based on this example :
#Service
public class Purchase {
#PersistenceContext
private EntityManager em;
#Autowired
private PurchaseDAO dao;
private String normalField;
.... // methods, operations, etc
}
Please help correct me if im mistaken :
The service class Purchase and the PurchaseDAO are singletons that are managed by spring
The service class's field normalField is not threadsafe, because singleton is a single object shared by many
Let's assume the #Repository-annotated-PurchaseDAO doesnt have any field, which means it's threadsafe, will be injected automatically by spring
The EntityManager instance is also a threadsafe property because #PersistenceContext will make sure that the entityManager of the current transaction will be used.
Thank you !
They are singletons by default (when they are Spring managed), unless you have configured them otherwise (from either the xml-configuration or with annotations you can set it with #Scope).
Yes and no. Yes, it's not safe in the sense that multiple threads can access and modify it simultaneously, and no because it depends on the data type and String is immutable (and at least said to be thread safe). It could possibly blow up if two different threads tried to store a new string in the member variable at the exact same moment.
Yes and no again. If the DAO has no internal state, yes, I'd say it's thread safe, but the objects it is handling might not be (although if you're using JPA-entities, they should be).
At least Hibernate's documentation says that EntityManagers are not threadsafe, but when using an EntityManager injected by Spring, it shouldn't be an issue.
SpringSource forums:
I have been looking for answers to the same question on the forum... The general concensus seems to be that while EntityManager is not thread safe per JPA specs, the EntityManager injected by Spring through its EntityManager proxy may be.
Hibernate:
An EntityManager is an inexpensive, non-threadsafe object that should be used once, for a single business process, a single unit of work, and then discarded. An EntityManager will not obtain a JDBC Connection (or a Datasource) unless it is needed, so you may safely open and close an EntityManager even if you are not sure that data access will be needed to serve a particular request.
Related
I'm working on a bunch of legacy code written by people before me and I'm confused about a particular kind of setup and wonder if this has ever worked to begin with.
There is a managed bean in spring that has a transactional method.
#Transactional(propagation = Propagation.REQUIRES_NEW, rollbackFor = Throwable.class)
public boolean updateDraftAndLivePublicationUsingFastDocumentsOfMySite(List<FastDocumentLite> fastDocumentLites, Long mySiteId) throws Exception { ... }
Now inside that method I find new instantiations calling update methods fe:
boolean firstFeed = new MySiteIdUpdate(publishing, siteDao, siteDomainService).update(siteId, fastDocumentLites.get(0).getMySiteId());
From my understanding on IOC this new class isn't managed by spring , it's just a variable in the bean. Now going further inside the update method you see another service gets called.
#Transactional(propagation=Propagation.REQUIRED, rollbackFor=Throwable.class)
public void activateSubdomainForSite(Long siteId, boolean activationOfSite)
So if there is a transaction open it should be propagated into this service. But here is what I don't get if that MySiteIdUpdate object isn't managed by spring does the first transaction move forward to the activateSubdomainForSite method ?? Or is another transaction being opened here. I looked in the logs and I believe it to be the latter but I rather ask the experts for a second oppinion before I proclame this legacy code to be complete rubbish to the project lead. I'm suffering with a StaleStateException somewhere further down the road and I'm hoping this has anything to do with it.
I think the code is correct, and the second #Transactional should reuse the existing transaction.
Because:
1) Spring Transaction handling is done either by Proxies or by AspectJ advices. If it is done by Proxies then it is required that MySiteIdUpdate invoke an instance that is injected (this is what you did). If you use AspectJ, then it should work anyway.
2) The association Transactions to the code that use is done by the Thread, this mean, as long as you "are" in the thread which started the transaction you can use it. (you do not start an new thread, so it should work)
An other way to explain: It is perfect legal when you have some method in your call hierarchy that does not belong to an spring bean. This should not make the transaction handling fail.
