Jeff, Thanks for your help with this. I'm not on the very latest mod_security version (few months old). I'll try upgrading it, and following up with mod_security community if that doesn't resolve it. Thanks for the tip about valgrind, I'll look into that (this is new territory for me).Thanks again, Paul Date: Sun, 27 Jul 2014 08:27:04 -0400 From: trawick@xxxxxxxxx To: users@xxxxxxxxxxxxxxxx Subject: Re: Segmentation Fault - too many proxy balancers On Sat, Jul 26, 2014 at 1:37 PM, Paul Beckett <paul_beckett@xxxxxxxxxxx> wrote:
So the crash is during request logging, in mod_security, as it passes a bad value to apr_global_mutex_lock(). Check if you have the latest version of mod_security and/or if those folks have seen this crash before.
Increasing some type of resource (e.g., number of proxy balancers) can push things around in memory and perhaps a latent bug is now exposed. Sometimes running httpd under a memory debugger like valgrind can catch bad use of memory at an earlier stage. Consider that if you can reproduce this in a test environment.
|