That sounds as if a buffer isn't flushed under all circumstances or the buffer is cleared by the second thread. As this problem persists across different vm implementations it might be a linux problem. (The 2.4 kernel between 2.4.8 and 2.4.18 is not very famous for stability) I currently have difficulties to imagine an error in any java code outside the core io libraries that can produce such behaviour.
Can you switch bak to the 16K Buffer and use an explicit out.flush() at the end of the page and see what happens ? > -----Urspr�ngliche Nachricht----- > Von: Matti H�r� [mailto:[EMAIL PROTECTED]] > Gesendet: Donnerstag, 11. April 2002 16:00 > An: [EMAIL PROTECTED] > Betreff: Re: Concurrency problem with Tomcat under Linux > We solved the problem by removing the buffer="16000" > attribute from the page directive and voila-everything > seems to work fine! > > The error happens if the following conditions are met: > > > > 1) The server operating system is Linux (my kernel version > is: Linux > > pollux 2.4.9-13 #1 Tue Oct 30 20:11:04 EST 2001 i686 unknown) > > The error does not appear under windows Tomcat versions. > > 2) The page is large enough, so that the previous request is still > > executing while a new one is coming in. (need two separate > browsers in > > two separate machines for testing the concurrency problem) > > > > The error occurs with Sun's JDK 1.4, BlackDown's JDK, and > IBM's JDK, > > so the bug is propably not in the JDK implementation. > > > > browsers accessing simultaneously the page. -- To unsubscribe: <mailto:[EMAIL PROTECTED]> For additional commands: <mailto:[EMAIL PROTECTED]> Troubles with the list: <mailto:[EMAIL PROTECTED]>
