On 7/26/06, Nick Wilson <nick@xxxxx> wrote:
I don't know if no one saw it, or no one can help, but I'll try one more time, this is honestly my last hope and im not sure what else to do ________________________________ Im having trouble debugging a segmentation fault, the server is only running a phpAdsNew ads server. The system: Dual opteron box running gentoo 2.6 kernel Apache 2.0.58 PHP 5.1.4 When we hit our peak hours is when the seg faults start occurring: [Tue Jul 25 13:09:32 2006] [notice] child pid 20868 exit signal Segmentation fault (11) [Tue Jul 25 13:09:32 2006] [notice] child pid 20877 exit signal Segmentation fault (11) [Tue Jul 25 13:09:32 2006] [notice] child pid 20895 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20340 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20698 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20738 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20826 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20858 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20866 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20884 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20886 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20893 exit signal Segmentation fault (11) [Tue Jul 25 13:09:34 2006] [notice] child pid 20903 exit signal Segmentation fault (11) This will happen every 5-10 minutes in blocks like that. At first I thought it was a ram issue, but when I moved it to another web server running apache 2.0.58/php5.0.4 the same thing happens. I've attempted to setup a core dump directory using the CoreDumpDirectory directive, the directory is owned by the apache user and "ulimit –c unlimited" was run before starting apache, however nothing is ever outputted. I did read that this has something to do with apache starting as root and then switching to the correct user, but with kernel 2.4+ it should still output, maybe im wrong. Next I tried to setup mod_whatkilledus (http://people.apache.org/~trawick/mod_whatkilledus.c), I compiled apache with –enable-exception-hook, and I have the EnableExceptionHook directive on and the module is loaded, however as you can see from my log output mod_whatkilledus is not doing anything. Is there something im missing to make both of these work? Any help would be appreciated as this is driving me nuts =] -Nick Nick Wilson Web Developer Digitally Imported, Inc. http://www.di.fm http://www.sky.fm
first, digitally imported is a wonderful webcast! second: try this sequence: is the Harddrive the same? if so, try another drive. is the Distro the same? if so , and the HD dosent fix it, try with a new distro only. what does WhatKilledUs do? what about mod_systemstats (its with my apache (debian 2006.3.5), and i've found that can help... also, try a different version of apache. I run apache 2.1(something), and Have run 2.0.5x - it usually fails if CurrentClients > 150% of MaxClients. set the httpd.conf to have a maxClients set. that fixed it for me. -- "Space does not reflect society, it expresses it." -- Castells, M., Space of Flows, Space of Places: Materials for a Theory of Urbanism in the Information Age, in The Cybercities Reader, S. Graham, Editor. 2004, Routledge: London. p. 82-93.