Question

May I know by configuring the data source in Spring like this:

<bean id="dataSource" class="org.springframework.jndi.JndiObjectFactoryBean">
  <property name="jndiName" value="jdbc/dev"/>
  <property name="lookupOnStartup" value="false"/>
  <property name="cache" value="true"/>
  <property name="proxyInterface" value="javax.sql.DataSource"/>
</bean>

<bean id="sessionFactory" class="org.springframework.orm.hibernate3.LocalSessionFactoryBean">
  <property name="dataSource">
    <ref bean="dataSource"/>
  </property>
  ...
</bean>

And configuring my BOC and DAO object in Spring like this:

<bean id="Dao" class="com.dao.impl.DaoImpl">
  <property name="sessionFactory" ref="sessionFactory"/>
</bean>

<bean id="Bo" class="com.bo.impl.BoImpl">
  <property name="theDao">
    <ref local="Dao"/>
  </property>
</bean>

Currently I am testing it with 3 users, 1 successfully insert data into DB, 1 is hung, and 1 is missing in action, meaning there is no response, no log was capture in Websphere Application Server. With 3 users concurrently using the app has failed the test case, may I know how could I ensure all these are thread safe when come to a situation when there are 1000 users using the app concurrently?

---

UPDATE

In response to @Adrian Shum query:

Regarding the BO thing, I’m not sure what pattern is this. But I’m BOC is stand for Business Object Controller, the purpose of having this unit class is to separate the business logic from DAO object. Eventually this will end up the XHTML/JSP is the front-end, BO is the business controller, and DAO is concern about hibernate and query construction.

In order to retrieve the session factory, every DAO object must extends the HibernateDaoSupport, this is how Spring-Hibernate Integration work according to this tutorial. Here is some code snippet:

class DAO extends HibernateDaoSupport implements IDao {

  public void save( Pojo pojo ) {
    getHibernateTemplate().save(pojo);
  }

  public void update( Pojo pojo ) {
    getHibernateTemplate().update(pojo);
  }

  public void delete( Pojo pojo ) {
    getHibernateTemplate().delete(pojo);
  }
}

I know that Spring object are singleton by default. Does this means each thread will have only ONE object or the whole JVM instance will have only ONE object? What if I declare those BO and DAO object as session scope like this:

<bean id="Dao" class="com.dao.impl.DaoImpl" scope="session">
  <property name="sessionFactory" ref="sessionFactory"/>
</bean>

<bean id="Bo" class="com.bo.impl.BoImpl" scope="session">
  <property name="theDao">
    <ref local="Dao"/>
  </property>
</bean>

Regarding the data update or retrieval, this could happen as the 3 users that we are testing on is actaully targeting on the same record. There might be a lock as I notice that there is a function doing this code:

Query queryA = session.createQuery("Delete From TableA where fieldA = :theID");
queryA.setParameter("theID", "XX");
queryA.executeUpdate();

Query queryB = session.createQuery("Delete From TableB where fieldB = :theID");
queryB.setParameter("theID", "YY");
queryB.executeUpdate();

// update tableB object
session.save(tableBObj);

// update each tableA object
for(TableAObj obj : TableAObjList) {                
  session.save(obj);
  session.flush();
  session.evict(obj);
}

The TableA(slave) and TableB(master) has relationship in each other. I know there is a database design between TableA and TableB but this is beyond of this question. I’m just curious whether this function could cause the concurrent issue even though I made this class as singleton?

Answer 1

I have the problem resolved. It is due to the DB2 failed to handle concurrency issues by adding a new column into the table, and make it as a primary key.