My only suggestion, other than increasing certain counts, is try doing a full stop and start instead of a graceful restart. There was a bug not too long ago that caused the 'Bad file descriptor: apr_socket_accept:' output to the log file when a graceful restart, which when occurring, caused a lot of headaches for people trying to keep Apache up.<br>
<br>-- Ross<br><br><div class="gmail_quote">On Wed, Jul 2, 2008 at 12:50 PM, ZuG <<a href="mailto:ozugo@yahoo.com">ozugo@yahoo.com</a>> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
Sure, here's the apache error.log. Not very interesting. The crash, according to our network monitor script, happened about 3:30.<br>
<br>
Also, the problem probably isn't ssl because I have another server crashing that doesn't have it installed.<br>
<br>
[Tue Jul 01 15:48:36 2008] [debug] ssl_engine_kernel.c(1789): OpenSSL: Exit: error in SSLv2/v3 read client hello A<br>
[Tue Jul 01 15:48:36 2008] [info] [client <a href="http://69.89.98.133" target="_blank">69.89.98.133</a>] SSL handshake failed: HTTP spoken on HTTPS port; trying to send HTML error page<br>
[Tue Jul 01 15:48:36 2008] [info] SSL Library Error: 336027804 error:1407609C:SSL routines:SSL23_GET_CLIENT_HELLO:http request speaking HTTP to HTTPS port!?<br>
[Tue Jul 01 23:23:59 2008] [error] [client <a href="http://60.172.219.2" target="_blank">60.172.219.2</a>] File does not exist: /var/www/sharedip/myproxies<br>
[Wed Jul 02 03:30:03 2008] [notice] Graceful restart requested, doing restart <-- this is our network monitoring process trying to restart the server, it failed at about this time.<br>
[Wed Jul 02 03:30:03 2008] [error] (9)Bad file descriptor: apr_socket_accept: (client socket)<br>
[Wed Jul 02 03:30:03 2008] [error] (9)Bad file descriptor: apr_socket_accept: (client socket)<br>
[Wed Jul 02 03:30:03 2008] [error] (9)Bad file descriptor: apr_socket_accept: (client socket)<br>
[Wed Jul 02 03:30:03 2008] [error] (9)Bad file descriptor: apr_socket_accept: (client socket)<br>
[Wed Jul 02 03:30:03 2008] [error] (9)Bad file descriptor: apr_socket_accept: (client socket)<br>
/usr/sbin/httpd2-prefork: symbol lookup error: /usr/lib/php5/extensions/ffmpeg.so: undefined symbol: av_free_static<br>
[Wed Jul 02 03:30:04 2008] [info] mod_fcgid: Process manager 10854 stopped<br>
[Wed Jul 02 09:34:48 2008] [info] Init: Seeding PRNG with 136 bytes of entropy<br>
[Wed Jul 02 09:34:48 2008] [info] Loading certificate & private key of SSL-aware server<br>
[Wed Jul 02 09:34:48 2008] [debug] ssl_engine_pphrase.c(469): unencrypted RSA private key - pass phrase not required<br>
[Wed Jul 02 09:34:48 2008] [info] Init: Generating temporary RSA private keys (512/1024 bits)<br>
[Wed Jul 02 09:34:48 2008] [info] Init: Generating temporary DH parameters (512/1024 bits)<br>
[Wed Jul 02 09:34:48 2008] [info] Init: Initializing (virtual) servers for SSL<br>
[Wed Jul 02 09:34:48 2008] [info] Configuring server for SSL protocol<br>
<br>
<br>
<br>
<br>
--- On Wed, 7/2/08, Ross Smith <<a href="mailto:rsmith1916@gmail.com">rsmith1916@gmail.com</a>> wrote:<br>
<br>
> From: Ross Smith <<a href="mailto:rsmith1916@gmail.com">rsmith1916@gmail.com</a>><br>
> Subject: Re: [GLLUG] Apache crashing?<br>
> To: <a href="mailto:ozugo@yahoo.com">ozugo@yahoo.com</a><br>
> Cc: <a href="mailto:linux-user@egr.msu.edu">linux-user@egr.msu.edu</a><br>
> Date: Wednesday, July 2, 2008, 12:34 PM<br>
<div><div></div><div class="Wj3C7c">> On Wed, Jul 2, 2008 at 9:47 AM, ZuG <<a href="mailto:ozugo@yahoo.com">ozugo@yahoo.com</a>><br>
> wrote:<br>
><br>
> > Hey guys,<br>
> ><br>
> > I'm having trouble with my apache 2.2.3 system.<br>
> Installed ffmpeg,<br>
> > ffmpeg-php, flvtool2, and menconder to run dolphin CMS<br>
> system.<br>
> ><br>
> > Now, apache is crashing randomly, perhaps 2 times a<br>
> week. I turned the<br>
> > apache logging up fully to try to track it down, and<br>
> the apache logs just<br>
> > donīt show anything out of the ordinary. Everything<br>
> is working fine, and<br>
> > then nothing.<br>
> ><br>
> > Do I get to play the "uninstall stuff until it<br>
> stops" game, or is there a<br>
> > better suggestion as to how to track this down?<br>
> I'm pretty new to<br>
> > troubleshooting apache, though I know my way around<br>
> the machine pretty well.<br>
> ><br>
> ><br>
> > Erica<br>
><br>
><br>
> Any chance you could post the apache and/or messages log<br>
> file contents for<br>
> the time around when the crash occurs? There may be<br>
> something obscure that<br>
> an extra pair of eyes might help to find. Also, is<br>
> anything else crashing<br>
> or just apache? There is a good chance uninstalling ffmpeg<br>
> and the other<br>
> programs won't help your case. If anything, I would<br>
> suggest re-installing<br>
> Apache if you take that route. Really though, that's a<br>
> last route. Try<br>
> changing some settings within apache such as the increasing<br>
> the<br>
> 'MaxRequestPerChild' and 'ThreadsPerChild'<br>
> counts. It could be as simple as<br>
> a memory issue (hopefully).<br>
><br>
> -- Ross<br>
<br>
</div></div></blockquote><div><br>It's at the top :P <br></div></div><br>