>
>> EJB spec describ how to perform bean pooling/swapping.  But it seems to me
>> that if there is a reference to EJBHome/EJBObject, then you will have
>> one corresponding object running.  So for 1000 concurrent users, there
>> will be 1000+ EJBHome/EJBObject object on server although most of them
>> idle, while we may only have 100 bean instances or so doing real work.
>> If a client exits or crashes, its corresponding EJBHome/EJBObject object
>> will be gc-ed.  Is this understanding correct?  How expensive are to keep
>> those EJBHome/EJBObject objects around (compared with bean instance?)
>
>If you follow the spec word by word, then the above is correct. For
>example, if you do a find on a EntityBean and you get 1000 or so back,
>then if the implementor followed the spec literally it would create 1000
>EJBObjects.
>

When I read the ejb 1.1 entity state machine diagram p.102, the ejbOjbect does
not get associated with the bean for ejbFind (as I understand it).
For ejbFind the enity bean stays in the pooled state, it only moves to the ready
state (and is associated with an ejbObject) for ejbCreate, or ejbActivate.


                       ______
                      /_____/\
                     /____ \\ \      Carol McDonald
                    /_____\ \\ /     Sun Developer Relations
                   /_____/ \/ / /    One Network Drive, ubur01-101        ((
                  /_____/ /   \//\   Burlington, MA 01803-0901          C|~~|
                  \_____\//\   / /   [EMAIL PROTECTED]         `--'
                   \_____/ / /\ /    (781) 442-2601 PHONE                Java
                    \_____/ \\ \     (781) 442-1678 FAX
                     \_____\ \\
                      \_____\/


===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff EJB-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to