Date: Wed, 24 Aug 2005 10:03:52 -0700 From: Lei Sun <lei.sun@gmail.com> To: Dan Nelson <dnelson@allantgroup.com> Cc: questions@freebsd.org Subject: Re: oversized httpd process? Message-ID: <d396fddf050824100345063d98@mail.gmail.com> In-Reply-To: <20050824143133.GA88693@dan.emsphone.com> References: <d396fddf05082400304744960f@mail.gmail.com> <20050824143133.GA88693@dan.emsphone.com>
next in thread | previous in thread | raw e-mail | index | archive | help
Now that I think about it, I did install the eaccelerator for php, and configured it to take 128Mb for eaccelerator.shm_size. It also appears that when I reboot the machine A without pointing my browser to it once, the httpd processes are quite small. BUT... Even with the 2 php application that I installed (phpmyadmin, mediawiki), I don't think there should be that much caching going on. Since if I add the total size of the 2 application together, they won't even hit 20MB. In other words, my understanding would be: even if eaccelerator wants to cache them all, eaccelerator wouldn't be able to find that much stuff to cache, and it would always be less than 20MB. Am I expecting the right thing? On 8/24/05, Dan Nelson <dnelson@allantgroup.com> wrote: > In the last episode (Aug 24), Lei Sun said: > > I saw many posts from google regarding to this question, but there > > were no definite answers.. > > > > some say, it's mod_ssl, some say it's mod_perl, some say it is mm. > > But my case, it just doesn't make much sence to me at all. > > > > Here are the 2 test machines that I have, both have the exact same > > configuration > > > > A is a lot more powerful than B > > > > machine A, p4 3.0 2GB Mem > > machine B, p2 450Mhz 128MB Mem > > > > Both have mod_php, mod_ssl, and no traffic has been sent. > > > > Looking at the httpd sizes, I start to wonder ... How come Machine B > > only uses around 15Mb per httpd, while machine A takes 155Mb, and > > while they have exactly the same software, same configuration. >=20 > Try running lsof on both processes. Since SIZE is 155M but RES is a > lot smaller, there may be a large file being mmapped by one system and > not the other. >=20 > > machine A: > > PID USERNAME PRI NICE SIZE RES STATE C TIME WCPU CPU COM= MAND > > 440 root 96 0 155M 17412K select 0 0:02 0.00% 0.00% htt= pd > > > > machine B: > > PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMA= ND > > 50855 www 20 0 16348K 12K lockf 0:00 0.00% 0.00% httpd >=20 > -- > Dan Nelson > dnelson@allantgroup.com >
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?d396fddf050824100345063d98>