Re: apache24 reload dies with SIGSEGV

From: Jim Pazarena <fquest_at_paz.bz>
Date: Fri, 12 May 2023 05:48:06 UTC
On 2023-05-11 8:46 PM, Paul Procacci wrote:
> 
> 
> On Thu, May 11, 2023 at 11:35 PM Paul Procacci <pprocacci@gmail.com 
> <mailto:pprocacci@gmail.com>> wrote:
> 
> 
> 
>     On Thu, May 11, 2023 at 11:31 PM Jim Pazarena <fquest@paz.bz
>     <mailto:fquest@paz.bz>> wrote:
> 
>         I just loaded two servers with a fresh 13.2 install . Along with
>         a fresh
>         install of 2.4.57 Apache .
> 
>         both have public IPs
>         one is https , while the other is not .
> 
>         an "./apache24 start" works as expected . It passes the sanity
>         check ,
>         announces Syntax OK , and runs .
> 
>         However , if I issue an "./apache24 reload"
>         it still announces Syntax OK , followed with:
>         Performing a graceful restart
> 
>         but it does not start .
>         running "dmesg" , I see:
>         pid xxxx (httpd), jid 0, uid 0: existed on signal 11
> 
>         If I run an ".apache24 start"
>         it fires up as expected and runs as expected .
> 
>         What is a reload doing extra that a start does not do , that
>         could be
>         causing a SIGSEGV ?
>         oh yes, a restart works as expected
>         a reload dies .
> 
>         Suggestions would be most appreciated .
> 
> 
>     You need to generate a core and run gdb against it getting a
>     backtrace. -- This ultimately would provide you very specifically
>     where it's getting foul'd up.
> 
>     ~Paul
>     -- 
>     __________________
> 
>     :(){ :|:& };:
> 
> 
> https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=267435 
> <https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=267435>
> 
> As a follow up to my brief comment and given the bug report above (last 
> comment in particular), you don't happen to be running apache with 
> mod_php and ASLR are you?
> 
> ~Paul
> 

How about that . I do indeed . My symptoms described in that bugzilla 
are not quite the same tho .

I can avoid the reload and switch to a restart . Not optimal , but it 
works . Then I can wait to follow if this bugzill ais acted on .
Thanks for the reply .