关于springboot整合sharding-jdbc官网有4种方式，而网上千篇一律都是用.properties文件。我在这里主要是用yml+bean两种方式结合，bean主要是为了实现分片键分片规则更灵活。
1.1引入maven依赖这里有个小坑，我一开始引入是最新的sharding-jdbc和mybatis-plus的依赖，结果启动项目直接报错。所以个人不建议引用最新的包。
1.2配置分库分表和读写分离（核心）org.apache.shardingsphere sharding-jdbc-spring-boot-starter4.0.0-RC1
spring:
shardingsphere:
datasource:
names: "search-center-00,search-center-00-slave,search-center-01,search-center-01-slave"
search-center-00:
type: com.alibaba.druid.pool.DruidDataSource
driver-class-name: com.mysql.cj.jdbc.Driver
url: jdbc:mysql://localhost:3306/search_center_00?useUnicode=true&characterEncoding=utf8&serverTimezone=UTC
username: root
password: 123456
search-center-00-slave:
type: com.alibaba.druid.pool.DruidDataSource
driver-class-name: com.mysql.cj.jdbc.Driver
url: jdbc:mysql://localhost:3307/search_center_00?useUnicode=true&characterEncoding=utf8&serverTimezone=UTC
username: root
password: 123456
search-center-01:
type: com.alibaba.druid.pool.DruidDataSource
driver-class-name: com.mysql.cj.jdbc.Driver
url: jdbc:mysql://localhost:3306/search_center_01?useUnicode=true&characterEncoding=utf8&serverTimezone=UTC
username: root
password: 123456
search-center-01-slave:
type: com.alibaba.druid.pool.DruidDataSource
driver-class-name: com.mysql.cj.jdbc.Driver
url: jdbc:mysql://localhost:3307/search_center_01?useUnicode=true&characterEncoding=utf8&serverTimezone=UTC
username: root
password: 123456
sharding:
master-slave-rules:
search-center-00:
master-data-source-name: search-center-00
slave-data-source-names: search-center-00-slave
search-center-01:
master-data-source-name: search-center-01
slave-data-source-names: search-center-01-slave
tables:
user:
actual-data-nodes: search-center-$->{(0..1).collect{t ->t.toString().padLeft(2,'0')}}.user_$->{(0..1).collect{t ->t.toString().padLeft(2,'0')}}
database-strategy:
standard:
shardingColumn: id
precise-algorithm-class-name: com.gory.searchcenter.config.DatabaseAlgorithm
table-strategy:
standard:
sharding-column: id
precise-algorithm-class-name: com.gory.searchcenter.config.TableAlgorithm
props:
sql:
show: true
names:自定义数据库名称，然后对这些名称进行具体的配置，如：url，password，username等
tables:对每个表进行分库分表。例如，这里对user表进行分库分表，
actual-data-nodes：则是user表分库分表的节点分配。
database-strategy：是分库的策略，我这里是根据id，然后通过bean方式更加灵活配置分库规则，
table-strategy：则是分表的策略，配置和分库大概一致。
master-slave-rules:设置主从实现读写分离，
DatabaseAlgorithm
public class DatabaseAlgorithm implements PreciseShardingAlgorithm{ @Override public String doSharding(Collection databaseNames, PreciseShardingValue shardingValue) { Long shardResult = Math.abs((long) shardingValue.getValue().hashCode()) % 2; for (String databaseName : databaseNames) { String shardValue = String.format("%02d", shardResult); if (databaseName.endsWith(shardValue)){ System.out.println("shardingValue:"+shardingValue+", databaseName:" + databaseName); return databaseName; } } throw new UnsupportedOperationException(); } }
TableAlgorithm
public class TableAlgorithm implements PreciseShardingAlgorithm{ @Override public String doSharding(Collection tableNames, PreciseShardingValue shardingValue) { Long shardResult = Math.abs((long) shardingValue.getValue().hashCode()) / 2 % 2; for (String tableName : tableNames) { String shardValue = String.format("%02d", shardResult); if (tableName.endsWith(shardValue)){ System.out.println("shardingValue:"+shardingValue+", tableName:" + tableName); return tableName; } } throw new UnsupportedOperationException(); } }
随便说一下，这里的分片主键id是用雪花算法设值的。你的分片键的赋值最好和你的分片规则对应上。
二、整合mybatis-plus上一节已经说过了，引入最新的maven依赖会报错，所以我还是不建议用最新的。
2.1 引入maven依赖2.2 mybatis-plus的配置com.baomidou mybatis-plus-boot-starter3.2.0 com.baomidou mybatis-plus-generator3.5.1
mybatis-plus:
mapper-locations: classpath:/mapper/*.xml
configuration:
log-impl: org.apache.ibatis.logging.stdout.StdOutImpl
再启动类加上包扫表注解
@MapperScan({"com.gory.searchcenter.mapper"})
三、数据库主从配置
在上两节已经能完成分库分表了，但是因为数据库没设置主从，从表没有数据库，导致查询没有数据。因为此文章主要是为了学习，所以mysql的配置是在window环境的。
3.1my.ini配置 主库#主库id不能重复
server_id=1
#开启bin-log
log_bin=E:libmysql-5.7.30-winx64master-bin.log
#mysql清除过期日志的时间与从库设置一致
expire_logs_days = 1
#设置需要同步的数据库
binlog-do-db=search_center_00
binlog-do-db=search_center_01
------------------------------------------------------------ |
server_id=2
#开启bin-log
log_bin=E:libmysql-5.7.30-winx64-slaveslave-bin.log
expire_logs_days = 1
#设置需要同步的数据库
binlog-do-db=search_center_00
binlog-do-db=search_center_01
------------------------------------------------------------ |
值得一提，设置多个同步表，不能用逗号隔开。
3.1 sql配置 主库开通从服务器（ Slave ）访问主服务器（ Master ）的权限
grant replication slave on . to ‘root’@‘127.0.0.1’ identified by ‘123456’;
查看授权信息
show master status ;
从库停止同步
stop slave;
设置主库的信息，将从库和主库联系起来
change master to
master_host=’127.0.0.1’,
master_user=’root’,
master_password=’123456’,
master_log_file=’log-bin.000008’,
master_log_pos=107;
master_log_file和master_log_pos这两个 配置是和主 服务器 show master status对应的
启动同步
start salve;
查看slave启动状态
show slave status；
如果出现，则表明，启动成功。
Slave_IO_Running:Yes
Slave_SQL_Running:Yes
四、完整的项目代码完整的代码：https://github.com/gorylee/SearchCenter
