after reboot 502 Bad Gateway and php7.4-fpm.service not running, /run/php missing

Imorg · 8. Februar 2023

Greetings everyone.

Although for years i considered myself a happy and meanwhile even not quite inexperienced OMV profiteer i have recently found myself to be in a position where it seems that no thread is able to provide the necessary information to solve my problem. Therefore i kindly ask forgiveness if i may have overseen the solution.

The problem:

After reboot i can access the webgui but i cannot login. It gives me a 502 Bad gateway. Further investigation let me find out that the service php7.4-fpm has not started at all. The reason therefore was the missing socket which i can fix manually by using the following commands:

Code

mkdir --mode=07500 /run/php
chown www-data:www-data /run/php

omv-salt deploy run phpfpm

The first two lines i got from the starting procedure of the php7.4-fpm service

Code

nano /etc/init.d/php7.4-fpm

Now, as i understand it, the init of php7.4-fpm should be happening automatically on startup. When i do it manually php is running and i am able to login to the webgui just fine.

I have not tested to just use

Code

service php7.4-fpm start

omv-salt deploy run phpfpm

But my guess is that it would also work.

The question remains, why is it not starting automatically anymore?

I'm running OMV6 on a QNAP, installed with the debian operating system variant found here: https://docs.openmediavault.or…stallation/on_debian.html

Everything is up to date.

I'm also running docker with some websites but those don't seem to be the cause of my problem. Maybe another container i tried out is responsible: cups airprint. I tried two different of them but in the end i destroyed them both.

Any hint is appreciated!

macom · 9. Februar 2023

What is the output of

systemctl status php7.4-fpm

Imorg · 10. Februar 2023

Code

systemctl status php7.4-fpm

php7.4-fpm.service - The PHP 7.4 FastCGI Process Manager
     Loaded: loaded (/lib/systemd/system/php7.4-fpm.service; enabled; vendor preset: enabled)
     Active: active (running) since Wed 2023-02-08 16:34:22 CET; 1 day 8h ago
       Docs: man:php-fpm7.4(8)
    Process: 3548 ExecStartPost=/usr/lib/php/php-fpm-socket-helper install /run/php/php-fpm.sock /etc/php/7.4/fpm/pool.d/www.conf 74 (code=exited, status=0/SUCCESS)
   Main PID: 3545 (php-fpm7.4)
     Status: "Processes active: 0, idle: 9, Requests: 9443, slow: 0, Traffic: 1.7req/sec"
      Tasks: 10 (limit: 1090)
     Memory: 29.3M
        CPU: 22min 59.249s
     CGroup: /system.slice/php7.4-fpm.service
             ├─  3545 php-fpm: master process (/etc/php/7.4/fpm/php-fpm.conf)
             ├─  3546 php-fpm: pool www
             ├─  3547 php-fpm: pool www
             ├─330245 php-fpm: pool openmediavault-webgui
             ├─330251 php-fpm: pool openmediavault-webgui
             ├─330260 php-fpm: pool openmediavault-webgui
             ├─330261 php-fpm: pool openmediavault-webgui
             ├─330262 php-fpm: pool openmediavault-webgui
             ├─330263 php-fpm: pool openmediavault-webgui
             └─330426 php-fpm: pool openmediavault-webgui

Feb 08 16:34:22 nas systemd[1]: Starting The PHP 7.4 FastCGI Process Manager...
Feb 08 16:34:22 nas systemd[1]: Started The PHP 7.4 FastCGI Process Manager.
Feb 08 16:34:38 nas openmediavault-webgui[3557]: Authorized login from 192.168.17.70 [username=admin, user-agent=Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/10>
Feb 08 19:41:40 nas openmediavault-webgui[106167]: Unauthorized login attempt from 192.168.17.70 [username=admin, user-agent=Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/2010010>
Feb 08 19:41:43 nas openmediavault-webgui[106166]: Authorized login from 192.168.17.70 [username=admin, user-agent=Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:109.0) Gecko/20100101 Firefox/>
Feb 10 01:28:56 nas openmediavault-webgui[330245]: Authorized login from 192.168.17.70 [username=admin, user-ag...

Alles anzeigen

Another interesting fact: the QNAP's own display shows that the system is still "booting" although its up and running since 9 hours and 4 minutes !

The unauthorized login is because i made a mistake with my password once.

votdev · 10. Februar 2023

The /run/php directory should be automatically created by systemd during bootup thanks to the /usr/lib/tmpfiles.d/php7.4-fpm.conf configuration file.

votdev · 10. Februar 2023

Zitat von Imorg

Another interesting fact: the QNAP's own display shows that the system is still "booting" although its up and running since 9 hours and 4 minutes !

