Hello,
Philippe Durieux mentionned in a mail (Pb: with unexportObject) a problem with
the unexport() method: Its call led to the same trouble:
org.omg.CORBA.OBJECT_NOT_EXIST: minor code: 0 completed: No
But I do not know if anyone gave him an answer, and I am interested because I
have exactly the same problem.
I think that when removing a EJB from Jonas an unexport is probably performed
to unreference the object, and forbids any remote access attempt.
Another description of the bug:
- I am using jeremie 2.4 and have a similar trouble, but I am not using Jonas. I
am performing an unexport() on objects I want to forbid access, except through a
direct method call of another one exported .
A: remote object that has a getB() method
B: remote object accessed only through the getB() method.
- The first client that gets this B object succeeds accessing it (through the A
object), but if another one is trying to access at the same time it fails with
the org.omg.CORBA.OBJECT_NOT_EXIST exception.
- If my first client exists normally, another one can access the remote object.
But if the first client stops abnormally (CTRL-C for example), the remote object
is no more accessible.
- No simultaneous access succeeds.
Gérard;
Alexandre Lefebvre wrote:
> FYI, message posted on jonas-users
>
> ------------------------------------------------------------------------
>
> Subject: Bug in Jeremie
> Date: Mon, 19 Feb 2001 09:23:09 +0100
> From: "Blumreiter, Fa. Westernacher, IIPS, B" <[EMAIL PROTECTED]>
> To: "'[EMAIL PROTECTED]'" <[EMAIL PROTECTED]>
>
> Hi (Philippe),
>
> thanks for reacting so promptly to my bug report. Unfortunately, I'm not
> able yet to test jonas 2.2.7 because I am behind a firewall, so I have no
> access to cvs. I'm just waiting for the binary release.
>
> But there is another problem: You said about jonas 2.2.7:
>
> "
> Fix a minor bug when using jonas and Jeremie about
> session bean remove: Sometimes, you used to get the error message
> "ExportException: object already exported"
> (bug reported by Torsten Blumreiter)
> "
>
> I found this error only when using RMI. Is this fixed, too?
>
> With jeremie, though, I get a different error. I reported this one already
> to the mailing list but I think you weren't able to reproduce it. I have now
> modified the client of the session bean example shipped with jonas and I can
> reproduce it with the jonas example session bean.
> The modification is as follows:
>
> package sb;
>
> import javax.naming.InitialContext;
> import javax.naming.Context;
> import javax.transaction.UserTransaction;
> import javax.rmi.PortableRemoteObject;
>
> /**
> * Sample for Session Bean.
> * Usage:
> * java sb.ClientOp
> */
> // Here, the modification starts...
> // you just have to extend Thread and run it a couple of times to simulate
> several concurrent clients
> public class ClientOp extends Thread {
>
> public static void main(String args[]) {
> for (int i=1;i<10;i++)
> new ClientOp().start();
> }
>
> public void run () {
>
> Context initialContext = null;
> try {
> ...
> // ... and here it ends
>
> The rest ist the same. If I run the Client program at least twice, the
> second time (or maybe you have to try it a couple more times) I get the
> following output:
>
> C:\Programme\JONAS-2-2-6\examples\src\sb>java sb.ClientOp
> Create a bean
> Create a bean
> Create a bean
> Create a bean
> Create a bean
> Create a bean
> Create a bean
> Start a first transaction
> First request on the new bean
> Start a first transaction
> Start a first transaction
> Start a first transaction
> Start a first transaction
> Create a bean
> Start a first transaction
> Start a first transaction
> First request on the new bean
> First request on the new bean
> First request on the new bean
> First request on the new bean
> First request on the new bean
> Start a first transaction
> First request on the new bean
> First request on the new bean
> Second request on the bean
> Second request on the bean
> Second request on the bean
> Second request on the bean
> Commit the transaction
> Commit the transaction
> Second request on the bean
> Second request on the bean
> Second request on the bean
> Commit the transaction
> Commit the transaction
> Start a second transaction
> Commit the transaction
> exception during 1st Tx: java.rmi.MarshalException: error during
> marshalling/unm
> arshalling by stub; nested exception is:
> org.omg.CORBA.OBJECT_NOT_EXIST: minor code: 0 completed: No
>
> The first time after starting jonas, the error does not occur.
> By the way, the error does not occur if you uncomment the remove() call. I
> would be interested if you can reproduce it both on windows nt and unix. If
> so, are there any explanations?
>
> Thanks and regards
>
> Torsten
>
> ----
> To unsubscribe, send email to [EMAIL PROTECTED] and
> include in the body of the message "unsubscribe jonas-users".
> For general help, send email to [EMAIL PROTECTED] and
> include in the body of the message "help".
To unsubscribe, send email to [EMAIL PROTECTED] and
include in the body of the message "unsubscribe jonathan".
For general help, send email to [EMAIL PROTECTED] and
include in the body of the message "help".
