Enabling AspectJ for a Quartz Scheduler Job - spring

I have created bean which will be executed by a simple trigger quartz scheduler. When I enable AspectJ within my spring-context.xml , my scheduler job is not getting triggered
Here is a snippet of my spring-context.xml
<!-- Scheduler Factory -->
<bean class="org.springframework.scheduling.quartz.SchedulerFactoryBean">
<property name="triggers">
<list>
<ref bean="simpleTrigger" />
</list>
</property>
</bean>
<!-- Trigger -->
<bean id="simpleTrigger"
class="org.springframework.scheduling.quartz.SimpleTriggerFactoryBean">
<property name="jobDetail" ref="myJobDetail" />
<!-- 30 seconds -->
<property name="startDelay" value="10000" />
<!-- repeat every 50 seconds -->
<property name="repeatInterval" value="10000" />
</bean>
<!-- Job Details -->
<bean name="myJobDetail" class="org.springframework.scheduling.quartz.JobDetailBean">
<property name="jobClass" value="test.MyJob" />
<property name="jobDataAsMap">
<map>
.....
</map>
</property>
</bean>
<aop:aspectj-autoproxy />
....
I am not an expert in spring or aop. Can someone please explain what is wrong and how can we achieve it?
Without enabling AspectJ my scheduler gets triggered properly.
Thanks.

Do you have any aspects applied to the MyJob bean? If so, perhaps the aop mechanism creates a jdk dynamic proxy that cannot be recognised by the quartz api where you specify a specific job class of type MyJob. If this assumption is correct, then try:
<aop:aspectj-autoproxy proxy-target-class="true" />
That will instead use a CGLIB-based class proxy based on MyJob class.
I recommend reading http://docs.spring.io/spring/docs/4.1.x/spring-framework-reference/html/aop.html#aop-proxying in order to understand any implications that each proxying mechanism may have.
If that was the problem, it may be better to avoid completely having an aspect around the job class itself, and perhaps putting it around a bean that the job class will call to perform the desired functionality

Related

`mvc:message-converters` ignored in Spring 5

We used to have the following configuration to customize our message converters in Spring 4:
<mvc:annotation-driven conversion-service="conversionService" content-negotiation-manager="contentNegotiationManager">
<mvc:message-converters>
<bean class="org.springframework.http.converter.json.MappingJackson2HttpMessageConverter">
<property name="objectMapper" ref="myObjectMapper"/>
</bean>
</mvc:message-converters>
</mvc:annotation-driven>
But after migrating to Spring 5, the above configuration has lost any effect (the default message converters take place).
If I define the bean explicitly, it works:
<bean class="org.springframework.web.servlet.mvc.method.annotation.RequestMappingHandlerAdapter">
<property name="messageConverters">
<list>
<bean class="org.springframework.http.converter.json.MappingJackson2HttpMessageConverter">
<property name="objectMapper">
<bean class="org.springframework.http.converter.json.Jackson2ObjectMapperFactoryBean">
<property name="objectMapper" ref="myObjectMapper" />
</bean>
</property>
</bean>
<bean class="org.springframework.http.converter.ByteArrayHttannotation-drivenpMessageConverter"/>
<bean class="org.springframework.http.converter.xml.SourceHttpMessageConverter"/>
<bean class="org.springframework.http.converter.FormHttpMessageConverter"/>
<bean class="org.springframework.http.converter.StringHttpMessageConverter" />
</list>
</property>
</bean>
Any idea why the mvc:message-converters tag stopped working?
ps. While debugging the issue I also noticed several instances of RequestMappingHandlerAdapter being instantiated in all cases, even though I would expect them to be a singleton. The latter significantly obscures debugging, since I'm not sure which is the right one to attribute message converters to to investigate.
pps. Another report of this issue:
https://stackoverflow.com/a/49156186/657723
In this case mvc:message-converters was overridden by another, empty mvc:annotation-driven which happened to be unnoticeably included down the chain of child xml inclusions.
In any case, the major problem was the silent override, which should have better be an error.

Which is the default transaction manager the #Transactional uses?

We have configured multiple transaction managers:
<tx:annotation-driven transaction-manager="transactionManager1" />
<tx:annotation-driven transaction-manager="transactionManager2" />
<tx:annotation-driven transaction-manager="transactionManager3" />
<bean id="transactionManage1"
class="org.springframework.orm.hibernate4.HibernateTransactionManager">
<property name="sessionFactory" ref="sessionFactory1" />
</bean>
<bean id="transactionManager2"
class="org.springframework.orm.hibernate4.HibernateTransactionManager">
<property name="sessionFactory" ref="sessionFactory2" />
</bean>
<bean id="transactionManage3"
class="org.springframework.orm.hibernate4.HibernateTransactionManager">
<property name="sessionFactory" ref="sessionFactory3" />
</bean>
However,I don't see a name of transaction manager specified in #Transactional annotation in a service class. I am wondering which transaction manager will be used? The code is working and it looks like it used the first one - "transactionManager1" by default. Do I miss something?
I am wondering which transaction manager will be used?
The behaviour there is probably undefined, but it's most likely to use whichever <tx:annotation-driven> comes first in the context.
There's no reason you should need to do this. If you need fine-grained control over which tx manager to use, then you need to do it programmatically using something like TransactionTemplate. The #Transactional annotation is a convenience, but isn't as flexible as using the underlying API directly.

