Fastcgi error - FastCgiIpcDir problems in error log - apache2

Question

FastCgiIpcDir problems in error log
Hi,
I have noticed in my Apache error logs the following error (error 1):
[Wed Feb 08 14:00:06 2012] [alert] [client 41.185.88.175] (2)No such file or directory: FastCGI: failed to connect to (dynamic) server "/var/www/bin/php-splashpage-user/php-fastcgi": something is seriously wrong, any chance the socket/named_pipe directory was removed?, see the FastCgiIpcDir directive
Directly afterwards followed by this error (error 2):
[Wed Feb 08 14:00:06 2012] [error] [client 41.185.88.175] FastCGI: incomplete headers (0 bytes) received from server "/var/www/bin/php-splashpage-user/php-fastcgi"
How do I fix error 1?
I read about changing this due to the host system cleaning out the "/tmp" directory ( the default dir for fastCgiIpcDir if not defined ) periodically and thus obliterating communication with current active FastCGI services. So I decided to give it a go. I set the FastCgiIpcDir in fastcgi.conf file in hopes of success, but there is simply no change.
This is the contents of my fastcgi.conf file:
<IfModule mod_fastcgi.c>
FastCgiIpcDir /var/lib/apache2/fastcgi_test
FastCgiConfig -idle-timeout 60 -maxClassProcesses 1
FastCgiWrapper On
AddHandler php5-fcgi .php
Action php5-fcgi /cgi-bin/php-fastcgi
<Location "/cgi-bin/php-fastcgi">
Order Deny,Allow
Deny from All
Allow from env=REDIRECT_STATUS
Options ExecCGI
SetHandler fastcgi-script
</Location>
</IfModule>
Permissions and onwerships of /var/lib/apache2/fastcgi_test:
drwxr-xr-x 3 www-data www-data 4.0K 2012-02-08 09:20 fastcgi_test
My php wrapper script php_fastcgi has the following lines:
#!/bin/sh
PHP_FCGI_CHILDREN=120
export PHP_FCGI_CHILDREN
PHP_FCGI_MAX_REQUESTS=1000
export PHP_FCGI_MAX_REQUESTS
umask 0022
exec /usr/bin/php-cgi -d apc.shm_size=50
I am running PHP 5.3.1, Apache/2.2.14, Ubuntu 10.04.
Here's few things I've picked up so far:
Error 1 only appears at the beginning of an hour say 6 seconds just after the new hour
From working with mod_FastCgi I have learnt that increasing the child processes help relieve some of the "error 2" errors ( which cause the HTTP 500 error at random intervals). Currently I am not quite sure what the affect of error 1 would be, however if error 2 follows directly after then it's safe to say it's not a good thing.
There is very little, if any, full information on errors reported by Fastcgi with tried and tested solutions. Sadly I may just be adding onto the piles of Fastcgi errors posted on the web with no conclusion.
Your help, advice or tips in resolving error 1 would be readily appreciated.

Answer 1

I don't know how to make it work with a wrapper, and suexec, but you should try this:
http://blog.kmp.or.at/2013/06/apache-2-2-on-debian-wheezy-w-php-fpm-fastcgi-apc-and-a-kind-of-suexec/
The solution in the link does not even use suexec, nor the wrapper, and that way at least it works.
The steps required for that:
0) install php5-fpm, and apache-mpm-worker if not already installed
1) comment this line :
#FastCgiWrapper On
2) make an alias:
Alias /cgi-bin/php-fastcgi **/var/something**
3) fastcgiexternalserver:
FastCgiExternalServer **/var/something** -socket php5-fpm-site1user.sock
(The strongly typed paths must be the same)
4) set up the conf in php5-fpm/pool.d/site1user.conf
[site1user]
user = site1user
group = site1user
listen = /var/run/php5-fpm-site1user.sock
pm = dynamic
pm.max_children = 5
pm.start_servers = 2
pm.min_spare_servers = 1
pm.max_spare_servers = 3
chdir = /
5) restart fpm
service php5-fpm restart
6) for deeper understanding check my other answer here
https://serverfault.com/questions/524708/php5-fpm-apache2-on-wheezy-connect-failed-with-fastcgi/536277#536277