hi - thank you for your message.
> RMI/IIOP explicity doesn't support distributed garbage collection (CORBA
> doesn't provide support for it). See section 1.4.12 of the spec at
> http://www.omg.org/cgi-bin/doc?formal/01-06-07.pdf
I've read the doc - I've never really have done the
RMI-IIOP stuf, but this was interesting - thank you
very much.
> In practice, even if the EJB is passivated and reactived, the reference
> to the remote object should still be valid unless the remote server has,
> in the meantime, decided to destory it using
> PortableRemoteObject.unexportObject(object)...
ok., so it's more like no plain vanila RMI callbacks
are recommended in this(EJB's) case... got it,got it...
> >(question here: on the client side JVM, if I've held some
> >normal local object (hard) reference to the Remote Object, I
> >think the Remote Object may still get rmi-``unreferenced'',
> >but will it still be garbage collected then? if not so, 'rmi
> >object not found' exception will happen next time the rmi
> >stub did the method invocation?? --I think I may have to do
> >some experiment myself on this...)
I did some experiments with this today anyway, to have
found out that there is some ``time lag'' between the
``dgc unreferenced'' takes place and the actual local
GC takes place (depending on appropriate JVM properties
settings). and if the RMI method invocation was done
after the ``dgc unreferenced'' event *but before* the
local GC reclaim takes place, the RMI subsystem still
dispatches the method call into appropriate Remote Object
(...maybe I have to verify this again tomorrow; but looks
interesting...)
regards,
kenji
===========================================================================
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".
