Database schema evolution
For an early mall system, because the traffic is not too large, the data-bearing pressure is not high, we can have all the data on a MySQL server, this case is similar to the database schema shown below.
With the continuous increase in business volume, increasing pressure on the database, then we can make MySQL master-slave replication cluster, using MySQL Mycat complete separation of reading and writing, which, as the main library to write MySQL database, MySQL from the library library as a read, then the database schema may be as shown below,
On how to implement MySQL Mycat separate read and write, we can refer to " Mycat of --Mycat read and write separation from the copy on the basis of the main MySQL ," a text.
After a period of time, with the rise in business volume, we find for the mall business, the various different business pressure on the frequency of visits is different. At this point, we would even think of the independent store different data services out, made a separate database, and for the upper layer of applications, to access these databases, still like the same access to the same database. At this time, we will think of vertical database segmentation.
We can store the database into a user database, product database, order and warehouse database with the database. Since the mall business, orders and warehouse operations with business needs time to interact, early, we can order business database and warehouse distribution business into one database. Finally, we can store business is divided into user traffic, commodity business and business orders.
In this case, we can store a database schema as shown in FIG.
Wherein, for each part of the image above, we can make high availability mode, related to high availability mode you can refer to the following article.
" MyCat --Mycat achieve the separate read and write from the master copy based on MySQL "
" MyCat of --Mycat cluster deployment (+ myCat based HAProxy) "
《Mycat之——高可用负载均衡集群的实现（HAProxy + Keepalived + Mycat）》
本文，我们就基于一个简单的商城数据库来实现MySQL的垂直分库。
服务器规划
我们可以将服务器的规划简化成下面的表格所示。
| 主机名 | IP地址 | 安装的服务 | 业务数据库 |
|---|---|---|---|
| binghe151 | 192.168.175.151 | Mycat | shop（虚拟库） |
| binghe152 | 192.168.175.152 | MySQL | order_db（实体订单库） |
| binghe153 | 192.168.175.153 | MySQL | product_db（实体商品库） |
| binghe154 | 192.168.175.154 | MySQL | customer_db（实体用户库） |
安装MySQL
如果需要安装MySQL 8.x版本，则可以参考《MySQL之——源码编译MySQL8.x+升级gcc+升级cmake（亲测完整版）》
如果需要安装MySQL 5.x版本，则可以参考《MySQL之——CentOS6.5 编译安装MySQL5.6或5.7》
我这里，安装的是MySQL 8.x版本。安装其他的版本也可以。
创建实体数据库
- 创建order_db库
在binghe152服务器上创建订单库，SQL语句如下所示。
create database if not exists order_db;
- 创建product_db库
在binghe153服务器上创建商品库，SQL语句如下所示。
create database if not exists product_db;
- 创建customer_db库
在binghe154服务器上创建用户库，SQL语句如下所示。
create database if not exists customer_db;
有关各数据库的数据导入，大家可以下载【商城数据库SQL脚本文件.rar】文件，解压后，分别导入到相应的数据库即可。
创建Mycat连接MySQL的用户
分别在三个MySQL数据库中创建Mycat连接MySQL的用户，如果大家安装的是MySQL 8.x版本，则在MySQL命令行执行如下命令
CREATE USER 'mycat'@'192.168.175.%' IDENTIFIED BY 'mycat';
ALTER USER 'mycat'@'192.168.175.%' IDENTIFIED WITH mysql_native_password BY 'mycat';
GRANT SELECT, INSERT, UPDATE, DELETE ON *.* TO 'mycat'@'192.168.175.%';
FLUSH PRIVILEGES;
如果大家安装的是MySQL 5.x版本，则在MySQL命令行执行如下命令。
GRANT SELECT, INSERT, UPDATE, DELETE ON *.* TO 'mycat'@'192.168.175.%' IDENTIFIED BY 'mycat';
FLUSH PRIVILEGES;
下载并安装Mycat
我这里，下载的是Mycat的1.6.7.4 release版本，连接地址为：http://dl.mycat.io/1.6.7.4/ 。 下载后直接解压到binghe151服务器的/usr/local/mycat目录下即可。
这里，需要注意的是，Mycat是使用Java语言编写的，所以安装Mycat时，需要安装JDK，大家可以按照如下方式安装JDK
到JDK官网下载JDK 1.8版本，JDK1.8的下载地址为：https://www.oracle.com/technetwork/java/javase/downloads/jdk8-downloads-2133151.html。
我这里下载的JDK安装包版本为：jdk-8u212-linux-x64.tar.gz，如果JDK版本已更新，读者下载对应的版本即可。
将下载的JDK安装包解压到/usr/local/jdk1.8.0_212目录下。并配置JDK和Mycat的系统环境变量，如下所示。
vim /etc/profile
MYCAT_HOME=/usr/local/mycat
JAVA_HOME=/usr/local/jdk1.8.0_212
CLASS_PATH=.:$JAVA_HOME/lib
PATH=$JAVA_HOME/bin:$MYCAT_HOME/bin:$PATH
export JAVA_HOME MYCAT_HOME CLASS_PATH PATH
使系统环境变量生效，如下所示。
source /etc/profile