Spring, Websphere, JMS and Eclipselink Issue

I have an application that runs inside of Websphere, and I am having an issue with persisting JPA entities.
Previously, the application was setup with RESOURCE_LOCAL persistence units, with the Spring JpaTransactionManager, and transactions that were committed explicitly in code.
TransactionStatus transactionStatus = transactionManager.getTransaction( new DefaultTransactionDefinition() );
try {
entityManager.persist( someJpaEntity );
}
catch( Exception exception ) {
transactionManager.rollback( transactionStatus );
throw exception;
}
try {
transactionManager.commit( transactionStatus );
}
catch( TransactionException exception ) {
exception
}
I am working on an enhancement to the application that will allow calls through a Message Driven Pojo linked to a Websphere Queue. I was able to setup a configuration through spring that will allow my application to receive messages through a JMS queue. The spring config looks like:
<jee:jndi-lookup id="jmsConnectionFactory" jndi-name="QueueConnectionFactory"/>
<jee:jndi-lookup id="jmsQueue" jndi-name="DIQueue" />
<!-- A dynamic resolver -->
<bean id="jmsDestResolver" class="org.springframework.jms.support.destination.DynamicDestinationResolver"/>
<bean id="jmsQueueTemplate" class="org.springframework.jms.core.JmsTemplate">
<property name="connectionFactory">
<ref bean="jmsConnectionFactory"/>
</property>
<property name="destinationResolver">
<ref bean="jmsDestResolver"/>
</property>
</bean>
<bean id="messageListener" class="my.app.FileMessageListener" />
<bean id="exListener" class="my.app.JmsExceptionListener" />
<bean id="msgListenerContainer" class="org.springframework.jms.listener.DefaultMessageListenerContainer">
<property name="connectionFactory" ref="jmsConnectionFactory" />
<property name="destination" ref="jmsQueue" />
<property name="messageListener" ref="messageListener" />
<property name="transactionManager" ref="transactionManager" />
<property name="taskExecutor" ref="myTaskExecutor" />
<property name="exceptionListener" ref="exListener" />
</bean>
<bean id="myTaskExecutor" class="org.springframework.scheduling.commonj.WorkManagerTaskExecutor">
<property name="workManagerName" value="wm/default" />
</bean>
Not sure if there is an issue with my spring setup, but I do receive messages through my Active MQ broker, so that part I seem to be good with.
Now, the issue is, that when I get a message in through JMS, I would call the above code to insert the JPA entity. When the code would run, I would get the message "unable to commit a one phase resource in a two phase transaction", or something similar. What I came to understand is that the Spring JpaTransactionManager does not work with XA or JTA transactions.
So, I worked on moving to the Spring JtaTransactionManager.
I changed everything I Could think of over to use JTA, here is where I declare my transaction manager:
<bean id="transactionManager" class="org.springframework.transaction.jta.WebSphereUowTransactionManager"/>
Updated my persistence XML:
<persistence-unit name="AppUnit" transaction-type="JTA">
<provider>oracle.toplink.essentials.ejb.cmp3.EntityManagerFactoryProvider</provider>
<jta-data-source>java:APPDS</jta-data-source>
And still, nothing works. My code runs without exception, but nothing gets persisted to the database. The message gets pulled off of the JMS Queue, but no data.
Any suggestions?
I finally got this working, and figured I would post the answer.
There are actually 2 pieces to the puzzle.
First, in Websphere, you need to go to your app server -> TransactionService, and check / enable the "Accept Heuristic Hazard" checkbox. That definitely helped. I am running WAS 7.
The second thing, is you MUST set the property eclipselink.target-server on your persistence unit or your EntityManagerFactory.
That second item definitely did the trick. I tested with the property and without it. Without it, nothing persists. With it, everything works fine.
Here is my EntityManagerFactory, I am using a property placeholder to set the value of the eclipselink.target-server property:
<bean id="entityManagerFactory" class="org.springframework.orm.jpa.LocalContainerEntityManagerFactoryBean">
<property name="dataSource" ref="dataSource" />
<property name="persistenceUnitName" value="MyUnit" />
<property name="jpaVendorAdapter">
<bean class="org.springframework.orm.jpa.vendor.EclipseLinkJpaVendorAdapter">
<property name="databasePlatform">
<value>${app.databasePlatform}</value>
</property>
<property name="showSql">
<value>${app.showSql}</value>
</property>
</bean>
</property>
<!-- THIS DID THE TRICK -->
<property name="jpaPropertyMap">
<map>
<entry key="eclipselink.target-server" value="${app.targetServer}"/>
</map>
</property>
</bean>

