2021-11-14

服务集成分库分表

---

前言

现在互联网公司数据量膨胀，以前的数据库配置已经不能满足需求，现在分库分表已经成为一种潮流。

---

一、水平分割

1、水平分库

以字段为依据，按照一定策略，将一个库中的数据拆分到多个库中。每个库的结构都一样；数据都不一样；所有库的并集是全量数据。

2、水平分表

以字段为依据，按照一定策略，将一个表中的数据拆分到多个表中。每个表的结构都一样；数据都不一样；所有表的并集是全量数据。

二、Shard-jdbc 中间件

1.架构

2.架构特点

Sharding-JDBC直接封装JDBC API，旧代码迁移成本几乎为零。适用于任何基于Java的ORM框架，如Hibernate、Mybatis等 。而且可基于任何第三方的数据库连接池，如DBCP、C3P0、 BoneCP、Druid等。
Sharding-JDBC以jar包形式提供服务，无proxy代理层，无需额外部署，无其他依赖。分片策略灵活，可支持等号、between、in等多维度分片，也可支持多分片键。SQL解析功能完善，支持聚合、分组、排序、limit、or等查询。

三、项目案例

数据库配置

一台基础库映射（shard_one）
两台库做分库分表（shard_two，shard_three）。
表使用：table_one，table_two

spring:
  datasource:
    # 数据源：shard_one
    dataOne:
      type: com.alibaba.druid.pool.DruidDataSource
      druid:
        driverClassName: com.mysql.jdbc.Driver
        url: jdbc:mysql://localhost:3306/shard_one?useUnicode=true&characterEncoding=UTF8&zeroDateTimeBehavior=convertToNull&useSSL=false
        username: root
        password: 123
        initial-size: 10
        max-active: 100
        min-idle: 10
        max-wait: 60000
        pool-prepared-statements: true
        max-pool-prepared-statement-per-connection-size: 20
        time-between-eviction-runs-millis: 60000
        min-evictable-idle-time-millis: 300000
        max-evictable-idle-time-millis: 60000
        validation-query: SELECT 1 FROM DUAL
        # validation-query-timeout: 5000
        test-on-borrow: false
        test-on-return: false
        test-while-idle: true
        connectionProperties: druid.stat.mergeSql=true;druid.stat.slowSqlMillis=5000
    # 数据源：shard_two
    dataTwo:
      type: com.alibaba.druid.pool.DruidDataSource
      druid:
        driverClassName: com.mysql.jdbc.Driver
        url: jdbc:mysql://localhost:3306/shard_two?useUnicode=true&characterEncoding=UTF8&zeroDateTimeBehavior=convertToNull&useSSL=false
        username: root
        password: 123
        initial-size: 10
        max-active: 100
        min-idle: 10
        max-wait: 60000
        pool-prepared-statements: true
        max-pool-prepared-statement-per-connection-size: 20
        time-between-eviction-runs-millis: 60000
        min-evictable-idle-time-millis: 300000
        max-evictable-idle-time-millis: 60000
        validation-query: SELECT 1 FROM DUAL
        # validation-query-timeout: 5000
        test-on-borrow: false
        test-on-return: false
        test-while-idle: true
        connectionProperties: druid.stat.mergeSql=true;druid.stat.slowSqlMillis=5000
    # 数据源：shard_three
    dataThree:
      type: com.alibaba.druid.pool.DruidDataSource
      druid:
        driverClassName: com.mysql.jdbc.Driver
        url: jdbc:mysql://localhost:3306/shard_three?useUnicode=true&characterEncoding=UTF8&zeroDateTimeBehavior=convertToNull&useSSL=false
        username: root
        password: 123
        initial-size: 10
        max-active: 100
        min-idle: 10
        max-wait: 60000
        pool-prepared-statements: true
        max-pool-prepared-statement-per-connection-size: 20
        time-between-eviction-runs-millis: 60000
        min-evictable-idle-time-millis: 300000
        max-evictable-idle-time-millis: 60000
        validation-query: SELECT 1 FROM DUAL
        # validation-query-timeout: 5000
        test-on-borrow: false
        test-on-return: false
        test-while-idle: true
        connectionProperties: druid.stat.mergeSql=true;druid.stat.slowSqlMillis=5000

数据库分库策略

/**
 * 数据库映射计算
 */
public class DataSourceAlg implements PreciseShardingAlgorithm<string> {

    private static Logger LOG = LoggerFactory.getLogger(DataSourceAlg.class);
    @Override
    public String doSharding(Collection<string> names, PreciseShardingValue<string> value) {
        LOG.debug("分库算法参数 {},{}",names,value);
        int hash = HashUtil.rsHash(String.valueOf(value.getValue()));
        return "ds_" + ((hash % 2) + 2) ;
    }
}

