Hi!
My company is using Tomcat (running with Apache) as a Servlet environment
for some heavy duty web stuff. We have been (up to this time) using Tomcat
3.1 with mod_jserv. In an effort to evaluate how to speed up the system as a
whole, we've experimented with a number of configurations including
- Turning off the reloadable flag (a great benefit)
- Moving from JRE 1.2.2 to JRE 1.3 (also notable benefit)
- Moving from Tomcat 3.1 to Tomcat 3.2, both running mod_jserv
- Moving from Tomcat 3.2 with mod_jserv to Tomcat 3.2 with mod_jk
Now, I've found in the last two cases that performance was either unaffected
(within the bounds of our analysis error, which I'd put at 5%) or was
slightly better. However, from reading the documentation, I get the
impression that Tomcat 3.2 was supposed to be faster and that mod_jk was
supposed to be another level of enhancement on top of that.
My questions are thus:
1) How large of an increase in performance should I see between Tomcat 3.1
with mod_jserv and Tomcat 3.2 with mod_jserv? 10%, 25%, 100%?
2) How large an increase in performance should I see between Tomcat 3.2 with
mod_jserv and with mod_jk? 10%, 25%, 100%?
3) If I'm not seeing the expected performance gains, what should I be
looking to tune? My assumption was that the default setup for workers and
such was sufficient... given that we didn't play with the defaults for these
in Tomcat 3.1. What might explain our apparent lack of extra speed between
the two versions? Or is the performance better, but not by a large margin?
Thank you for any advice or information.
Tomb
------------------------------------------
Thomas R. S. Barclay
Voice: (613) 722-3232 ext 349
e-mail: [EMAIL PROTECTED]
2001: To the New Millenium! The next thousand years
are MINE.
------------------------------------------
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
