One other suggestion: You might try putting Apache in front of the
Tomcat to see if it still drops the connections. If the behavior stops,
it might be something in the httpd connector. If the behavior persists,
then I would look at the app or, more likely, the other "network" access
(is there firewalls? VPN? some weird network setup?).

Ben

On Fri, 2003-03-21 at 13:51, Wouter Bijlsma wrote:
> There's nothing in the logfiles (we are only using catalina.out) that would suggest 
> that Tomcat even knows about a connection problem. The server just goes on as if 
> nothing happened, but the messages it sends to the 'dead client' are never 
> delivered. This might indeed be a load/config issue, although some users experienced 
> the problem even when only 3 users were logged in. The confusing part is that we 
> just cannot reproduce the bug on a fast network (not even with 15 users logged in), 
> while neither the client nor the server use time-sensitive code for sending or 
> receiving messages.
> 
> We are using tomcat 4.0.6 on a Debian/testing linux machine, kernel version 2.4.20
> 
> Wouter Bijlsma
> 
> On 21 Mar 2003 13:16:19 -0600
> Ben Ricker <[EMAIL PROTECTED]> wrote:
> 
> > Sounds like a load issue coupled with a config issue. How about some
> > logs entries? DO you see anything in the logs you setup? Catalina.out?
> > Context log?
> > 
> > Also, some idea what OS and version would also be helpful. It is nearly
> > impossible to tell (at least for me).
> > 
> > Ben Ricker
> > 
> > On Fri, 2003-03-21 at 13:01, Wouter Bijlsma wrote:
> > > Is it possible that tomcat sometimes randomly kills a request, does
> > > not process a request or queues a request indefinitely, thereby
> > > blocking a client that tries to read back the response to its request?
> > > Could this be possible when a servlet communicates with the clients
> > > using only raw data and InputStreams/OutputStreams? And: is it a good
> > > idea to develop a game server for playing games like chess, chequers
> > > or draughts with tomcat, considering the fact that tomcats main
> > > purpose is to act as web application server using a strict
> > > request-response paradigm. Is it safe to assume that every request
> > > made by a client always yields a proper response, also when using raw
> > > data transfers? Also when there's a really high volume of requests to
> > > a single servlet?
> > > 
> > > Don't get me wrong: I do *not* think there's something wrong with
> > > tomcat! As a matter of fact I have really good experiences with it
> > > using it for web applications!!
> > > 
> > > Kind Regards,
> > > 
> > >   Wouter Bijlsma,
> > >   [EMAIL PROTECTED]
> > > 
> > > 
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > > For additional commands, e-mail: [EMAIL PROTECTED]
> > -- 
> > Ben Ricker <[EMAIL PROTECTED]>
> > Wellinx.com
> > 
> > 
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> > 
> > 
> > 
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
-- 
Ben Ricker <[EMAIL PROTECTED]>
Wellinx.com


---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to