Hello all,
I recently had to switch from prefork to worker on a large production
system, in order to optimize performance and limit memory usage under
high load.
In principle, it is working fine, however I see that over time some
processes are "hanging" even though they should have been killed off
(recycled).
The worker table in server-status looks like this:
W....G........WW.....GW.........................................
..........W......W..............................................
..........W.......W.............................................
............WG..........W.......................................
....................W...........................................
..........W..WW.WWW....W........................................
WWW_____W_____________W_W.......................................
__________________W______.......................................
................................................................
................................................................
................................................................
................................................................
................................................................
................................................................
................................................................
................................................................
My question is: Will the top 6 processes be killed "for real" at any
stage, or will the server stop accepting connections once the
remaining processes have gone into the same state? I can manually
kill off the hanging processes (kill -9), but this is risky at best
as I don't know the PIDs (and am unsure how to find this information).
About the setup:
HP DL385 dual dualcore opteron servers
FreeBSD 6.1-RELEASE
Apache version 2.0.58_1
Mod_jk version 1.2.15
Tomcat version 5.5.12_3
The apache server is configured with the minimum number of modules,
no PHP or other sensitive stuff. The only indication I have on where
the problem may lie is that I haven't been able to reproduce when
simulating load on a path that isn't mod_jk-mounted, however
reproducing has proven difficult in any case.
Does anyone have any suggestions? Or am I worried without reason?
With best regards,
Eirik Øverby
Unicore AS
Oslo, Norway
---------------------------------------------------------------------
The official User-To-User support forum of the Apache HTTP Server Project.
See <URL:http://httpd.apache.org/userslist.html> for more info.
To unsubscribe, e-mail: [EMAIL PROTECTED]
" from the digest: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
