Distributed transaction have been integrated into ShardingDataSource, you can use TransactionTypeHolder to modify transaction type before creating ShardingConnection.
<dependency>
<groupId>org.apache.shardingsphere</groupId>
<artifactId>sharding-transaction-xa-core</artifactId>
<version>${shardingsphere.version}</version>
</dependency>
TransactionTypeHolder.set(TransactionType.XA);
Default ShardingSphere XA transaction manager is Atomikos. xa_tx.log generated in the project log is necessary for the recovery when XA crashes. Please do not delete it.
Or you can add jta.properties in classpath of the program to customize Atomikos configurations.
For detailed configuration rules, please refer to the official documentation of Atomikos.
Currently, we have integrated saga and seata into shardingsphere.
<!-- saga transaction -->
<dependency>
<groupId>io.shardingsphere</groupId>
<artifactId>sharding-transaction-base-saga</artifactId>
<version>${shardingsphere-spi-impl.version}</version>
</dependency>
${shardingsphere-spi-impl.version} mentioned has not been posted to the central maven repository, so you need to install it by yourself.Project Address: shardingsphere-spi-impl
<!-- seata at transaction -->
<dependency>
<groupId>org.apache.shardingsphere</groupId>
<artifactId>sharding-transaction-base-seata-at</artifactId>
<version>${sharding-sphere.version}</version>
</dependency>
TransactionTypeHolder.set(TransactionType.BASE);
You can add saga.properties in the project classpath to customize Saga configurations. When saga.persistence.enabled=true, Saga engine will persist event log through JDBC.
Configuration properties and explanations are as follow:
| Property | Default Value | Explanation |
|---|---|---|
| saga.actuator.executor.size | 5 | Saga actuator thread pool size |
| saga.actuator.transaction.max.retries | 5 | Maximum retry times |
| saga.actuator.compensation.max.retries | 5 | Maximum compensation times |
| saga.actuator.transaction.retry.delay.milliseconds | 5000 | Retry interval |
| saga.actuator.compensation.retry.delay.milliseconds | 3000 | Compensation interval |
| saga.persistence.enabled | false | Persistence for event log |
| saga.persistence.ds.url | No | JDBC url |
| saga.persistence.ds.username | No | User name |
| saga.persistence.ds.password | No | Password |
| saga.persistence.ds.max.pool.size | 50 | Maximum connection |
| saga.persistence.ds.min.pool.size | 1 | Minimum connection |
| saga.persistence.ds.max.life.time.milliseconds | 0 (unrestricted) | Maximum life time (millisecond) |
| saga.persistence.ds.idle.timeout.milliseconds | 60 * 1000 | Idle timeout (millisecond) |
| saga.persistence.ds.connection.timeout.milliseconds | 30 * 1000 | Connection timeout (millisecond) |
Saga event log table structure
-- MySQL init table SQL
CREATE TABLE IF NOT EXISTS saga_event(
id BIGINT AUTO_INCREMENT PRIMARY KEY,
saga_id VARCHAR(255) null,
type VARCHAR(255) null,
content_json TEXT null,
create_time TIMESTAMP DEFAULT CURRENT_TIMESTAMP,
INDEX saga_id_index(saga_id)
)ENGINE=InnoDB DEFAULT CHARSET=utf8
saga_event table will be created automatically when your put this DDL into schema-init.sql located in classpath.
1.Download seata server according to seata-work-shop, refer to step6 and step7 is OK.
2.Create undo_log table in each physical database.(only mysql supported before seata version 0.8.X)
CREATE TABLE IF NOT EXISTS `undo_log` (
`id` bigint(20) NOT NULL AUTO_INCREMENT,
`branch_id` bigint(20) NOT NULL,
`xid` varchar(100) NOT NULL,
`rollback_info` longblob NOT NULL,
`log_status` int(11) NOT NULL,
`log_created` datetime NOT NULL,
`log_modified` datetime NOT NULL,
`ext` varchar(100) DEFAULT NULL,
PRIMARY KEY (`id`),
UNIQUE KEY `ux_undo_log` (`xid`,`branch_id`)
) ENGINE=InnoDB AUTO_INCREMENT=1 DEFAULT CHARSET=utf8;
3.Config seata.conf file in classpath.
client {
application.id = raw-jdbc ## application unique id.
transaction.service.group = raw-jdbc-group ## transaction group
}