I'm trying to get a grip using MongoDB with the Spring Data MongoDB framework. I tried severeal approaches to connect to my local DB and insert + retrieve some collections and documents, using the official Spring Reference Documentation and some simple examples like this Hello-World-Demo.
Actually, for the beginning I'm going to use the MongoTemplate to keep it simple. But now I'run into the following problem.
When I use the Spring Configuration with annotations configure the setting needed to connect to my local DB, everything works fine.
Otherwise When I use XML for the configuration, I run into an java.lang.NullPointerException at com.mongodb.DBTCPConnector.getClusterDescription
Here are my configuration files and the example code for connecting to the DB:
Use Case 1 - Spring Configuration with annotations:
//package, imports etc. here
#Configuration
public class MongoConfiguration {
public #Bean MongoDbFactory mongoDbFactory() throws Exception {
return new SimpleMongoDbFactory(new MongoClient(), "Test1");
}
public #Bean MongoTemplate mongoTemplate() throws Exception {
return new MongoTemplate(mongoDbFactory());
}
}
Using the configuration class like this ...
AnnotationConfigApplicationContext ctx = new AnnotationConfigApplicationContext(MongoConfiguration.class);
MongoOperations mongoOperation = (MongoOperations) ctx.getBean("mongoTemplate");
ctx.close();
for (String s : mongoOperation.getCollectionNames()) {
System.out.println(s);
}
.. creates this output:
documents
leute
system.indexes
system.users
Use Case 2 - XML configuration (SpringConfig2.xml):
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:context="http://www.springframework.org/schema/context"
xmlns:mongo="http://www.springframework.org/schema/data/mongo"
xsi:schemaLocation="http://www.springframework.org/schema/context
http://www.springframework.org/schema/context/spring-context-3.0.xsd
http://www.springframework.org/schema/data/mongo
http://www.springframework.org/schema/data/mongo/spring-mongo-1.0.xsd
http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans-3.0.xsd">
<mongo:mongo host="127.0.0.1" port="27017" />
<mongo:db-factory dbname="Test1" />
<bean id="mongoTemplate" class="org.springframework.data.mongodb.core.MongoTemplate">
<constructor-arg name="mongoDbFactory" ref="mongoDbFactory" />
</bean>
</beans>
Using the configuration file like this ...
GenericXmlApplicationContext ctx = new GenericXmlApplicationContext("SpringConfig2.xml");
MongoOperations mongoOperation = (MongoOperations) ctx.getBean("mongoTemplate");
ctx.close();
for (String s : mongoOperation.getCollectionNames()) {
System.out.println(s);
}
.. results in this error
java.lang.NullPointerException
at com.mongodb.DBTCPConnector.getClusterDescription(DBTCPConnector.java:404)
at com.mongodb.DBTCPConnector.getMaxBsonObjectSize(DBTCPConnector.java:653)
at com.mongodb.Mongo.getMaxBsonObjectSize(Mongo.java:641)
at com.mongodb.DBCollectionImpl.find(DBCollectionImpl.java:81)
at com.mongodb.DBCollectionImpl.find(DBCollectionImpl.java:66)
at com.mongodb.DB.getCollectionNames(DB.java:510)
at org.springframework.data.mongodb.core.MongoTemplate$13.doInDB(MongoTemplate.java:1501)
at org.springframework.data.mongodb.core.MongoTemplate$13.doInDB(MongoTemplate.java:1499)
at org.springframework.data.mongodb.core.MongoTemplate.execute(MongoTemplate.java:394)
at org.springframework.data.mongodb.core.MongoTemplate.getCollectionNames(MongoTemplate.java:1499)
at test.main(Test.java:28)
When debugging DBTCPConnector.getClusterDescription, it seems that in the second case the private class variable cluster is for some reason not instantiated, leading to the described error.
What I'd like to know is: am I doing anything wrong within my XML-configuration or when using this config / context? Why does using XML-configuration end in an error, while using annotation configuration just works fine?
Basically (in the end) I just "copy+paste"'d the code examples from the official references for Spring / Spring Data MongoDB.
I'd appreciate any help / suggestions :)
Thanks to the hint from Tushar Mishra in the comments I was able to track down the origin of the error and why it occurs in one case but not in the other.
I'll try to explain in short, maybe it'll save someone some research time (or remember me if I perhaps run into this or a similar error again).
