Issue Nginx 502 Bad gateway on Php 7.2 with SSL

[Kyobul]

If I set a web hosting for a domain to PHP 7.2 handled by Nginx with SSL enabled I'm always getting a 502 Bad Gateway when accessing the website.

If I switch to lower version of PHP - 7.1 - everything is fine with no single other change.

Error log is recv() failed (104: Connection reset by peer) while reading response header from upstream

Plesk Onyx is latest, did plesk repair instlallation, digged the Internet about 2 hours and since I'm not a sysadmin, I have no clue.

Any ideas ?

 

[Bitpalast]

Is the service "plesk-php72-fpm" running? Can you start it with
# service plesk.php72-fpm restart
?

 

[Kyobul]

Thank you for your reply. Yes, indeed, the plesk-php72-fpm service is not running and there is no way it can be started.

I have uninstalled and re-installed the PHP 7.2 with the plesk installer. Nothing changes that.

Error message on # service plesk.php72-fpm restart is :

Redirecting to /bin/systemctl restart plesk.php72-fpm.service
Failed to restart plesk.php72-fpm.service: Unit not found

 

[Bitpalast]

Sorry for the typo, it is
# service plesk-php72-fpm restart

 

[Kyobul]

Thank you. Here is what I get :

# service plesk-php72-fpm restart
Redirecting to /bin/systemctl restart plesk-php72-fpm.service
Job for plesk-php72-fpm.service failed because the control process exited with error code. See "systemctl status plesk-php72-fpm.service" and "journalctl -xe" for details.


# /bin/systemctl status plesk-php72-fpm.service -l
● plesk-php72-fpm.service - The PHP 7.2.2 FastCGI Process Manager
Loaded: loaded (/usr/lib/systemd/system/plesk-php72-fpm.service; disabled; vendor preset: disabled)
Active: failed (Result: exit-code) since Thu 2018-02-22 23:50:21 CET; 1min 55s ago
Process: 23263 ExecStart=/opt/plesk/php/7.2/sbin/php-fpm --nodaemonize (code=exited, status=78)
Main PID: 23263 (code=exited, status=78)

Feb 22 23:50:21 s20801255 systemd[1]: Starting The PHP 7.2.2 FastCGI Process Manager...
Feb 22 23:50:21 s20801255 php-fpm[23263]: [22-Feb-2018 23:50:21] WARNING: Nothing matches the include pattern '/opt/plesk/php/7.2/etc/php-fpm.d/*.conf' from /opt/plesk/php/7.2/etc/php-fpm.conf at line 119.
Feb 22 23:50:21 s20801255 php-fpm[23263]: [22-Feb-2018 23:50:21] ERROR: No pool defined. at least one pool section must be specified in config file
Feb 22 23:50:21 s20801255 php-fpm[23263]: [22-Feb-2018 23:50:21] ERROR: failed to post process the configuration
Feb 22 23:50:21 s20801255 php-fpm[23263]: [22-Feb-2018 23:50:21] ERROR: FPM initialization failed
Feb 22 23:50:21 s20801255 systemd[1]: plesk-php72-fpm.service: main process exited, code=exited, status=78/n/a
Feb 22 23:50:21 s20801255 systemd[1]: Failed to start The PHP 7.2.2 FastCGI Process Manager.
Feb 22 23:50:21 s20801255 systemd[1]: Unit plesk-php72-fpm.service entered failed state.
Feb 22 23:50:21 s20801255 systemd[1]: plesk-php72-fpm.service failed.

 

[Bitpalast]

Normally in this situation the service does not start, because none of the websites is configured to use PHP 7.2 FPM. You can try to do this: Toggle the PHP setting between PHP 5.6 and PHP 7.2, then try to start the service again:
# service plesk-php72-fpm start
It should then start, because toggeling the configuration will create the missing configuration files for the PHP 7.2 FPM service.

 

[Kyobul]

Thank you for your reply. Indeed, no site was running since I wasn't able to get it running on PHP 7.2. Switching it back to PHP 7.2 (which creates config files and makes PHP.7.2-FPM running) helps nothing. So, to be sure I a proprer / new website installation : creating brand new subscrption and installing basic wordpress on it, no SSL on it.

When configured to PHP 7.2 with Nginx - always shows 502 Bad Gateway Nginx, log tells :
recv() failed (104: Connection reset by peer) while reading response header from upstream

When it is switched to PHP 7.1 with Nginx - eveything works just nice.

I suppose I'll be just sticking with PHP 7.1....

 

[Bitpalast]

Switching it back to PHP 7.2 (which creates config files and makes PHP.7.2-FPM running) helps nothing.

If the config files exist as you are saying, how can the startup routine still display
"ERROR: No pool defined. at least one pool section must be specified in config file"
as you are saying, too?

Do the config files really exist?

 

[Kyobul]

The error "ERROR: No pool defined. at least one pool section must be specified in config file" was logged when I was indeed not having a single site running on PHP 7.2
It is not relevant to my last message.