process corruption on CentOS 5.6

[Maxim Dounin]

Hello!

On Wed, Sep 21, 2011 at 11:03:31PM +0300, Avishai Ish-Shalom wrote:

> On CentOS 5.6, nginx 1.1.4 configured to work with php-fpm using fastcgi
> port. nginx was compiled with: configure --prefix=/usr/local/nginx
> --with-http_ssl_module --with-http_realip_module
> --with-http_geoip_module --with-http_flv_module
> --with-http_gzip_static_module --with-http_stub_status_module
> --with-http_addition_module --with-http_image_filter_module
> --with-http_sub_module --with-http_degradation_module
> --add-module=addons/ngx_upstream_fair --with-cc-opt="-D
> NGX_UPSTREAM_KEEPALIVE_PATCHED"

Just a side note: you don't need --with-cc-opt="-D 
NGX_UPSTREAM_KEEPALIVE_PATCHED"'.

> 
> every few seconds, i get:
> 
> *** glibc detected *** nginx: worker process: double free or corruption
> (!prev): 0x000000001d069800 ***

Could you please provide backtrace?  See here for details:
http://wiki.nginx.org/Debugging

Alternatively, please try the patch from here:
http://mailman.nginx.org/pipermail/nginx-devel/2011-September/001233.html

(And provide backtrace if it doesn't help.)

Maxim Dounin