Am wondering why the socket is being closed before the client can send
its close message to the broker; I wonder if the broker's
InactivityMonitor is thinking that the client has been inactive for
too long and discarding the connection?

James

On 4/7/06, Gerdes, Mike <[EMAIL PROTECTED]> wrote:
>
> I have made a few testings and it seems to be some kind of a timing problem 
> or so. I have enabled full ssl debugging and compared both output files with 
> diff. There are some different ports and some lines are in a different 
> order...nothing special here so I ignored that.
> Then there is a point in the case when no error occurs, where a thread sends 
> data. This is not the case when the error occurs. Ok this error happens on 
> two different machines with different configurations.
> So from this I would say, that the error occurs, when the socket is closed 
> before one thread can send its message. As this seams to be a timing problem, 
> it would explain, why the error doesn't happen in all cases.
>
>
> -----Ursprüngliche Nachricht-----
> Von: James Strachan [mailto:[EMAIL PROTECTED]
> Gesendet: Donnerstag, 6. April 2006 14:00
> An: [email protected]
> Betreff: Re: a strange ssl error
>
>
>
> Thanks for the heads up Mike. Its a bit surprising the close-thing; as
> really all that happens with a close is we asynchronously send a close
> command, then shut down the socket - but it appears from your stack
> trace that the write of the close command fails while flushing the
> buffer to the socket due to the socket already being closed which I
> don't quite understand. Bizarre :)
>
> If you can think of anything else to help us nail down this
> strangeness please do let us know.
>
> BTW I wonder if its anything to do with tcpNoDelayEnabled setting?
> (i.e. whether we send complete packets or wait for them to fill up
> etc).
>
> http://activemq.org/Configuring+Wire+Formats
>
> On 4/6/06, Gerdes, Mike <[EMAIL PROTECTED]> wrote:
> >
> > thanks for the fast answer. Ok there is no part of my network that closes 
> > the connection, but your hint looks close to the error. It doesn't happen 
> > with tcp, but SSL throws sockettimeout messages all the time and calls a 
> > close after the message is transfered. The timeouts don't cause the 
> > connection to be closed while it is idle, so I guess there might be 
> > something in the code that closes the connection or so.
> > The funny thing is that everything works fine as long as no close is 
> > called. That means I can send and recieve messages without any problems all 
> > the time, so the connection is not totally dropped. But when I want to 
> > close it then the error occurs.
> > It is not that critical, as I can just leave all connections open, but it 
> > might be a problem in an enviroment with many clients and when the 
> > application that gets the exception crashes.
> >
> > anyway thanks again james
> >
> > -----Ursprüngliche Nachricht-----
> > Von: James Strachan [mailto:[EMAIL PROTECTED]
> > Gesendet: Donnerstag, 6. April 2006 12:57
> > An: [email protected]
> > Betreff: Re: a strange ssl error
> >
> >
> >
> > On 4/6/06, Gerdes, Mike <[EMAIL PROTECTED]> wrote:
> > >
> > > hi,
> > >
> > > so in my experiments with ssl, which looked promising, I have encountered 
> > > a new error. To be honest I have no clue what causes this error or why it 
> > > is caused at all.
> > > First this error happens in compination of ssl and jms, second the error 
> > > only happens in about 50% of all cases, third and the strangest thing is, 
> > > that the error only happens when a connection.close(); is in class file 
> > > and I haven't noticed it when the connection is not closed. Also it looks 
> > > like the error is happening shortly before the connection.close(); 
> > > command is executed. At least the error is displayed right in the output 
> > > of a loop that runs before the connection.close();.
> > >
> > > I am totally confused by this error and and argh....
> >
> > From the stack trace it looks like you are trying to close a client
> > connection but that fails because the socket has already been closed
> > by someone else - though I've no bright ideas why that might be the
> > case I'm afraid. Could it be a firewall or some other part of your
> > network is simply just dropping the underlying socket? Or are there
> > any broker side warnings/errors that is causing it to drop the socket?
> > Or was the client just inactive for too long?
> >
> > --
> >
> > James
> > -------
> > http://radio.weblogs.com/0112098/
> >
> >
> >
> > This mail has originated outside your organization,
> > either from an external partner or the Global Internet.
> > Keep this in mind if you answer this message.
> >
> >
> > This e-mail is intended only for the above addressee. It may contain
> > privileged information. If you are not the addressee you must not copy,
> > distribute, disclose or use any of the information in it. If you have
> > received it in error please delete it and immediately notify the sender.
> > Security Notice: all e-mail, sent to or from this address, may be
> > accessed by someone other than the recipient, for system management and
> > security reasons. This access is controlled under Regulation of
> > Investigatory Powers Act 2000, Lawful Business Practises.
> >
>
>
> --
>
> James
> -------
> http://radio.weblogs.com/0112098/
>
>
>
> This mail has originated outside your organization,
> either from an external partner or the Global Internet.
> Keep this in mind if you answer this message.
>
> This mail has originated outside your organization, either from an external 
> partner or the Global Internet. Keep this in mind if you answer this message.
>


--

James
-------
http://radio.weblogs.com/0112098/

Reply via email to