由于Mycat垂直分库不需要配置Mycat的分片规则，也就不需要配置rule.xml文件，此时只需要配置schema.xml文件和server.xml文件。
配置schema.xml文件
编辑后的schema.xml文件内容如下所示。
<?xml version="1.0"?>
<!DOCTYPE mycat:schema SYSTEM "schema.dtd">
<mycat:schema xmlns:mycat="http://io.mycat/">
<schema name="shop" checkSQLschema="false" sqlMaxLimit="1000">
<table name="order_master" primaryKey="order_id" dataNode = "ordb"/>
<table name="order_detail" primaryKey="order_detail_id" dataNode = "ordb"/>
<table name="order_cart" primaryKey="cart_id" dataNode = "ordb"/>
<table name="order_customer_addr" primaryKey="customer_addr_id" dataNode = "ordb"/>
<table name="region_info" primaryKey="region_id" dataNode = "ordb,prodb,custdb" type="global"/>
<table name="serial" primaryKey="id" dataNode = "ordb"/>
<table name="shipping_info" primaryKey="ship_id" dataNode = "ordb"/>
<table name="warehouse_info" primaryKey="w_id" dataNode = "ordb"/>
<table name="warehouse_proudct" primaryKey="wp_id" dataNode = "ordb"/>
<table name="product_brand_info" primaryKey="brand_id" dataNode = "prodb"/>
<table name="product_category" primaryKey="category_id" dataNode = "prodb"/>
<table name="product_comment" primaryKey="comment_id" dataNode = "prodb"/>
<table name="product_info" primaryKey="product_id" dataNode = "prodb"/>
<table name="product_pic_info" primaryKey="product_pic_id" dataNode = "prodb"/>
<table name="product_supplier_info" primaryKey="supplier_id" dataNode = "prodb"/>
<table name="customer_balance_log" primaryKey="balance_id" dataNode = "custdb"/>
<table name="customer_inf" primaryKey="customer_inf_id" dataNode = "custdb"/>
<table name="customer_level_inf" primaryKey="customer_level" dataNode = "custdb"/>
<table name="customer_login" primaryKey="customer_id" dataNode = "custdb"/>
<table name="customer_login_log" primaryKey="login_id" dataNode = "custdb"/>
<table name="customer_point_log" primaryKey="point_id" dataNode = "custdb"/>
</schema>
<dataNode name="ordb" dataHost="binghe152" database="order_db" />
<dataNode name="prodb" dataHost="binghe153" database="product_db" />
<dataNode name="custdb" dataHost="binghe154" database="customer_db" />
<dataHost name="binghe152" maxCon="1000" minCon="10" balance="1"
writeType="0" dbType="mysql" dbDriver="native" switchType="1" slaveThreshold="100">
<heartbeat>select user()</heartbeat>
<writeHost host="binghe52" url="192.168.175.152:3306" user="mycat" password="mycat"/>
</dataHost>
<dataHost name="binghe153" maxCon="1000" minCon="10" balance="1"
writeType="0" dbType="mysql" dbDriver="native" switchType="1" slaveThreshold="100">
<heartbeat>select user()</heartbeat>
<writeHost host="binghe53" url="192.168.175.153:3306" user="mycat" password="mycat"/>
</dataHost>
<dataHost name="binghe154" maxCon="1000" minCon="10" balance="1"
writeType="0" dbType="mysql" dbDriver="native" switchType="1" slaveThreshold="100">
<heartbeat>select user()</heartbeat>
<writeHost host="binghe54" url="192.168.175.154:3306" user="mycat" password="mycat"/>
</dataHost>
</mycat:schema>
由于地域信息表region_info在每个业务库中可能都需要使用到，此时，我们可以将地域信息表region_info配置成全局表，将其配置到每个业务库中。
如下所示。
<table name="region_info" primaryKey="region_id" dataNode = "ordb,prodb,custdb" type="global"/>
另外，还需要在每个实体业务库中，创建地域信息表region_info，并将数据导入到region_info表。此时，通过Mycat管理region_info表时，数据会分别同步到每个业务数据库中。
注意：当将某个数据表在Mycat中配置成全局表时，只有通过Mycat对这个表进行增加、删除和修改操作时，数据才会同步到每个实体数据库中。
配置server.xml文件
编辑后的server.xml文件如下所示。
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE mycat:server SYSTEM "server.dtd">
<mycat:server xmlns:mycat="http://io.mycat/">
<system>
<property name="useHandshakeV10">1</property>
<property name="defaultSqlParser">druidparser</property>
<property name="serverPort">3307</property>
<property name="managerPort">3308</property>
<property name="nonePasswordLogin">0</property>
<property name="bindIp">0.0.0.0</property>
<property name="charset">utf8mb4</property>
<property name="frontWriteQueueSize">2048</property>
<property name="txIsolation">2</property>
<property name="processors">2</property>
<property name="idleTimeout">1800000</property>
<property name="sqlExecuteTimeout">300</property>