Your hardware incl. the LED display is proprietary, so it can only be accessed and controlled via Qnap software.

Imorg · 10. Februar 2023

The content of /usr/lib/tmpfiles.d/php7.4-fpm.conf:

Code

#Type Path                  Mode UID      GID      Age Argument
    d /run/php              0755 www-data www-data -   -

Zitat

Your hardware incl. the LED display is proprietary, so it can only be accessed and controlled via Qnap software.

That makes sense.

Imorg · 10. Februar 2023

That last thing made my curious. Found out: To control the LCD you just have to

Code

#Connect to LCD
stty -F /dev/ttyS1 1200

#Turn it on
echo -e "M^\x1" > /dev/ttyS1

#Turn it off
echo -e "M^\x0" > /dev/ttyS1

#Write in Line 1 (Apparently has to have exactly 16 characters)
echo -e "M\f\x0 HELLO WORLD 1234" > /dev/ttyS1

#Write in Line 2 (Apparently has to have exactly 16 characters)
echo -e "M\f\x1 HELLO WORLD 1234" > /dev/ttyS

Alles anzeigen

Thanks to paulvanleest from https://forum.qnap.com/viewtopic.php?t=19180

Imorg · 10. Februar 2023

I just tried do start the service php7.4-fpm after reboot with no success:

Code

root@nas:~# service php7.4-fpm start
Job for php7.4-fpm.service failed because the control process exited with error code.
See "systemctl status php7.4-fpm.service" and "journalctl -xe" for details.

systemctl:

Code

root@nas:~# systemctl status php7.4-fpm.service
● php7.4-fpm.service - The PHP 7.4 FastCGI Process Manager
     Loaded: loaded (/lib/systemd/system/php7.4-fpm.service; enabled; vendor preset: enabled)
     Active: failed (Result: exit-code) since Fri 2023-02-10 22:00:06 CET; 8s ago
       Docs: man:php-fpm7.4(8)
    Process: 3377 ExecStart=/usr/sbin/php-fpm7.4 --nodaemonize --fpm-config /etc/php/7.4/fpm/php-fpm.conf (code=exited, status=78)
    Process: 3389 ExecStopPost=/usr/lib/php/php-fpm-socket-helper remove /run/php/php-fpm.sock /etc/php/7.4/fpm/pool.d/www.conf 74 (code=exited, status=0/SUCCESS)
   Main PID: 3377 (code=exited, status=78)
        CPU: 201ms

Feb 10 22:00:06 nas systemd[1]: Starting The PHP 7.4 FastCGI Process Manager...
Feb 10 22:00:06 nas php-fpm7.4[3377]: [10-Feb-2023 22:00:06] ERROR: unable to bind listening socket for address '/run/php/php7.4-fpm-openmediavault-webgui.sock': No such file or directory >
Feb 10 22:00:06 nas php-fpm7.4[3377]: [10-Feb-2023 22:00:06] ERROR: FPM initialization failed
Feb 10 22:00:06 nas systemd[1]: php7.4-fpm.service: Main process exited, code=exited, status=78/CONFIG
Feb 10 22:00:06 nas systemd[1]: php7.4-fpm.service: Failed with result 'exit-code'.
Feb 10 22:00:06 nas systemd[1]: Failed to start The PHP 7.4 FastCGI Process Manager.

Alles anzeigen

journalctl (too big to post it here, therefore i share it this way. One version even has syntax highlighting!

journalctl

KliefertCloud

cloud.kliefert.net

(If not available my nas is probably rebooting or down for maintenance - i will add some RAM soon)

but with

Code

root@nas:~# mkdir --mode=07500 /run/php && chown www-data:www-data /run/php && omv-salt deploy run phpfpm
             Summary for nas.fritz.box
              -------------
              Succeeded: 13 (changed=1)
              Failed:     0
              -------------
              Total states run:     13
              Total run time:    1.211 s
----------
          ID: test_phpfpm_service_config
    Function: cmd.run
        Name: php-fpm7.4 --test
      Result: True
     Comment: Command "php-fpm7.4 --test" run
     Started: 22:14:54.832325
    Duration: 553.705 ms
     Changes:   
              ----------
              pid:
                  5971
              retcode:
                  0
              stderr:
                  [10-Feb-2023 22:14:55] NOTICE: configuration file /etc/php/7.4/fpm/php-fpm.conf test is successful
              stdout:
----------
          ID: restart_phpfpm_service
    Function: service.running
        Name: php7.4-fpm
      Result: True
     Comment: Service php7.4-fpm is already enabled, and is running
     Started: 22:14:55.448810
    Duration: 920.903 ms
     Changes:   
              ----------
              php7.4-fpm:
                  True

Alles anzeigen

everything went back to normal again...

