That seems a bit odd. have you reproduced this on more than one system? and more than one server?
I typically will run benchmarks on atleast two different systems using a third machine as the client to make sure the numbers are an accurate reflection of the performance. Try doing the same benchmark with just apache on the same exact system to establish a baseline. Once you have that baseline, you can then compare the relative performance. There are many other factors that can cause the performance numbers you see. peter Han Wang wrote: > > I'm just accessing a static html file. > > Han > > -----Original Message----- > From: peter lin [mailto:[EMAIL PROTECTED]] > Sent: Thursday, May 16, 2002 4:15 PM > To: Tomcat Users List > Subject: Re: Concurrent Accesses > > I've managed to get Tomcat to handle 100+ concurrent requests getting > dynamic pages that look at request parameters to determine which > header/footer to display. Are your tests getting static HTML files? or > are they getting dynamic pages which query a database? > > peter lin > > Han Wang wrote: > > > > Tomcat Users, > > > > The question is this: > > How many concurrent accesses can tomcat 4.0.3 support? > > > > Now, I know this may be a function of processing power, but my own tests > are > > showing it much lower than expected. > > > > Basically did the following: > > Created a simple java app that spawns off n threads, each thread making an > > access to a file hosted by tomcat. Code snipped shown below: > > > > At n > 17, I start getting > > > > java.net.ConnectException: Connection refused: connect > > > > The number of connetions refused is roughly proportional to the number of > > thread over 17. > > > > Is there some configuration I have to change? > > I already modified server.xml, > > > > connector acceptCount="100", minProcessors="50", maxProcessors="100" from > > the default values, with NO effect. > > > > I also tried using Apache HTTPD and ajp13 (using the same connector > settings > > as above), with no effect. > > > > I can't believe this low number is truely the # concurrent accesses that > > Tomcat can handle, this must be an artifical limit somewhere. > > > > Please advise (and thanks in advance) > > Han Wang > > [EMAIL PROTECTED] > > > > SAMPLE CODE: > > > > public class TestClient extends Thread { > > public void run() { > > URL url = new URL("http://localhost:8080/axis/";); > > InputStream in = url.openStream(); > > URLConnection conn = url.openConnection(); > > int len = conn.getContentLength(); > > } > > > > public static void main(String[] args) { > > int n = Integer.parseInt(args[0]); > > TestClient [] threads = new TestClient[n]; > > > > for (int i = 0; i<n ; i++) { > > threads[i] = new TestClient(); > > } > > > > for (int i = 0; i<n ; i++) { > > threads[i].start(); > > } > > > > for ( int i=0; i<n; i++ ) { > > threads[i].join(); > > > > } > > } > > } > > > > -- > > To unsubscribe, e-mail: > <mailto:[EMAIL PROTECTED]> > > For additional commands, e-mail: > <mailto:[EMAIL PROTECTED]> > > -- > To unsubscribe, e-mail: > <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: > <mailto:[EMAIL PROTECTED]> > > -- > To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> > For additional commands, e-mail: <mailto:[EMAIL PROTECTED]> -- To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]> For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
