Re: Failing FastCGI PHP
| From: | Frank de Bot | Date: | Sat, 13 May 2006 21:39:30 +0000 |
| Subject: | Re: Failing FastCGI PHP | ||
| References: | 1 2 | Groups: | php.general |
| Request: | Send a blank email to php-general+get-236123@lists.php.net to get a copy of this message | ||
I'll start by compiling php with --enable-debug
At the moment I get backtrace results like this:
#0 0x48c7d95b in memcpy () from /usr/lib/libc_r.so.4
#1 0x8977280 in ?? ()
#2 0x10 in ?? ()
#3 0x894e500 in ?? ()
#4 0x894dc00 in ?? ()
#5 0x8962000 in ?? ()
#6 0x8960200 in ?? ()
#7 0x894e4e0 in ?? ()
#8 0x894e4c0 in ?? ()
etc etc etc...
Thus useless :P
At the moment it has been running remarkably stable for a few days. So now and then a failure. Probably because of a request being given to a php-fastcgi process handling a slow upload of a file... :-/
Frank de Bot
chris smith wrote:
On 5/11/06, Frank de Bot <php@searchy.nl> wrote:Hi, I have made a apache setup with PHP running as FastCGI (apache 2.0.58, mod_fastcgi 2.4.2, php 5.1.4) PHP is called directly as FastCGI without a wrapper script. On a given moment the PHP FastCGI application reaches dynamicMaxClassProcs. After that happens the PHP FastCGI application totally crashes and can only be recovered by restarting apache. PHP Crashes with: "terminated due to uncaught signal '10' ((null)), a core file may have been generated. "Are you able to get a core dump and what does it show? Here's some doco on how to get a backtrace: http://bugs.php.net/bugs-generating-backtrace.php That will give us some clues on whether it's php or fastcgi..