From: Steve Nelson (sanelson@gmail.com)
Date: Mon Jul 25 2005 - 06:14:27 EDT
Hello All,
I have two webservers, each of which is handling a similar number of
threads (about 600).
However one machine's ns-httpd has consistently been using 80%+ CPU
for the whole weekend, with %sys greater than %usr, the other has
rarely used more than 2%.
Looking deeper, the 'problem' machine has been performing around
600,000 system calls a second, the other around 2000. Similarly,
context switching on the problem machine is excessive, and on the 'ok'
one is fine.
What steps can I take to investigate this further and determine what
is making these system calls? What sort of methodology should I be
emplying to establish what is the difference between the good machine
and the bad machine?
In terms of network activity the two are very similar, according to
netstat -in - what else should I look at?
Should I truss the ns-httpd process? What other options are available?
Thanks in advance,
Steve Nelson
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers
This archive was generated by hypermail 2.1.7 : Wed Apr 09 2008 - 23:31:10 EDT