PHP-FPM error

I have been using PHP-FPM 7.3.21 without any problems.

Today I updated a whole bunch of packages and afterwards PHP-FPM began to cause errors with my wordpress website.

I switched to the PHP-CGI version and that works.

How can I diagnose what is up?

By reading logs. 

Cynic by name and cynic by nature...

Obviously but where is the path for the logs...

Cynic by name and cynic by nature...

Yup!
When in doubt.. revert to *nix basics

Code: [Select]

ls -l /var/log

Code: [Select]

find / -name "*domlogs*" -printOr you use the GUI given to you by the developers: login as user and select logs. Though this is likely only to give a partial story.
If the developers would see fit to reinstate the official CSF GUI, then you could also see the Apache log in real time, assuming that you  use Apache.

Thanks... I was posting here because I had seemingly exhausted all options of finding where the output of errors when using PHP-FPM were being logged.

I will take a look at what you recommended but up until now I cannot find the output that can show why my website is down when using PHP-FPM

I am using Apache- Nginx- Varnish

And I have looked at error logs and access logs for all of them but nothing indicates an error..

Back to web basics, this time.
Likely you won't have disabled this method.. if your installed application doesn't have something similar.
Create a file in your user's public_html called for example myphpinfo.php

Code: [Select]

<?php phpinfo(); ?>

Code: [Select]

chown user:user myphpinfo.phphttp://userdomain.com/myphpinfo.php
Scroll down or search for error_log
example:    /home/user/public_html/shop/Log/errors-20201116.txt  <- will depend on whether your application modifies the default.

P.S. Life is too short to debug the complexities of running such a stack. I used nginx reverse proxy for a couple of years and though useful, when problems arose it just created more headaches. Varnish just adds complexity. KISS philosophy - I am veering toward PHP-FPM, rather than suexec/su_php/suhosin/FastCGI.

Yes I agree too complex :-(

I did find this though in my nginx output:

[Mon Nov 16 16:32:36.633747 2020] [proxy:error] [pid 19940:tid 139994467337984] (2)No such file or directory: AH02454: FCGI: attempt to connect to Unix domain socket /opt/alt/php-fpm73/usr/var/sockets/xxxxx.sock (*) failed
[Mon Nov 16 16:32:36.633780 2020] [proxy_fcgi:error] [pid 19940:tid 139994467337984] [client xx.xxx.xx.xxx:44980] AH01079: failed to make connection to backend: httpd-UDS

xxxxx.sock is a user account.

You could try Googling the clearing of nginx reverse proxy cache, plus varnish cache, then rebooting. Though it may be a permissions issue.
My advice: stick with Apache/PHP-FPM and tune your database/Apache. Add an opcache/memcache/redis and you're good to go. Are the performance gains from the other approaches really worth it, if they're even noticeable by end users?

 

I had a similar problem this morning.

It seems that Apache is using the wrong conf file (not the CWP one) and that /etc/init.d/httpd is missing

Guys, make sure you have CWPpro activate.
Also, you can try to rebuild VirtualHosts without PHP-FPM and then go back to PHP-FPM.

correction removing httpd did also removed my init-scripts.
This worked for me:

Code: [Select]

rpm -e httpd --nodeps --justdb
rpm -e httpd-tools --nodeps --justdb
yum reinstall cwp-httpd