When closing either the AnnotationConfigApplicationContext-object (UC1) or the AnnotationConfigApplicationContext-object (UC2) with ctx.close(), from somewhere
org.springframework.beans.factory.DisposableBean#destroy() is invoked. As far as I understood, within there used singleton beans get destroyed by invoking the destroy()-methods
of each of those beans.
In short: closing the XXApplicationContext-object also destroys the MongoFactoryBean and with it closes the connection to the MongoDB.
So using c**tx.close()** at the described position leads to using a MongoOperations-object on a closed connection at the next line, resulting in the described NullPointerException.
And for why the sample code with Annotations (UC1) runs fine, whereas the sample code configured with XML (UC2) just breaks with an NullPointerException:
In UC1, the org.springframework.beans.factory.DisposableBean#destroy()-call is interupted by some DisposableBeanMethodInterceptor, which tries to redirect to a dispose() method.
But there is no method implemented in MongoFactoryBean, so the Mongo-connection stays alive and can be used even after ctx.close() was invoked. I think the connection will get killed later by the garbage collector, leading to the same error.
I'm not sure what to make of this, yet. If I should implement a dispose()-method myself or something alike. But at least I figured out, why this code sample behave different, although supposed to do the same thing.
Maybe it helps someone. Thanks again, Tushar, for giving a hint to the right direction.
Related
While reading some advanced book on developing the enterprise applications, I constantly see the following pattern that could be illustrated by the following example:.
public interface Oracle {
String defineMeaningOfTheLife();
}
public class BookwormOracle implements Oracle {
public String defineMeaningOfTheLife() {
return "Use life, dude!";
}
}
And the main function:
public static void main(String[] args) {
XmlBeanDefinitionReader rdr = new XmlBeanDefinitionReader(factory);
rdr.loadBeanDefinitions(new ClassPathResource(
"META-INF/spring/xml-bean-factory-config.xml"));
Oracle oracle = (Oracle) factory.getBean("oracle");
System.out.println(oracle.defineMeaningOfTheLife());
}
And the xml config:
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans.xsd">
<bean id="oracle" name="wiseworm" class="BookwormOracle" />
As far as I understood, it is not possible to instantiate the interface. But, using Spring framework it seems to me that it is possible to do so. Since, how does the Spring framework manage to do it? From pure Java perspective the code
Oracle oracle = new Oracle();
is rather wrong.
Spring also needs an implementation of the interface to instanciate the bean and make it available to your application. It is actually what is defined in your context file (xml-bean-factory-config.xml):
<bean id="oracle" name="wiseworm" class="BookwormOracle" />
In the spring context, you define which implementation of the interface Oracle you want to create. When your main method call:
Oracle oracle = (Oracle) factory.getBean("oracle");
it asks to Spring to get the bean with id "oracle" which is an implementation of your Oracleinterface.
You shouldfirst understand about DI(Dependency Injection) and IoC(Inversion of Control) . Please google it .
I would recommend you this article from Martin Fowler on Ioc.
http://martinfowler.com/bliki/InversionOfControl.htmlenter link description here
Thanks
This line <bean id="oracle" name="wiseworm" class="BookwormOracle" /> is equal to the below java code.
BookwormOracle oracle = new BookwormOracle();
It just happened to have the name of the variable as oracle in the spring configuration, rather spring actually initializing the concrete class BookwormOracle . Later you are just asking the spring container to get that bean with the name oracle with below line.
Oracle oracle = (Oracle) factory.getBean("oracle");
I have a maven mutli module project :
Project
-ProjectDAO
-Projectx
ProjectDAO uses spring + hibernate
From Projectx I am trying to use something like below:
public class TesMessage implements ITesMessage {
#Autowired
private IGlobal iGlobal;
...
iGlobal.getSomeMethod();
}
.. With the above code I get Null Pointer exception, Am i missing anything?
I have this in my appContext.xml
<context:component-scan base-package="com.test.nty.dal">
<context:include-filter type="annotation"
expression="org.springframework.stereotype.Repository" />
</context:component-scan>
Thanks
Believe the exception: Yes, you're missing something.
You don't show how you've annotated the IGlobal interface. The exception suggests that it's not under Spring's control. You have to let the Spring app context handle its creation as well as your TextMessage.
I would question this design. I don't see why every text message would need something global. Looks like a singleton bottleneck to me.
