instance after modifying the configuration

docker restart mysql-slave

Connect the master and slave databases

• Connect to the master database mysql client , View the status of the primary database

show master status;

The master database status is shown below ：

• Get into mysql-slave In the container

docker exec -it mysql-slave /bin/bash

• Use... In containers mysql To connect to the client

mysql -uroot -proot

• Configure master slave replication in the slave database

change master to master_host='192.168.6.132', master_user='slave', master_password='123456', master_port=3307, master_log_file='wdj-mysql-bin.000001', master_log_pos=785, master_connect_retry=30;  

• Master slave copy command parameter description

1. master_host： The main database IP Address ;
2. master_port： The running port of the primary database ;
3. master_user： A user account created in the master database to synchronize data ;
4. master_password： The user password created in the master database for synchronizing data ;
5. master_log_file： Specify the log file to copy data from the database , By looking at the status of the master data , obtain File Parameters ;
6. master_log_pos： Specify where to start copying data from the database , By looking at the status of the master data , obtain Position Parameters ;
7. master_connect_retry： The connection failed and the time interval between retries , The unit is in seconds .

• Check the master-slave synchronization status
  
  The above figure shows no Indicates that the connection has not been made
• Turn on master-slave synchronization

start slave;

• Check the database status and find that it has been synchronized
  
  Status has changed from no Turn into yes 了

Master slave replication test

There are many ways to test master-slave replication , You can create a database in the main instance , See if there is the database in the instance , If there is , Indicates that the master-slave replication has been successfully established .

• Create a database in the main instance testzhucong
  
  After creating the database in the main database , Now check the database from the Library
  
  The newly created database of the master database has also been created synchronously from the database , Since then, a simple Mysql The practical operation of master-slave replication .