Spring MVC from 2.5 to 3.0

We are migrating our web application from Spring 2.5 to Spring 3.0.5. Looks like all the Controller classes (the classes in org.springframework.web.servlet.mvc package: BaseCommandController, AbstractCommandController, SimpleFormController, etc) have been deprecated. We used those classes heavily in our application. These classes have been replaced by #Controller annotation instead.
I have a few questions regarding Spring 3.0.x Controller configuration...
We used the following XML to create a Controller bean in Spring 2.5. If <context:component-scan> is used instead to create the Controller beans, then how do I wire-in the dependencies? If possible I would like to avoid Auto-Wiring.
<bean id="photosetViewController" class="com.xyz.web.PhotosetViewController"
p:photosetManager-ref="photosetManager"
p:photoManager-ref="photoManager" />
We have created 3 Interceptors. We use SimpleUrlHandlerMapping to map these Interceptors to different request URLs and also to map URLs to Controllers. How do we achieve the same in Spring 3.0.x?
<bean class="org.springframework.web.servlet.handler.SimpleUrlHandlerMapping">
<property name="alwaysUseFullPath" value="true" />
<property name="interceptors">
<list>
<ref bean="httpRedirectInterceptor"/>
<ref bean="loginInterceptor"/>
</list>
</property>
<property name="mappings">
<value>
/abc.html=abcLookupController
/photoupload.html=photoUploadController
</value>
</property>
</bean>
<bean class="org.springframework.web.servlet.handler.SimpleUrlHandlerMapping">
<property name="alwaysUseFullPath" value="true" />
<property name="interceptors">
<list>
<ref bean="httpRedirectInterceptor"/>
<ref bean="loginInterceptor"/>
<ref bean="userMenuInterceptor" />
</list>
</property>
<property name="mappings">
<value>
/index.html=homepageController
/success.html=simpleController
</value>
</property>
</bean>
In case of SimpleFormControllers we used different methods like initBinder, referenceData, formBackingObject. We also used command objects and validation classes. How do we achieve the same in Spring 3.0.x?
<bean id="photosetAddController" class="com.xyz.web.PhotosetAddController"
p:photosetManager-ref="photosetManager"
p:photosetTypeManager-ref="photosetTypeManager"
p:stateManager-ref="stateManager" >
<property name="validator">
<bean class="com.xyz.form.PhotosetAddValidator" />
</property>
<property name="formView" value="photosetadd" />
<property name="successView" value="redirect:/photoset" />
</bean>
Any pointers are greatly appreciated.
As skaffman noted - your code will work fine without any modifications. I'll answer your questions in short:
You can use #Resource(name="beanName") for name-based injection. But autowiring is also a good option. Either #Autowired, or #javax.inject.Inject. Note that you can use #Qualifier("name") to differentiate between beans with the same interface/base class. You can even use the javax.inject qualifiers (read about all these in the docs)
interceptor mappings can stay the same
There is #InitBinder, which denotes the initBinder method. I can't tell about the other.

host/role dependent spring configuration

I have a cluster of servers running a spring application. Some of the spring components need to be configured differently depending on the role their server is playing (primary, secondary, etc). I don't want to maintain separate spring config files for each role, rather I want to dynamically detect this when the application is started. Its almost like I want conditional bean instantiation (which doesn't exist in spring).
Q: What is the best way to achieve this type of configuration?
Example: Only the primary node in a cluster should create durable subscription to a JMS broker (which requires a globally unique JMS clientID). I can detect if the current host has this role by looking up the running server's hostname in a database and start this container manually (if my node happens to be primary); however, I don't want every node in the cluster to create a durable subscription (by instantiating this bean).
<bean id="auditrecordListenerContainer"
class="org.springframework.jms.listener.DefaultMessageListenerContainer">
<property name="concurrentConsumers" value="1" />
<property name="clientID" value="${server-hostname}" />
<property name="durable" value="true" />
<!-- only started on the primary node: via application listener -->
<property name="autoStartup" value="false" />
</bean>
Note, however there is no ${server-hostname} property in the spring container (at least that I know of)
If your code already conditionally starts the appropriate services based on object properties, you can use utility methods in the following manner:
<!-- Factory methods to determine properties -->
<bean id="hostname" class="MyUtil" factory-method="determineHostName"/>
<bean id="isHost" class="MyUtil" factory-method="isHost"/>
<bean id="auditrecordListenerContainer"
class="org.springframework.jms.listener.DefaultMessageListenerContainer">
<property name="concurrentConsumers" value="1" />
<property name="durable" value="true" />
<!-- Reference properties here -->
<property name="hostname" ref="hostname" />
<property name="autoStartup" ref="isHost" />
</bean>
To use a property of a singleton bean instead, use a PropertyPathFactoryBean:
<bean id="config" class="MyConfig"/>
<util:property-path id="hostname" path="config.hostname"/>
<util:property-path id="isHost" path="config.host"/>
You can implement a conditional instantiation logic
as a FactoryBean

Resources