Three implementation schemes of MySQL read-write separation

Hello everyone , I meet you again , I'm your friend, Quan Jun .

MySQL Three implementation schemes of read-write separation

List of articles

• MySQL Three implementation schemes of read-write separation
  • One 、 Building a “ One master and two slaves ” Of MySQL colony
  • Two 、 Read write separation implementation ： Scheme 1
    • 2.1 Configure multiple data sources
    • 2.2 Use AbstractRoutingDataSource
    • 2.3 The disadvantages of this version ：
  • 3、 ... and 、 Read write separation implementation ： Option two
    • 3.1 adopt ShardingSphere-jdbc Read and write separation
    • 3.2 The disadvantages of this version ：
  • Four 、 Read write separation implementation ： Option three
    • 4.1 adopt ShardingSphere-Proxy Read and write separation

One 、 Building a “ One master and two slaves ” Of MySQL colony

First, build one mysql colony ( One master and two slaves ), Semi-synchronous replication ：mysql Semi-synchronous replication , Three nodes .

Two 、 Read write separation implementation ： Scheme 1

2.1 Configure multiple data sources

（1）、 be based on Spring/Spring Boot, Configure multiple data sources ( for example 2 individual ,master and slave)

（2）、 According to the concrete Service Method will manipulate data , Inject different data sources ,1.0 edition

By configuring multiple data sources , stay service Layer realizes read-write separation .

2.2 Use AbstractRoutingDataSource

（3）、 To improve the 1.1： Based on operation AbstractRoutingDataSource And custom annotations readOnly And Class , Simplify automatic data source switching

（4）、 Improve the second 1.2： Support configuring multiple slave libraries ;

Use AbstractRoutingDataSource And custom annotations .

（5）、 Improve three times 1.3： Support load balancing of multiple slave libraries

Ideas ： stay myreadwritesep-abstract-rounting-v1 On the basis of , More than a read Data source of type , Then judge the annotated name by read When , stay read01 and read02 Choose... At random .

--  Create user ,  Execute... On the primary node 
create schema performance;
CREATE USER 'performance'@'%' IDENTIFIED BY 'performance^pw';
GRANT ALL ON performance.* TO 'performance'@'%';
flush privileges;

--  Execute... On the primary node 
create table performance.t_model_info(
  m_id    int(11)  not null primary key auto_increment comment ' Primary key , Self increasing ',
  model_type varchar(128) not null unique comment ' Module type ',
  model_name varchar(128) not null unique comment ' Module name ',
  model_status tinyint(1)  not null default 0 comment '0  Enable , 1  Is not enabled '
)Engine=InnoDB AUTO_INCREMENT=1 default CHARSET=utf8mb4;

Copy

2.3 The disadvantages of this version ：

• Invasive strong ;
• “ After writing and reading ” Inconsistency ,server Method , Go to the main library first insert sentence , And then immediately select, But I can't find the data ;

3、 ... and 、 Read write separation implementation ： Option two

3.1 adopt ShardingSphere-jdbc Read and write separation

improvement v1.0,ShardingSphere-jdbc Of Master-Slave function

1）SQL Parsing and transaction management , Automatic read-write separation

• First of all , Manage everything ;
• second , do SQL analysis , Automatic read-write separation ;

2） solve ” After writing and reading ” The problem of inconsistency

• If in a transaction , Write before they are read , The framework has an optimization ; In a business , The first few are poison , Just read it normally , As long as you meet someone who writes , No matter how many reads follow , Go to the main warehouse . This is the perfect solution “ After writing and reading ” The problem of inconsistency .

adopt Shardingsphere-jdbc The framework realizes the separation of reading and writing .

3.2 The disadvantages of this version ：

• It's still invasive to the business （ This set of things , It needs to be configured into the business code ）;
• It is not friendly to the transformation of existing old systems ;

Four 、 Read write separation implementation ： Option three

4.1 adopt ShardingSphere-Proxy Read and write separation

improvement v2.0, MyCat/ShardingSphere-Proxy Of Master-Slave function

1） You need to deploy a middleware , Rules are configured in middleware

2） Simulate one MySQL The server , No intrusion into the business system

shardingsphere-proxy Read and write separation .

Publisher ： Full stack programmer stack length , Reprint please indicate the source ：https://javaforall.cn/126809.html Link to the original text ：https://javaforall.cn