In the new project that I have joined, they keep using the terms Hibernate and JPA interchangeably. So, I tried to dive down into the code and try to understand how this whole thing works (I am new to Spring, JPA and Hibernate world). I'll try to put the code here for better understanding:
1) There is a #Configuration class where they have the following:
#Resource
private HibernateJpaVendorAdapter hibernateOracleJpaVendorAdapter;
LocalContainerEntityManagerFactoryBean entityManager =
new LocalContainerEntityManagerFactoryBean();
entityManager.setJpaVendorAdapter(hibernateOracleJpaVendorAdapter);
entityManager.setPersistenceUnitName("abc");
.
.
So, in this configuration class, we are returning an EntityManagerFactory.
2) Then there is a persistor class marked #Persistor, where a method of repository is invoked (for example, for a save operation):
blahblahRepository.save(blahblahEntity, abcdef);
3) Finally there is a repository class which is annotated #Repository. Then again, they have this piece of code:
#PersistenceContext(unitName = "same as the name in persistence.xml")
protected EntityManager entityManager;
The "save" method wraps around the persist method of JPA:
getEntityManager().persist(entityObject);
My questions are as follows:
1) There is no word about Hibernate other than in the hibernateJpaVendorAdapter. I searched the entire workspace and it showed just 3 occurences of the word hibernate, all in the configuration file.
2) From whatever knowledge I have, one should use either an EntityManagerFactory or an EntityManager but we are doing both?
Hibernate is one of the implementations of the JPA spec. Since your project chose Hibernate as its JPA implementation, it uses the JPA API, which delegates the Hibernate. Just like when you use the JDBC API, which delegates to a specific Oracle or PostgreSQL driver.
EntityManagerFactory, as its name indicates, is a factory for EntityManager. I don't see why you wouldn't use both. EntityManager is the main interface of the JPA API, used to execute all database operations (find, persist, merge, query, etc.). EntityManagerFactory must be configured before asking it to create an EntityManager.
I'm using JAX-RS to provide an HTTP-based interface to manage a data model. The data model is stored in a database and interacted with via JPA.
This allows me to modify the interface to the data model to suit REST clients and mostly seems to work quite well. However, I'm not sure how to handle the scenario where a method provided by a JAX-RS resource requires a transaction, which affects the JPA get, update, commit-on-tx-end pattern, because there is only a transaction wrapping the get operation, so the update is never committed. I can see the same problem occurring if a single REST operation requires multiple JPA operations.
As I'm using Spring's transaction support, the obvious thing to do is to apply #Transactional to these methods in the JAX-RS resources. However, in order for this to work, Spring needs to manage the lifecycle of the JAX-RS resources, and the useage examples I'm aware of have resources being created via `new' when needed, which makes me a little nervous anyway.
I can think of the following solutions:
update my JPA methods to provide a transaction-managed version of everything I want to do from my REST interface atomically. Should work, keeps transactions out of the JAX-RS layer, but prevents the get, update, commit-on-tx-end pattern and means I need to create a very granular JPA interface.
Inject Resource objects; but they are typically stateful holding at least the ID of the object being interacted with
Ditch the hierarchy of resources and inject big, stateless super resources at the root that manage the entire hierarchy from that root; not cohesive, big services
Have a hierarchy of injected, stateless, transaction-supporting helper objects that 'shadow' the actual resources; the resources are instantiated and hold ths state but delegate method invocations to the helper objects
Anyone got any suggestions? It's quite possible I've missed some key point somewhere.
Update - to work around the lack of a transaction around the get, update, commit-on-tx-close flow, I can expose the EntityManager merge(object) method and call it manually. Not neat and doesn't solve the larger problem though.
Update 2 #skaffman
Code example:
In JPA service layer, injected, annotations work
public class MyEntityJPAService {
...
#Transactional(readOnly=true) // do in transaction
public MyEntity getMyEntity(final String id) {
return em.find(MyEntity.class, id);
}
In JAX-RS resource, created by new, no transactions
public class MyEntityResource {
...
private MyEntityJPAService jpa;
...
#Transactional // not injected so not effective
public void updateMyEntity(final String id, final MyEntityRepresentation rep) {
MyEntity entity = jpa.getMyEntity(id);
MyEntity.setSomeField(rep.getSomeField());
// no transaction commit, change not saved...