I am spawning a thread which will keep pulling the chunk of records from database and putting them into the queue. This thread will be started on the server load. I want this thread to be active all the time. If there are no records in the database I want it to wait and check again after some time. I was thinking of using spring task scheduler to schedule this but not sure if that is right because I only want my task to be started once. What will be the good way of implementing this in Spring ?
Also, i need to have a boundary check that if my thread goes down (because of any error or exception condition) it should be re-instantiated after some time.
I can do all this in java by using thread communication methods but just trying if there is something available in Spring or Java for such scenarios.
Any suggestions or pointer will help.
You can use the #Scheduled annotation to run jobs. First create a class with a method that is annotated with #Scheduled.
Class
public class GitHubJob {
#Scheduled(fixedDelay = 604800000)
public void perform() {
//do Something
}
}
Then register this class in your configuration files.
spring-context.xml
<?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:tx="http://www.springframework.org/schema/tx"
xmlns:task="http://www.springframework.org/schema/task"
xmlns:context="http://www.springframework.org/schema/context"
xsi:schemaLocation="http://www.springframework.org/schema/task http://www.springframework.org/schema/task/spring-task-3.1.xsd
http://www.springframework.org/schema/beans http://www.springframework.org/schema/beans/spring-beans.xsd
http://www.springframework.org/schema/context http://www.springframework.org/schema/context/spring-context-3.1.xsd
http://www.springframework.org/schema/tx http://www.springframework.org/schema/tx/spring-tx-3.0.xsd">
<tx:annotation-driven/>
<task:annotation-driven scheduler="myScheduler"/>
<task:scheduler id="myScheduler" pool-size="10"/>
<bean id="gitHubJob" class="org.tothought.spring.jobs.GitHubJob"/>
</beans>
For more about scheduling visit the Spring Docs.
#Scheduled(fixedDelay=3600000)
private void refreshValues() {
[...]
}
That runs a task once every hour. It needs to be void and accept no arguments. If you use Spring's Java configuration you'll also need to add the annotation #EnableScheduling to one of your #Configuration classes.
You can try using Quartz scheduler. http://quartz-scheduler.org/
That will allow you to specify the duration of time between task executions. You can set a flag (boolean) that says if the class has run before to avoid duplicating code that you only want to run the 'first' time.
Spring has out-ot-the-box support for scheduling tasks. Here are for the docs for the latest 3.2.x version of spring but check the docs for the version you are using. Looks like it uses Quartz under the hood for scheduling tasks.
I think your requirement is just regular senario which quartz or spring scheduling framework supports very well. but you want to create a spececial approach to impl it. my suggestion is to keep it simple and stupid. spring scheudling will take advantage your worker thread by pooling instead of runing it all the time.
As of statistics, it 's easy to check worker class's log. if you want see the statistics in web console, you need to record worker's log in database. I am sure you could make it easily in normal way.
Let's say I have following Spring config (version of Spring is 3.0.3):
<?xml version="1.0" encoding="UTF-8"?>
<beans xmlns="http://www.springframework.org/schema/beans"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xsi:schemaLocation="http://www.springframework.org/schema/beans
http://www.springframework.org/schema/beans/spring-beans.xsd">
<bean id="theFactoryBean" class="wax.MyFactoryBean"/>
<bean id="factoryBeanUser" class="wax.FactoryBeanUser">
<!-- what should be placed here?-->
</bean>
</beans>
I have instance of FactoryBean implementation and some other instance. I need Spring to inject to other instance FactoryBean, not the object it produces.
There are two possible ways to solve it.
First one, obvious and malfunctional:
<bean id="factoryBeanUser" class="wax.FactoryBeanUser">
<property name="myFactoryBean" ref="&theFactoryBean"/>
</bean>
With this config Spring throws following exception on start:
[skipped irrelevant part]
Caused by: org.xml.sax.SAXParseException: The reference to entity "theFactoryBean" must end with the ';' delimiter.
at com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.createSAXParseException(ErrorHandlerWrapper.java:195)
at com.sun.org.apache.xerces.internal.util.ErrorHandlerWrapper.fatalError(ErrorHandlerWrapper.java:174)
at com.sun.org.apache.xerces.internal.impl.XMLErrorReporter.reportError(XMLErrorReporter.java:388)
at com.sun.org.apache.xerces.internal.impl.XMLScanner.reportFatalError(XMLScanner.java:1414)
I found this solution Spring: Getting FactoryBean object instead of FactoryBean.getObject(), this question is maked as answered and four people voted for it. So I assume that this solution might work, but currently there is something wrong in my code.
