Ok…. I'm running Tomcat 4.1.18 as a standalone server on a Linux box.
I've configured 6 virtual hosts in the Tomcat-Standalone server having a
non-SSL Coyote HTTP/1.1 Connector (probably obvious) and a Coyote/JK2
AJP 1.3 Connector (I don't think this is used, but I'm not
sure….unfortunately reveals my lack of experise!).
I'm confused with the number of java processes spawned when I start
Tomcat. Here is what I see:
# of initial
minProcessors | maxProcessors | java processes | comments
1 | 1 | 25 | can't
get a page
1 | 2 | 27 | pages
served to 1 session
2 | 4 | 31
4 | 4 | 31 | slow
response, 1 session at a time
4 | 7 | 31 | 3
sessions respond, 4 add'l java procs created
5 | 7 | 31 | same
as above
4 | 20 | 31 | 5
sessions respond (all tested), 9 add'l java procs
So….. First question: why are there so many java processes when I
specify only a few minProcessors (i.e. 1 minProc, 1 maxProc, 6 virtual
hosts - I would expect 6 java sessions)? Second question: why isn't the
number of initial java processes a linear relationship relative to the
number of minProcessors specified?
I'm concerned about this because in the end, only about half the
processes are ever used, but they all take up a significant amount of
memory. How do I reduce the memory utilization?
Thanks,
Scott Queen
