Atomikos Forum

We had been on 4.0.0M4 for a very long time and are trying out 4.0.2 right now. For the upgrade we have simply changing Maven dependency version from 4.0.0M4 to 4.0.2. We have started seeing a large number of operation time outs after the upgrade. Scenarios that work fine with 4.0.0M4 time out after upgrading to 4.0.2. We have not made any other changes like configuration settings.

We are using Atomikos with Spring Data JPA and Hibernate. The relevant configuration is shown below:

    <bean class="org.springframework.orm.jpa.LocalContainerEntityManagerFactoryBean">
      <property name="jpaProperties">
        <props>
          ...
          <prop key="hibernate.transaction.jta.platform">com.atomikos.icatch.jta.hibernate4.AtomikosPlatform</prop>
        </props>
      </property>
      <property name="jtaDataSource">
        <bean class="com.atomikos.jdbc.AtomikosDataSourceBean" init-method="init" destroy-method="close">
          <property name="uniqueResourceName" value="dataSource"/>
          <property name="xaDataSource">
            <bean class="com.mysql.jdbc.jdbc2.optional.MysqlXADataSource" lazy-init="true">
              <property name="password" value="someunknownpassword"/>
              <property name="url" value="jdbc:mysql://localhost/somedatabase"/>
              <property name="user" value="someuser"/>
            </bean>
          </property>
          <property name="borrowConnectionTimeout" value="3000"/>
          <property name="defaultIsolationLevel" value="4"/>
          <property name="maxPoolSize" value="25"/>
          <property name="minPoolSize" value="5"/>
        </bean>
      </property>
    </bean>

The error is:

    org.springframework.transaction.UnexpectedRollbackException: JTA transaction unexpectedly rolled back (maybe due to a timeout); nested exception is javax.transaction.RollbackException: Prepare: NO vote
Caused by: javax.transaction.RollbackException: Prepare: NO vote
Caused by: com.atomikos.icatch.RollbackException: Prepare: NO vote
        at com.atomikos.icatch.imp.ActiveStateHandler.prepare(ActiveStateHandler.java:202)
        at com.atomikos.icatch.imp.CoordinatorImp.prepare(CoordinatorImp.java:523)
        at com.atomikos.icatch.imp.CoordinatorImp.terminate(CoordinatorImp.java:687)
        at com.atomikos.icatch.imp.CompositeTransactionImp.commit(CompositeTransactionImp.java:282)
        at com.atomikos.icatch.jta.TransactionImp.commit(TransactionImp.java:172)

I can confirm that downgrading to 4.0.0M4 gets rid of the time outs immediately for exactly the same set of operations.

4.0.1 has similar behaviour with time outs for even shorter operations.

Has a new time out value been introduced or the value for an existing one changed in 4.0.1?

We changed things but hard to see what is causing this.

Can you check with:

Atomikos INFO logs in 4.0.0M4 versus
Atomikos DEBUG logs in 4.0.2

...so the timeouts are shown,
and post the results here?

Thanks

Any update on the requested information?

Currently the 4.0.3 is available, and next week we will publish the final 4.0.4 community release after which fixes are reserved for customers.

So this is your last opportunity for us to fix this in the short term without a support budget.

Thanks!