Second one, working but awkward:
public class FactoryBeanUser implements ApplicationContextAware{
private MyFactoryBean myFactoryBean;
public void setApplicationContext(ApplicationContext applicationContext) throws BeansException {
myFactoryBean = (MyFactoryBean)applicationContext.getBean("&theFactoryBean");
}
}
My question is it possible to make first approach functional or I should stick with a second way?
It seems the XML parser interprets the ampersand (&) as a start of an XML-entity. You can try using ref="&theFactoryBean".
The spring docs is not clear whether this syntax is allowed in an xml file, or only with programatic lookup. But then the xml configuration is used by the app context, so I assume the & should work (although it seems it has not been the best choice for a special symbol)
Here's why I'd suggest another thing - if you really need the factory bean rather than its product, create another bean, that does not implement FactoryBean, define a method createObject() or something like that, and use it in all factories that need it.
A sidenote - better reference the xsd with the version included:
http://www.springframework.org/schema/beans/spring-beans-3.0.xsd
I have an application that I am currently writing that will use Spring and Hibernate. In my services layer I have injected a DAO that will do some very basic CRUD-ing actions. For grins, I have created a method annotated as follows:
#Transactional(readOnly = false, propogation=Propogation.REQUIRES_NEW)
public void doSomeWork(Dao dao, Entity e){
//do some searching
dao.persist(e);
dao.findAll(Entity.clz);
}
The dao persist method looks like this:
public void persist(Entity e){
session.saveOrUpdate(e); //This has already been built using a SessionFactory
}
The dao findAll method looks like this
public void findAll(Class clz) {
session.createCriteria(clz).list();
}
Now, everything seems to run, OK. After I insert (persist) my object, I can see it using the findAll method (along with the new Primary Key ID that it was assigned by the Data Store), however, when the "doSomeWork" method completes, my data does not actually get persisted to the underlying datastore (Oracle 10g).
If, however, I remove the #Transactional annotations and use Hibernate's session.getTransaction().begin() and session.getTransaction().commit() (or rollback), the code works as I would anticipate.
So, my underlying question would then be: Does Hibernate not actually use Spring's transaction management for actual transaction management?
In my bean-config file I am declaring a TransactionManager bean, a SessionFactory bean, and I am also including in the config file.
What could I possibly be missing, aside for a better working-knowledge of Spring and Hibernate?
Don't forget to add <tx:annotation-driven> for #Transactional support
sounds like spring doesnt actually inject the transaction handling code.
do you have something like this in your config file, to tell spring where to look for annotated classes?
<beans xmlns:context="http://www.springframework.org/schema/context" ...
xsi:schemaLocation="http://www.springframework.org/schema/context
http://www.springframework.org/schema/context/spring-context-3.0.xsd ..." >
...
<context:annotation-config/>
<context:component-scan base-package="mypackage.dao.impl"/>
<bean name="transactionManager" class="org.springframework.orm.hibernate3.HibernateTransactionManager">
<property name="sessionFactory" ref="sessionFactory"/>
</bean>
</beans>
what method do you use to obtain the session object from the session factory? Are you using openSession(), or getCurrentSession(), or perhaps something else? This matters because you need to session to be bound to the spring transaction (I doubt if openSession is good for your scenario)
I suggest that you use Spring's hibernateTemplate to invoke saveOrUpdate and persist, instead of using the session object. This way you are guaranteed that it will be bound to the transaction, and, as spring promises, you won't need to change the code if you ever change the transaction management strategy.
Ok, well, thanks to everyone who responded... it helped me to figure out what I am doing wrong...
In my overzealous "proof-of-concepting" it never really dawned on me what was going on until I realized that my "simple java class with a main method that will be doing all my work" isn't managed by spring, therefore no real transaction management. This will in no way behave as a product application would, being managed by an app-server with controller beans, services, etc.
Then it dawned on me... "services"... I'm going to have a services layer, that's where the transaction support will live. So, right as rain, I created a simple service bean (marked with #Transactional) and it works just as I would hope it would. So, I call methods on my service, which calls methods on my Dao and bam!!! it works.
Thanks again to everyone who helped me to come to this conclusion.