<property name="useSqlStat">0</property>
<property name="useGlobleTableCheck">0</property>
<property name="sequenceHandlerType">2</property>
<property name="defaultMaxLimit">1000</property>
<property name="maxPacketSize">104857600</property>
</system>
<user name="mycat" defaultAccount="true">
<property name="usingDecrypt">1</property>
<property name="password">cTwf23RrpBCEmalp/nx0BAKenNhvNs2NSr9nYiMzHADeEDEfwVWlI6hBDccJjNBJqJxnunHFp5ae63PPnMfGYA==</property>
<property name="schemas">shop</property>
</user>
</mycat:server>
其中，在server.xml中，使用了Mycat对密码的加密功能。user标签下的password标签的值为登录Mycat的密码，此值使用Mycat提供的加密类对密码明文进行了加密。此类存在于Mycat安装目录下的lib目录下的Mycat-server-xxx-release.jar中的io.mycat.util.DecryptUtil类，可以使用如下命令对密码进行加密。
java java -cp /usr/local/mycat/lib/Mycat-server-xxx-release.jar io.mycat.util.DecryptUtil 0:mycat:mycat
其中0:mycat:mycat为运行Jar包的参数，0表示应用程序连接Mycat时使用密文密码；第一个mycat代表连接Mycat的用户名，也就是说为哪个用户的密码加密；第二个mycat代表需要加密的密码。
加密后的结果数据如下所示
cTwf23RrpBCEmalp/nx0BAKenNhvNs2NSr9nYiMzHADeEDEfwVWlI6hBDccJjNBJqJxnunHFp5ae63PPnMfGYA==
即user标签下的password属性的值。
如果按照上述方式为连接Mycat的密码加密后，需要在user标签下配置如下选项，否则无法正确连接Mycat。
<property name="usingDecrypt">1</property>
另外，在需要使用MySQL 8.x的mysql命令连接Mycat时，在server.xml文件的system标签下必须配置如下选项。
<property name="useHandshakeV10">1</property>
<property name="defaultSqlParser">druidparser</property>
否则，MySQL 8.x的mysql命令连接Mycat会失败。
连接Mycat
使用MySQL 8.x中的mysql命令连接Mycat，如下所示。
[root@binghe151 ~]# mysql -umycat -pmycat -h192.168.175.151 -P3307 --default-auth=mysql_native_password
mysql: [Warning] Using a password on the command line interface can be insecure.
Welcome to the MySQL monitor. Commands end with ; or \g.
Your MySQL connection id is 2
Server version: 5.6.29-mycat-xxx-release-20200228205020 MyCat Server (OpenCloudDB)
Copyright (c) 2000, 2019, Oracle and/or its affiliates. All rights reserved.
Oracle is a registered trademark of Oracle Corporation and/or its
affiliates. Other names may be trademarks of their respective
owners.
Type 'help;' or '\h' for help. Type '\c' to clear the current input statement.
mysql>
The use MySQL8.x mysql connection command Mycat, be noted that to add -default-auth = mysql_native_password option mysql command later. If you are using MySQL 5.x the mysql command line, you do not need to add -default-auth = mysql_native_password option on the command later.
Next, the logical view database and logic table under the myCat, as shown below.
mysql> SHOW DATABASES;
+----------+
| DATABASE |
+----------+
| shop |
+----------+
1 row in set (0.00 sec)
mysql> USE shop;
Database changed
mysql> SHOW TABLES;
+-----------------------+
| Tables in shop |
+-----------------------+
| customer_balance_log |
| customer_inf |
| customer_level_inf |
| customer_login |
| customer_login_log |
| customer_point_log |
| order_cart |
| order_customer_addr |
| order_detail |
| order_master |
| product_brand_info |
| product_category |
| product_comment |
| product_info |
| product_pic_info |
| product_supplier_info |
| region_info |
| serial |
| shipping_info |
| warehouse_info |
| warehouse_proudct |
+-----------------------+
21 rows in set (0.00 sec)
mysql> SELECT product_id, product_code, product_name FROM product_info LIMIT 10;
+------------+------------------+---------------------------------------+
| product_id | product_code | product_name |
+------------+------------------+---------------------------------------+
| 1 | 3700000000000001 | [Columbia]打底裤示例商品-1 |
| 2 | 3600000000000001 | [TheNorthFace]小脚裤示例商品-1 |
| 3 | 3500000000000001 | [李宁]九分裤示例商品-1 |
| 4 | 3400000000000001 | [LOWA]哈伦裤示例商品-1 |
| 5 | 3300000000000001 | [JACK&JONES]连体裤示例商品-1 |
| 6 | 3200000000000001 | [诺诗兰]牛仔裤示例商品-1 |
| 7 | 3100000000000001 | [骆驼]休闲裤示例商品-1 |
| 8 | 3000000000000001 | [金狐狸]风衣示例商品-1 |
| 9 | 2900000000000001 | [Columbia]小西装示例商品-1 |
| 10 | 2800000000000001 | [李宁]外套示例商品-1 |
+------------+------------------+---------------------------------------+
10 rows in set (0.01 sec)
So far, it completes the vertical component library based Mycat operation for MySQL.
