nginx 1.0.2 intermittent segfault error 4

[Maxim Dounin]

Hello!

On Sat, May 14, 2011 at 11:16:22AM -0400, inph wrote:

> Hi,
> 
> I've noticed some nginx segfaults over the last week or so with 1.0.1
> and 1.0.2. I havent tested any other version for long enough.
> The segfaults are not immediate and sometimes it requires several hours
> to reproduce through normal use.
> 
> Here is the debug info, if you require anymore information please ask.
> 
> [12702.705897] nginx[1769]: segfault at 50545448 ip 0807d1fa sp bfc241d0
> error 4 in nginx[8048000+82000]

[...]

>    location ~ \.php$ {
>         try_files      $uri =404;
>         access_log     /var/log/nginx/upstream.log  upstream;
>         fastcgi_pass   unix:/var/run/php-fpm.sock;
>         fastcgi_index  index.php;
>         include        fastcgi.conf;
>         fastcgi_intercept_errors        on;
>         fastcgi_ignore_client_abort     on;

Yes, thank you for your report.  The problem is with 
fastcgi_ignore_client_abort, it doesn't handle POST requests 
correctly.

Switching fastcgi_ignore_client_abort back to default (off) should 
resolve this issue.

Maxim Dounin