数据表1分表策略

/**
 * 分表算法
 */
public class TableOneAlg implements PreciseShardingAlgorithm<string> {
    private static Logger LOG = LoggerFactory.getLogger(TableOneAlg.class);
    /**
     * 该表每个库分5张表
     */
    @Override
    public String doSharding(Collection<string> names, PreciseShardingValue<string> value) {
        LOG.debug("分表算法参数 {},{}",names,value);
        int hash = HashUtil.rsHash(String.valueOf(value.getValue()));
        return "table_one_" + (hash % 5+1);
    }
}

数据表2分表策略

/**
 * 分表算法
 */
public class TableTwoAlg implements PreciseShardingAlgorithm<string> {
    private static Logger LOG = LoggerFactory.getLogger(TableTwoAlg.class);
    /**
     * 该表每个库分5张表
     */
    @Override
    public String doSharding(Collection<string> names, PreciseShardingValue<string> value) {
        LOG.debug("分表算法参数 {},{}",names,value);
        int hash = HashUtil.rsHash(String.valueOf(value.getValue()));
        return "table_two_" + (hash % 5+1);
    }
}

数据源集成配置

/**
 * 数据库分库分表配置
 */
@Configuration
public class ShardJdbcConfig {
    // 省略了 druid 配置，源码中有
    /**
     * Shard-JDBC 分库配置
     */
    @Bean
    public DataSource dataSource (@Autowired DruidDataSource dataOneSource,
                                  @Autowired DruidDataSource dataTwoSource,
                                  @Autowired DruidDataSource dataThreeSource) throws Exception {
        ShardingRuleConfiguration shardJdbcConfig = new ShardingRuleConfiguration();
        shardJdbcConfig.getTableRuleConfigs().add(getTableRule01());
        shardJdbcConfig.getTableRuleConfigs().add(getTableRule02());
        shardJdbcConfig.setDefaultDataSourceName("ds_0");
        Map<string,datasource> dataMap = new LinkedHashMap&lt;&gt;() ;
        dataMap.put("ds_0",dataOneSource) ;
        dataMap.put("ds_2",dataTwoSource) ;
        dataMap.put("ds_3",dataThreeSource) ;
        Properties prop = new Properties();
        return ShardingDataSourceFactory.createDataSource(dataMap, shardJdbcConfig, new HashMap&lt;&gt;(), prop);
    }

    /**
     * Shard-JDBC 分表配置
     */
    private static TableRuleConfiguration getTableRule01() {
        TableRuleConfiguration result = new TableRuleConfiguration();
        result.setLogicTable("table_one");
        result.setActualDataNodes("ds_${2..3}.table_one_${1..5}");
        result.setDatabaseShardingStrategyConfig(new StandardShardingStrategyConfiguration("phone", new DataSourceAlg()));
        result.setTableShardingStrategyConfig(new StandardShardingStrategyConfiguration("phone", new TableOneAlg()));
        return result;
    }
    private static TableRuleConfiguration getTableRule02() {
        TableRuleConfiguration result = new TableRuleConfiguration();
        result.setLogicTable("table_two");
        result.setActualDataNodes("ds_${2..3}.table_two_${1..5}");
        result.setDatabaseShardingStrategyConfig(new StandardShardingStrategyConfiguration("phone", new DataSourceAlg()));
        result.setTableShardingStrategyConfig(new StandardShardingStrategyConfiguration("phone", new TableTwoAlg()));
        return result;
    }
}

测试

@RestController
public class ShardController {
    @Resource
    private ShardService shardService ;
    /**
     * 1、建表流程
     */
    @RequestMapping("/createTable")
    public String createTable (){
        shardService.createTable();
        return "success" ;
    }
    /**
     * 2、生成表 table_one 数据
     */
    @RequestMapping("/insertOne")
    public String insertOne (){
        shardService.insertOne();
        return "SUCCESS" ;
    }
    /**
     * 3、生成表 table_two 数据
     */
    @RequestMapping("/insertTwo")
    public String insertTwo (){
        shardService.insertTwo();
        return "SUCCESS" ;
    }
    /**
     * 4、查询表 table_one 数据
     */
    @RequestMapping("/selectOneByPhone/{phone}")
    public TableOne selectOneByPhone (@PathVariable("phone") String phone){
        return shardService.selectOneByPhone(phone);
    }
    /**
     * 5、查询表 table_one 数据
     */
    @RequestMapping("/selectTwoByPhone/{phone}")
    public TableTwo selectTwoByPhone (@PathVariable("phone") String phone){
        return shardService.selectTwoByPhone(phone);
    }
}

---