Hi,

Thanks for the email.  The more I thought about it
last night the more I would agree.  Doing a true
timout would be a large headache for both the
container and the J2EE developer.

I do wish the docs were clearer on this as part of the
project I got asked to look at is relying on the
container interrupting the transaction .  Thankfully I
did not write it ;-).

Thanks again,
David

--- Eric Williams <[EMAIL PROTECTED]> wrote:
> If my memory of Weblogic is correct, a transaction
> timeout will not
> interrupt a thread that goes beyond the timeout. The
> only thing the timeout
> will do is to set a flag on the transaction to
> indicate that it has timed
> out. So, the next time you attempt a transactional
> operation (e.g., a JDBC
> call or a UserTransaction.commit()), you'll be made
> aware of the timeout.
> Interrupt semantics are extremely hard to get right
> in Java. I tried to do
> this in a server I wrote once. When I interrupted a
> thread that was blocked
> on a socket I/O inside an Oracle JDBC [type 4]
> connection, it managed to
> hose the Connection every time. My suspicion is that
> this is the reason
> Weblogic does not provide interrupt semantics for
> transaction timeouts.
>
> To get "interrupt" semantics, you have to code that
> into your application.
> Some JDBC drivers / databases allow a maximum
> timeout to be specified
> (e.g., wait only N milliseconds for a row lock). In
> your test case, you
> would have to timeout the socket read in code. I'm
> not certain, but there
> may also be a way to get a timeout from the iPlanet
> IIOP layer (usually,
> these things are available in ORBs, but are
> configured in a proprietary
> way).
>
> -eric
> Eric Williams
>
>
>
>
>
>                         David Jones            To:
>  [EMAIL PROTECTED]
>                         <davidbrummy@YAHO      cc:
>                         O.COM>
> Subject:    Re: EJB Timeout
>                         Sent by: A
>                         mailing list for
>                         Enterprise
>                         JavaBeans
>                         development
>                         <EJB-INTEREST@JAV
>                         A.SUN.COM>
>
>
>                         05/08/2002 04:36
>                         PM
>                         Please respond to
>                         David Jones
>
>
>
>
>
> Hi,
>
> We have been struggling at lot with this issue at
> the
> moment.  It seems that Weblogic ignores its Tx
> timeout
> to a large extent.
>
> We ran some testing where an EJB opened a socket to
> a
> test program and tried to read a byte.  The Test
> programs thread sleeps for 10 seconds and the
> returns
> a byte thus unblocking the client.  We set the
> timeout
> for 5 seconds in both the console and deployement
> descriptor.
>
> We found the transaction never times out until we
> either call another EJB or the transaction attempts
> to
> commit. In both cases when the container does
> finally
> attempt to do a commit the container rolls back and
> the client gets a TransactionRollBackException.
>
> I would also be interested as well if anyone has a
> solution to this.
> David
>
> --- Scott Miao <[EMAIL PROTECTED]> wrote:
> > Hi! Everyone,
> >         I am having a problem that I am not sure
> how
> > to solve this.
> > Hopefully I can get some help from expert out
> there.
> > The situation I am
> > running into is that my application that runs
> inside
> > the Weblogic server
> > need to make a method call to a EJB which runs on
> > iPlanet on a different
> > Unix box. My Weblogic Server will stop running
> > (freeze) when the iPlanet app
> > server had a core dump. I did a thread dumping on
> > the weblogic side. All
> > "ExecuteThread" are had same message as I show
> > below.
> >
> > "ExecuteThread: '12' for queue: 'default'" daemon
> > prio=5 tid=0x76e790
> > nid=0x173 waiting on monitor
> [0xd65f000..0xd65fdc0]
> >         at java.lang.Object.wait(Native Method)
> >         at java.lang.Object.wait(Object.java:420)
> >         at
> >
>
com.sun.corba.se.internal.orbutil.Condition.wait(Condition.java:28)
> >         at
> >
>
com.sun.corba.se.internal.iiop.IIOPConnection.send(IIOPConnection.java:1002)
>
> >         at
> >
>
com.sun.corba.se.internal.iiop.IIOPOutputStream.invoke(IIOPOutputStream.java
>
> > :76)
> >         at
> >
>
com.sun.corba.se.internal.iiop.ClientRequestImpl.invoke(ClientRequestImpl.ja
>
> > va:87)
> >         at
> >
>
com.sun.corba.se.internal.corba.ClientDelegate.invoke(ClientDelegate.java:15
>
> > 8)
> >         at
> >
>
com.sun.corba.se.internal.corba.ClientDelegate.invoke(ClientDelegate.java:19
>
> > 8)
> >         at
> >
>
org.omg.CORBA.portable.ObjectImpl._invoke(ObjectImpl.java:459)
> > at......
> >
> > Any idea on how to broken the pipe between
> Weblogic
> > and iPlanet when iplanet
> > had core dump.
> >
> >
>
===========================================================================
> > 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".
> >
>
>
> =====
> David J. Jones, <[EMAIL PROTECTED]>,
> Virgin Mobile USA,
> 8th Floor,
> 22 Fourth Street,
> San Francisco,
> CA, 94103,               Work: 415 932 5470.
> USA.                     Fax:  415 358 4999.
>
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Health - your guide to health and wellness
> http://health.yahoo.com
>
>
===========================================================================
> 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".
>
>
===========================================================================
> 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".
>
=== message truncated ===


=====
David J. Jones, <[EMAIL PROTECTED]>,
Virgin Mobile USA,
8th Floor,
22 Fourth Street,
San Francisco,
CA, 94103,               Work: 415 932 5470.
USA.                     Fax:  415 358 4999.

__________________________________________________
Do You Yahoo!?
Yahoo! Shopping - Mother's Day is May 12th!
http://shopping.yahoo.com

===========================================================================
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