Imorg · 13. Februar 2023

I put this in my crontab:

Code

@reboot sleep 120 && mkdir --mode=07500 /run/php && chown www-data:www-data /run/php && omv-salt deploy run phpfpm

and now i do not even notice this unusual bug anymore

garyt · 7. März 2023

I have experienced this issue when upgrading from OMV4 to OMV6.

The issue is that the PHP package now puts its openmediavault socket in the wrong place.

Open up the OpenMediaVault-specific PHP configuration:

sudo vi /etc/php/7.4/fpm/pool.d/openmediavault-webgui.conf

Edit the (4th) line (that starts with "listen ="):

listen = /run/php/php7.4-fpm-openmediavault-webgui.sock

(Originally, it said something along the lines of "/var/run/..." for me without mentioning the PHP version number. That was wrong.)

Then restart PHP-fpm:

sudo systemctl restart php7.4-fpm

This works on OpenMediaVault 6.3.1.

Most probably, the underlying issue is that during the "openmediavault" package installation the job running the salt/omv/deploy/phpfpm/10webgui.sls function is failing or something. I didn't look at the installation logs.

Imorg · 20. März 2023

Unfortunately this solution doesnt fit here, because the mentioned file

/etc/php/7.4/fpm/pool.d/openmediavault-webgui.conf

already looks like this:

listen = /run/php/php7.4-fpm-openmediavault-webgui.sock

garyt · 5. April 2023

That's unfortunate. Apparently, you have a different bug. (I should've read through your post in more detail.)

I wonder what the new 6.3.7 version's "Improve Salt configuration." brought to the table...

xcraft · 5. Mai 2023

@author

were you able to fix yours the "official way" or still using your crontab fix? I'm have the same exact issue. I thought mine started after two power outages in a span of 15 minutes but no because after failing to fix my issue I started from scratch and after everything setup, I tried to reboot then the issue again. Now I suspect a recent update with something like lib***, I forgot, there were 7 of them.

I already uninstalled flashmemory plugin and tried all other "fixes" I found in this forum but still the issue remains. I am now using the crontab fix but this at most a dirty fix.

systemctl status php7.4-fpm

Code

● php7.4-fpm.service - The PHP 7.4 FastCGI Process Manager
     Loaded: loaded (/lib/systemd/system/php7.4-fpm.service; enabled; vendor preset: enabled)
     Active: failed (Result: exit-code) since Fri 2023-05-05 09:59:10 PST; 17s ago
       Docs: man:php-fpm7.4(8)
    Process: 18109 ExecStart=/usr/sbin/php-fpm7.4 --nodaemonize --fpm-config /etc/php/7.4/fpm/php-fpm.conf >
    Process: 18110 ExecStopPost=/usr/lib/php/php-fpm-socket-helper remove /run/php/php-fpm.sock /etc/php/7.>
   Main PID: 18109 (code=exited, status=78)
        CPU: 131ms

May 05 09:59:10 omvraspi systemd[1]: Starting The PHP 7.4 FastCGI Process Manager...
May 05 09:59:10 omvraspi php-fpm7.4[18109]: [05-May-2023 09:59:10] ERROR: unable to bind listening socket f>
May 05 09:59:10 omvraspi php-fpm7.4[18109]: [05-May-2023 09:59:10] ERROR: FPM initialization failed
May 05 09:59:10 omvraspi systemd[1]: php7.4-fpm.service: Main process exited, code=exited, status=78/CONFIG
May 05 09:59:10 omvraspi systemd[1]: php7.4-fpm.service: Failed with result 'exit-code'.
May 05 09:59:10 omvraspi systemd[1]: Failed to start The PHP 7.4 FastCGI Process Manager.

Alles anzeigen

I don't know what else to provide, please help. Also a question, how important is the GUI anyway, after you setup everything, the mounts, the services? Will PHP being down affect any service aside from WEB-GUI?

Edit: NVM, fixed mine. It was an outside process messing up OMV installation

after reboot 502 Bad Gateway and php7.4-fpm.service not running, /run/php missing

crashtest 8. Februar 2023

Jetzt mitmachen!

Tags