are you using a single system as the client or multiple? I've been able to simulate upward of 100 threads in jmeter against tomcat for a static file using my Gateway laptop which is a 1.4ghz PentiumM with 1Gb RAM. beyond 100 threads, the throughput tends to go down. I don't remember the exact settings I used, but it was roughly
max 150 min 100 spare 50 since your webpage hits the database, I would suggest using JMeter to test your database first and figure out the peak queries/second. Even if you have 4 tomcat servers on the front end, if your database can only handle 400 queries/sec, it only going to kill the database. at that point, tomcat is just tell the user internal server error. I hope that helps peter On Wed, 15 Sep 2004 11:11:07 +0200, Jorge Sopena <[EMAIL PROTECTED]> wrote: > > My case is teh first one, there will be peaks of requests. > I mean it's quite likely that in any moment I could receive this > quantity of requests. > For example, I very likely scenario is to receive in 2-3 minutes around > 100000 requests, what is in average 500req/second > > I know I will need more hardware for the final environment. > But my current problem is that I can't get connections with Tomcat in a > quickly way unless I send them separately. > I can send 15 requests concurrently, then I get > java.net.ConnectionException. > I knew I wouldn´t reach my goal in my machine (Pentimu IV, 2.4GHz, > 512MB) but I find the results very poor. > > So I wondered if Tomcat can reach to this numbers in another environment? > > Thanks > > Jorge > > > > > Elmar Haneke wrote: > > >> Thanks for the calculations, but more exactly I want the server to > >> attend 1000 requests/second. I mean I don't want any of those > >> requests to be rejected. It's probably the response time were > >> several seconds, but with better hardware I could manage that. > > > > > > Is it your intention to handle the (perhaps rare) case where 1000 > > requests are coming within the same second or are you planning to > > manage 1000 Requests every second (60000 requests per minute)? > > > > In the second case it won't help if each request remains active > > several seconds - this would cause to have several thousand > > connections open simultaneously - an very hard job for any kernel. > > > > Elmar > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > > > -- > ============================================================== > Jorge Sopena Torres > SIDSA > (Semiconductores Investigación y Diseño, S.A.) > > Parque Tecnológico de Madrid > c/ Torres Quevedo, nº 1 > 28760 TRES CANTOS (Madrid) (SPAIN) > > Phone : +34 91 803 5052 > Fax: +34 91 803 9557 > > e-mail: [EMAIL PROTECTED] > URL: http://www.sidsa.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]