Great! Now I'll be able to use Jeremie for everthing. :)
- Keith
-----Original Message-----
From: Bruno Dumant <[EMAIL PROTECTED]>
To: Keith L. Musser <[EMAIL PROTECTED]>
Cc: jonas-users <[EMAIL PROTECTED]>; jonathan
<[EMAIL PROTECTED]>
Date: Thursday, November 09, 2000 6:15 AM
Subject: Re: GIOP Message error
Hi,
I have found a way to correct this bug. It seems that write actions to
socket output streams need to be synchronized, although this fact is not
documented (at least, I have not found the related documentation).
The correction is available on the CVS repository, and will be available
in
the next Jonathan release. The nex Jonas release should include it too.
Thanks for having pointed out this bug and helped us reproduce it !
Bruno
"Keith L. Musser" wrote:
> Hi,
>
> I'm having a problem similar to the one reported by someone else to
the
> Jonathan mailing list:
> http://www.objectweb.org/messages/Jonathan/0036.html
>
> No solution to the problem was posted on the list.
>
> Here's my problem. I run my application using Jeremie it works fine
for
> several minutes. (I'm automatically hitting the same page - invoking
> the same session and entity beans, repeatedly.)
>
> After a while, Jonas crashes with the stack trace below.
>
> org.objectweb.jonathan.apis.kernal.InternalException: GIOP Message
> error.
> at
>
org.objectweb.david.libs.protocols.giop.GIOPProtocol$ServerSession_Low.s
> end(GIOPProtocol.java:720)
> at
>
org.objectweb.jonathan.libs.protocols.tcpip.TcpIpProtocol$Session.run(Tc
> pIpProtocol.java:431)
> at
>
org.objectweb.jonathan.libs.resources.JScheduler$JJob.run(JScheduler.jav
> a:252)
>
> Jonas hangs at this point and must be restarted.
>
> Elsewhere, at a different time (i.e. different crash of the system,
but
> I suspect they are related), my client reported
> "java.io.StreamCorruptedException: Type code out of range, is 73. It
> was reading an object from an ObjectInputStream, ultimately called
from
>
org.objectweb.jeremie.libs.presentation.std.StdMarshallerFactory$StdUnMa
> rshaller.readReference(StdMarshallerFactory.java:1009), which occurs
> inside the stub for my EJB object.
>
> I suspect both problems are related to thread safety. For example,
> perhaps multiple threads are putting data into the same stream,
> corrupting the result. The crash then occurs when the other end tries
> to read the stream. (I have no proof of this; it's just a hunch based
> on my experience with protocols and multi-threaded programming.)
>
> I have no problems like this when I run using RMI instead of Jeremie.
>
> Does anyone have experience with this? Has this problem been fixed?
>
> Keith L. Musser
> Integrated Dynamics, Inc.
> 812-371-7777
> email: [EMAIL PROTECTED]
>
> ----
> 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".
--
*******************************************************************
Bruno Dumant
DTL/ASR
france telecom R&D
38-40 rue du g�n�ral Leclerc
92794 Issy Moulineaux Cedex 9
FRANCE
mailto:[EMAIL PROTECTED]
tel: +33 1 45 29 46 03
fax: +33 1 45 29 66 04
*******************************************************************
----
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".
