Hello everybody,

As of some time ago I started seeing "exit signal Segmentation fault
(11)" in /var/log/apache2/error.log. I'm running the current stable
Debian, php4 and Apache. We have quite a few users and usually hear
about any errors. But although I see a lot of these in the error log,
no one seems to be complaining.

Now, the child dies a couple of seconds later than the last access by
that child according to /var/log/apache2/access.log. So, the page must
be served completely and several seconds later, the child dies. What's
the best way to debug this one? I've enabled core dumping but no cores
are found under /tmp -- is this where they are supposed to appear? The
pages it happens to seem to be kind of random.

I've seen a lot of discussion about "Segmentation fault (11)" but I
didn't see what the resolution was. Sorry if this has been asked
before...

Best regards,
Sergei

On 11 Apr, 19:42, surge <gera... [at] gmail.com> wrote:
> Hello everybody,
>
> As of some time ago I started seeing "exit signal Segmentation fault
> (11)" in /var/log/apache2/error.log. I'm running the current stable
> Debian, php4 and Apache. We have quite a few users and usually hear
> about any errors. But although I see a lot of these in the error log,
> no one seems to be complaining.
>
> Now, the child dies a couple of seconds later than the last access by
> that child according to /var/log/apache2/access.log. So, the page must
> be served completely and several seconds later, the child dies. What's
> the best way to debug this one? I've enabled core dumping but no cores
> are found under /tmp -- is this where they are supposed to appear? The
> pages it happens to seem to be kind of random.
>
> I've seen a lot of discussion about "Segmentation fault (11)" but I
> didn't see what the resolution was. Sorry if this has been asked
> before...
>
> Best regards,
> Sergei

Read this:

http://aplawrence.com/Linux/limit_core_files.html

Check ulimit and permissions for webserver uid.

C.

Thanks for the reply. I've read the article and tweaked my settings.
Now I'm waiting for a core dump, which, as luck would have it, hasn't
happened all day :)

On Apr 14, 5:18 pm, surge <gera... [at] gmail.com> wrote:
> Thanks for the reply. I've read the article and tweaked my settings.
> Now I'm waiting for a core dump, which, as luck would have it, hasn't
> happened all day :)


Hmm, the cores are still not dumped even though "kill -s SIGSEGV $$"
does create a core in the right directory. Any ideas?