}
I have a few suggestions
Introduce a layer between your JPA and JAX-RS layers. This layer would consist of Spring-managed #Transactional beans, and would compose the various business-level operations from their component JPA calls. This is somewhat similar to your (1), but keeps the JPA layer simple.
Replace JAX-RS with Spring-MVC, which provides the same (or similar) functionality, including #PathVariable, #ResponseBody, etc.
Programmatically wrap your JAX-RS objects in transactional proxies using TransactionProxyFactorybean. This would detct your #Transactional annotations and generate a proxy that honours them.
Use #Configurable and AspectJ LTW to allow Spring to honour #Transactional even if you create the object using `new. See 8.8.1 Using AspectJ to dependency inject domain objects with Spring
I have a problem with my entity manager in my application. I have two DAO clasess like this:
#Repository
public abstract class DaoA
{
protected ClassA persistentClass;
#PersistenceContext(name="my.persistence", type=PersistenceContextType.EXTENDED)
protected EntityManager entityManager;
-------------- some typical action for DAO --------------
}
Second DAO is for ClassB and looks similar to DaoA. The rest of things are done for me by the Spring framework.
When I'm debugging the application I recognize that both DAO objects have different instances of EntityManager. In the result my two different DAOs are connected with different PersistenceContext.
Question is if this is correct behaviour or not? I would like to have the same PersistenceContext for all my DAO classes. Please give me a hint if this is possible and if I understood the JPA correctly?
Regards Hsd
It's a correct behaviour of EXTENDED persistence context, therefore you don't need it to be EXTENDED in this case.
In most cases you need a TRANSACTIONAL persistence context, that is the default mode when type is omitted:
#PersistenceContext(name="my.persistence")
protected EntityManager entityManager;
In this mode persistence context is associated with the transaction, so that all DAOs will share the same persistence context when working inside the same transaction. I guess it's what you need.
I'm developing my first app with JPA/Hibernate and Spring. My first attempt at a DAO class looks like this:
#Repository(value = "userDao")
public class UserDaoJpa implements UserDao {
#PersistenceContext
private EntityManager em;
public User getUser(Long id) {
return em.find(User.class, id);
}
public List getUsers() {
Query query = em.createQuery("select e from User e");
return query.getResultList();
}
}
I also found some examples using JpaDaoSupport and JpaTemplate. Which design do you prefer? Is there anything wrong with my example?
I'd say your approach looks totally sound. Personally I don't use JpaDaoSupport or JpaTemplate because you can do everything you need with the EntityManager and Criteria Queries.
Quote from the JavaDoc of JpaTemplate:
JpaTemplate mainly exists as a sibling of JdoTemplate and HibernateTemplate, offering the same style for people used to it. For newly started projects, consider adopting the standard JPA style of coding data access objects instead, based on a "shared EntityManager" reference injected via a Spring bean definition or the JPA PersistenceContext annotation.
I prefer the template-less approach (i.e. your current approach) because
it's less invasive, you don't tie DAOs to Spring
templates don't offer much value with APIs that use unchecked exceptions
And this is the Spring recommendation, as summarized in the blog post "So should you still use Spring's HibernateTemplate and/or JpaTemplate??" and the official javadoc:
The real question is: which approach to choose??
(...)
So in short (as the JavaDoc for
HibernateTemplate and
JpaTemplate already mention)
I'd recommend you to start using the
Session and/or EntityManager API
directly if you're starting to use
Hibernate or JPA respectively on a new
project–remember: Spring tries to be
non-invasive, this is another great
example!
I, personally, prefer your approach - inject EntityManager and use it directly. But JpaTemplate is also a good option. I don't like it, because adds yet another, unnecessary layer of abstraction.
I don't know if there's a "standard" approach.
If you're using JPA, you have your choice of implementations: Hibernate, TopLink, etc.
If you deploy to Google App Engine, you'll use JPA talking to BigTable.
So if your objectives are to maximize portability, stick with the JPA standard, and not tie yourself to a particular implementation like Hibernate, make sure that your DAOs only use JPA constructs.