Hi,
I try to understand how Tomcat 3.1 handles requests. I made the
following test:
I send a large amount of requests (JavaScript reloads, 5 per second over
a period of several minutes) from one user to Tomcat.
Result WITHOUT ThreadPool (SimpleTcpConnector):
the number of java processes (threads) increases. If I don't stop the
server crashes. If I stop after a while Tomcat needs some time to
regenerate respectively to work off the threads. Then I can continue.
Result WITH ThreadPool (PoolTcpConnector):
the number of java processes doen't increase to nirvana. If the limit is
adapted (on my 166 MHz ;-) max_threads=15 is ok) Tomcat does not crash.
But: when I stop the requests it needs nevertheless some time to
regenerate. Why? Does it kill the threads then and my PC is too slow? I
thought, all threads are killed immediately if there are more than
max_threads. After a while I can continue.
Can someone explain me in easy words these behaviours? Thank you.
What about Tomcat 3.2 (announced today) and thread pooling? Is it
similar?
Kai M�ller
--
Mediadom audiovisuelle Medien GmbH
Merheimer Str. 151
D-50733 Koeln
