This version is still in development and is not considered stable yet. For the latest stable version, please use Spring Framework 6.2.1!

Understanding the Spring Framework Transaction Abstraction

The key to the Spring transaction abstraction is the notion of a transaction strategy. A transaction strategy is defined by a TransactionManager, specifically the org.springframework.transaction.PlatformTransactionManager interface for imperative transaction management and the org.springframework.transaction.ReactiveTransactionManager interface for reactive transaction management. The following listing shows the definition of the PlatformTransactionManager API:

public interface PlatformTransactionManager extends TransactionManager {

	TransactionStatus getTransaction(TransactionDefinition definition) throws TransactionException;

	void commit(TransactionStatus status) throws TransactionException;

	void rollback(TransactionStatus status) throws TransactionException;
}

This is primarily a service provider interface (SPI), although you can use it programmatically from your application code. Because PlatformTransactionManager is an interface, it can be easily mocked or stubbed as necessary. It is not tied to a lookup strategy, such as JNDI. PlatformTransactionManager implementations are defined like any other object (or bean) in the Spring Framework IoC container. This benefit alone makes Spring Framework transactions a worthwhile abstraction, even when you work with JTA. You can test transactional code much more easily than if it used JTA directly.

Again, in keeping with Spring’s philosophy, the TransactionException that can be thrown by any of the PlatformTransactionManager interface’s methods is unchecked (that is, it extends the java.lang.RuntimeException class). Transaction infrastructure failures are almost invariably fatal. In rare cases where application code can actually recover from a transaction failure, the application developer can still choose to catch and handle TransactionException. The salient point is that developers are not forced to do so.

The getTransaction(..) method returns a TransactionStatus object, depending on a TransactionDefinition parameter. The returned TransactionStatus might represent a new transaction or can represent an existing transaction, if a matching transaction exists in the current call stack. The implication in this latter case is that, as with Jakarta EE transaction contexts, a TransactionStatus is associated with a thread of execution.

Spring also provides a transaction management abstraction for reactive applications that make use of reactive types or Kotlin Coroutines. The following listing shows the transaction strategy defined by org.springframework.transaction.ReactiveTransactionManager:

public interface ReactiveTransactionManager extends TransactionManager {

	Mono<ReactiveTransaction> getReactiveTransaction(TransactionDefinition definition) throws TransactionException;

	Mono<Void> commit(ReactiveTransaction status) throws TransactionException;

	Mono<Void> rollback(ReactiveTransaction status) throws TransactionException;
}

The reactive transaction manager is primarily a service provider interface (SPI), although you can use it programmatically from your application code. Because ReactiveTransactionManager is an interface, it can be easily mocked or stubbed as necessary.

The TransactionDefinition interface specifies:

  • Propagation: Typically, all code within a transaction scope runs in that transaction. However, you can specify the behavior if a transactional method is run when a transaction context already exists. For example, code can continue running in the existing transaction (the common case), or the existing transaction can be suspended and a new transaction created. Spring offers all of the transaction propagation options familiar from EJB CMT. To read about the semantics of transaction propagation in Spring, see Transaction Propagation.

  • Isolation: The degree to which this transaction is isolated from the work of other transactions. For example, can this transaction see uncommitted writes from other transactions?

  • Timeout: How long this transaction runs before timing out and being automatically rolled back by the underlying transaction infrastructure.

  • Read-only status: You can use a read-only transaction when your code reads but does not modify data. Read-only transactions can be a useful optimization in some cases, such as when you use Hibernate.

These settings reflect standard transactional concepts. If necessary, refer to resources that discuss transaction isolation levels and other core transaction concepts. Understanding these concepts is essential to using the Spring Framework or any transaction management solution.

The TransactionStatus interface provides a simple way for transactional code to control transaction execution and query transaction status. The concepts should be familiar, as they are common to all transaction APIs. The following listing shows the TransactionStatus interface:

public interface TransactionStatus extends TransactionExecution, SavepointManager, Flushable {

	@Override
	boolean isNewTransaction();

	boolean hasSavepoint();

	@Override
	void setRollbackOnly();

	@Override
	boolean isRollbackOnly();

	void flush();

	@Override
	boolean isCompleted();
}

Regardless of whether you opt for declarative or programmatic transaction management in Spring, defining the correct TransactionManager implementation is absolutely essential. You typically define this implementation through dependency injection.

TransactionManager implementations normally require knowledge of the environment in which they work: JDBC, JTA, Hibernate, and so on. The following examples show how you can define a local PlatformTransactionManager implementation (in this case, with plain JDBC.)

