Java EJB Interview questions Part11

Question Can a Session Bean be defined without ejbCreate() method? (EJB)

Answer The ejbCreate() methods is part of the bean's lifecycle, so, the compiler will not return an error because there is no ejbCreate() method.

However, the J2EE spec is explicit:

the home interface of a Stateless Session Bean must have a single create() method with no arguments, while the session bean class must contain exactly one ejbCreate() method, also without arguments.

Stateful Session Beans can have arguments (more than one create method) stateful beans can contain multiple ejbCreate() as long as they match with the home interface definition

You need a reference to your EJBObject to startwith. For that Sun insists on putting a method for creating that reference (create method in the home interface). The EJBObject does matter here. Not the actual bean.



Question Is it possible to share an HttpSession between a JSP and EJB? What happens when I change a value in the HttpSession from inside an EJB? (EJB)

Answer You can pass the HttpSession as parameter to an EJB method, only if all objects in session are serializable.This has to be consider as "passed-by-value", that means that it's read-only in the EJB. If anything is altered from inside the EJB, it won't be reflected back to the HttpSession of the Servlet Container.The "pass-by-reference" can be used between EJBs Remote Interfaces, as they are remote references. While it IS possible to pass an HttpSession as a parameter to an EJB object, it is considered to be "bad practice (1)" in terms of object oriented design. This is because you are creating an unnecessary coupling between back-end objects (ejbs) and front-end objects (HttpSession). Create a higher-level of abstraction for your ejb's api. Rather than passing the whole, fat, HttpSession (which carries with it a bunch of http semantics), create a class that acts as a value object (or structure) that holds all the data you need to pass back and forth between front-end/back-end. Consider the case where your ejb needs to support a non-http-based client. This higher level of abstraction will be flexible enough to support it. (1) Core J2EE design patterns (2001)



Question Is there any way to read values from an entity bean without locking it for the rest of the transaction (e.g. read-only transactions)? We have a key-value map bean which deadlocks during some concurrent reads. Isolation levels seem to affect the database only, and we need to work within a transaction. (EJB)

Answer The only thing that comes to (my) mind is that you could write a 'group accessor' - a method that returns a single object containing all of your entity bean's attributes (or all interesting attributes). This method could then be placed in a 'Requires New' transaction. This way, the current transaction would be suspended for the duration of the call to the entity bean and the entity bean's fetch/operate/commit cycle will be in a separate transaction and any locks should be released immediately. Depending on the granularity of what you need to pull out of the map, the group accessor might be overkill.



Question What is the difference between a "Coarse Grained" Entity Bean and a "Fine Grained" Entity Bean? (EJB)

Answer A 'fine grained' entity bean is pretty much directly mapped to one relational table, in third normal form.

A 'coarse grained' entity bean is larger and more complex, either because its attributes include values or lists from other tables, or because it 'owns' one or more sets of dependent objects. Note that the coarse grained bean might be mapped to a single table or flat file, but that single table is going to be pretty ugly, with data copied from other tables, repeated field groups, columns that are dependent on non-key fields, etc.

Fine grained entities are generally considered a liability in large systems because they will tend to increase the load on several of the EJB server's subsystems (there will be more objects exported through the distribution layer, more objects participating in transactions, more skeletons in memory, more EJB Objects in memory, etc.) The other side of the coin is that the 1.1 spec doesn't mandate CMP Error! No index entries found.support for dependent objects (or even indicate how they should be supported), which makes it more difficult to do coarse grained objects with CMP. The EJB 2.0 specification improves this in a huge way.