It is divided into single data source configuration and multi data source configuration. ShardingSphere-JDBC Supports all JDBC drivers and database connection pools.
In this example, the database driver is MySQL, and connection pool is HikariCP, which can be replaced with other database drivers and connection pools. When using ShardingSphere JDBC, the property name of the JDBC pool depends on the definition of the respective JDBC pool, and is not defined by ShardingSphere. For related processing, please refer to the class org.apache.shardingsphere.infra.datasource.pool.creator.DataSourcePoolCreator
. For example, with Alibaba Druid 1.2.9, using url
instead of jdbcUrl
in the example below is the expected behavior.
dataSources: # Data sources configuration, multiple <data-source-name> available
<data-source-name>: # Data source name
dataSourceClassName: # Data source class name
driverClassName: # Class name of database driver, ref property of connection pool
jdbcUrl: # Database URL, ref property of connection pool
username: # Database username, ref property of connection pool
password: # Database password, ref property of connection pool
# ... Other properties for data source pool
dataSources:
ds_1:
dataSourceClassName: com.zaxxer.hikari.HikariDataSource
driverClassName: com.mysql.jdbc.Driver
jdbcUrl: jdbc:mysql://localhost:3306/ds_1
username: root
password:
ds_2:
dataSourceClassName: com.zaxxer.hikari.HikariDataSource
driverClassName: com.mysql.jdbc.Driver
jdbcUrl: jdbc:mysql://localhost:3306/ds_2
username: root
password:
# Configure other data sources