You can define a JDBC DataSource by creating a bean similar to the following:

<bean id="dataSource" class="org.apache.commons.dbcp.BasicDataSource" destroy-method="close">
	<property name="driverClassName" value="${jdbc.driverClassName}" />
	<property name="url" value="${jdbc.url}" />
	<property name="username" value="${jdbc.username}" />
	<property name="password" value="${jdbc.password}" />
</bean>

The related PlatformTransactionManager bean definition then has a reference to the DataSource definition. It should resemble the following example:

<bean id="txManager" class="org.springframework.jdbc.datasource.DataSourceTransactionManager">
	<property name="dataSource" ref="dataSource"/>
</bean>

If you use JTA in a Jakarta EE container, then you use a container DataSource, obtained through JNDI, in conjunction with Spring’s JtaTransactionManager. The following example shows what the JTA and JNDI lookup version would look like:

<?xml version="1.0" encoding="UTF-8"?>
<beans xmlns="http://www.springframework.org/schema/beans"
	xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
	xmlns:jee="http://www.springframework.org/schema/jee"
	xsi:schemaLocation="
		http://www.springframework.org/schema/beans
		https://www.springframework.org/schema/beans/spring-beans.xsd
		http://www.springframework.org/schema/jee
		https://www.springframework.org/schema/jee/spring-jee.xsd">

	<jee:jndi-lookup id="dataSource" jndi-name="jdbc/jpetstore"/>

	<bean id="txManager" class="org.springframework.transaction.jta.JtaTransactionManager" />

	<!-- other <bean/> definitions here -->

</beans>

The JtaTransactionManager does not need to know about the DataSource (or any other specific resources) because it uses the container’s global transaction management infrastructure.

The preceding definition of the dataSource bean uses the <jndi-lookup/> tag from the jee namespace. For more information see The JEE Schema.
If you use JTA, your transaction manager definition should look the same, regardless of what data access technology you use, be it JDBC, Hibernate JPA, or any other supported technology. This is due to the fact that JTA transactions are global transactions, which can enlist any transactional resource.

In all Spring transaction setups, application code does not need to change. You can change how transactions are managed merely by changing configuration, even if that change means moving from local to global transactions or vice versa.

Hibernate Transaction Setup

You can also easily use Hibernate local transactions, as shown in the following examples. In this case, you need to define a Hibernate LocalSessionFactoryBean, which your application code can use to obtain Hibernate Session instances.

The DataSource bean definition is similar to the local JDBC example shown previously and, thus, is not shown in the following example.

If the DataSource (used by any non-JTA transaction manager) is looked up through JNDI and managed by a Jakarta EE container, it should be non-transactional, because the Spring Framework (rather than the Jakarta EE container) manages the transactions.

The txManager bean in this case is of the HibernateTransactionManager type. In the same way as the DataSourceTransactionManager needs a reference to the DataSource, the HibernateTransactionManager needs a reference to the SessionFactory. The following example declares sessionFactory and txManager beans:

<bean id="sessionFactory" class="org.springframework.orm.hibernate5.LocalSessionFactoryBean">
	<property name="dataSource" ref="dataSource"/>
	<property name="mappingResources">
		<list>
			<value>org/springframework/samples/petclinic/hibernate/petclinic.hbm.xml</value>
		</list>
	</property>
	<property name="hibernateProperties">
		<value>
			hibernate.dialect=${hibernate.dialect}
		</value>
	</property>
</bean>

<bean id="txManager" class="org.springframework.orm.hibernate5.HibernateTransactionManager">
	<property name="sessionFactory" ref="sessionFactory"/>
</bean>

If you use Hibernate and Jakarta EE container-managed JTA transactions, you should use the same JtaTransactionManager as in the previous JTA example for JDBC, as the following example shows. Also, it is recommended to make Hibernate aware of JTA through its transaction coordinator and possibly also its connection release mode configuration:

<bean id="sessionFactory" class="org.springframework.orm.hibernate5.LocalSessionFactoryBean">
	<property name="dataSource" ref="dataSource"/>
	<property name="mappingResources">
		<list>
			<value>org/springframework/samples/petclinic/hibernate/petclinic.hbm.xml</value>
		</list>
	</property>
	<property name="hibernateProperties">
		<value>
			hibernate.dialect=${hibernate.dialect}
			hibernate.transaction.coordinator_class=jta
			hibernate.connection.handling_mode=DELAYED_ACQUISITION_AND_RELEASE_AFTER_STATEMENT
		</value>
	</property>
</bean>

<bean id="txManager" class="org.springframework.transaction.jta.JtaTransactionManager"/>

Or alternatively, you may pass the JtaTransactionManager into your LocalSessionFactoryBean for enforcing the same defaults:

<bean id="sessionFactory" class="org.springframework.orm.hibernate5.LocalSessionFactoryBean">
	<property name="dataSource" ref="dataSource"/>
	<property name="mappingResources">
		<list>
			<value>org/springframework/samples/petclinic/hibernate/petclinic.hbm.xml</value>
		</list>
	</property>
	<property name="hibernateProperties">
		<value>
			hibernate.dialect=${hibernate.dialect}
		</value>
	</property>
	<property name="jtaTransactionManager" ref="txManager"/>
</bean>

<bean id="txManager" class="org.springframework.transaction.jta.JtaTransactionManager"/>