Start crond with s6 overlay

[vladaurosh]

Is there an existing issue for this?

• I have searched the existing open and closed issues

Is your feature request related to a problem? Please describe

Hi @jokob-sk
Just noticed in https://github.com/jokob-sk/NetAlertX/blob/main/dockerfiles/init.sh#L129 that crond is started in the background.

Describe the solution you'd like

Since we already use s6 overlay I think it would be good idea to move it into https://github.com/jokob-sk/NetAlertX/blob/main/dockerfiles/start.sh

Probably this would do:

# s6 overlay setup
mkdir -p /etc/s6-overlay/s6-rc.d/{SetupOneshot,crond/dependencies.d,php-fpm/dependencies.d,nginx/dependencies.d,$APP_NAME/dependencies.d}
echo "oneshot" > /etc/s6-overlay/s6-rc.d/SetupOneshot/type
echo "longrun" > /etc/s6-overlay/s6-rc.d/crond/type
echo "longrun" > /etc/s6-overlay/s6-rc.d/php-fpm/type
echo "longrun" > /etc/s6-overlay/s6-rc.d/nginx/type
echo "longrun" > /etc/s6-overlay/s6-rc.d/$APP_NAME/type
echo -e "${INSTALL_DIR}/dockerfiles/init.sh" > /etc/s6-overlay/s6-rc.d/SetupOneshot/up
echo -e "#!/bin/execlineb -P\n/usr/sbin/crond -f -d 8" > /etc/s6-overlay/s6-rc.d/crond/run
echo -e "#!/bin/execlineb -P\n/usr/sbin/php-fpm83 -F" > /etc/s6-overlay/s6-rc.d/php-fpm/run
echo -e '#!/bin/execlineb -P\nnginx -g "daemon off;"' > /etc/s6-overlay/s6-rc.d/nginx/run
...
touch /etc/s6-overlay/s6-rc.d/user/contents.d/{SetupOneshot,crond,php-fpm,nginx,$APP_NAME} /etc/s6-overlay/s6-rc.d/{crond,php-fpm,nginx,$APP_NAME}/dependencies.d/SetupOneshot
touch /etc/s6-overlay/s6-rc.d/nginx/dependencies.d/php-fpm
touch /etc/s6-overlay/s6-rc.d/$APP_NAME/dependencies.d/nginx

Also I've noticed:

mkdir -p /etc/s6-overlay/s6-rc.d/{SetupOneshot,php-fpm/dependencies.d,nginx/dependencies.d}
mkdir -p /etc/s6-overlay/s6-rc.d/{SetupOneshot,php-fpm/dependencies.d,nginx/dependencies.d,$APP_NAME/dependencies.d}

First line is not needed so block should be replaced with:
mkdir -p /etc/s6-overlay/s6-rc.d/{SetupOneshot,crond/dependencies.d,php-fpm/dependencies.d,nginx/dependencies.d,$APP_NAME/dependencies.d}

Same goes for:

touch /etc/s6-overlay/s6-rc.d/user/contents.d/{SetupOneshot,php-fpm,nginx} /etc/s6-overlay/s6-rc.d/{php-fpm,nginx}/dependencies.d/SetupOneshot
touch /etc/s6-overlay/s6-rc.d/user/contents.d/{SetupOneshot,php-fpm,nginx,$APP_NAME} /etc/s6-overlay/s6-rc.d/{php-fpm,nginx,$APP_NAME}/dependencies.d/SetupOneshot

First line is not needed, these 2 should be replaced with:
touch /etc/s6-overlay/s6-rc.d/user/contents.d/{SetupOneshot,crond,php-fpm,nginx,$APP_NAME} /etc/s6-overlay/s6-rc.d/{crond,php-fpm,nginx,$APP_NAME}/dependencies.d/SetupOneshot

Describe alternatives you've considered

N/A

Anything else?

N/A

Am I willing to test this? 🧪

• I will do my best to test this feature on the netlertx-dev image when requested within 48h and report bugs to help deliver a great user experience for everyone and not to break existing installations.

Can I help implement this? 👩‍💻👨‍💻

• Yes
• No

[jokob-sk]

Thanks @vladaurosh for the ongoing help 🙏 just pushed these changes into -dev in the above commit

[jokob-sk]

Had to roll back the cron s6 implementation approach as it didn't trigger the scripts it was triggering previously - you can test this by trying to restart the app in Maintenance -> Logs -> Restart Server

[vladaurosh]

Ok, what script needs to be triggered and what does "Maintenance -> Logs -> Restart Server" restart? The python application?

[jokob-sk]

This is the corontab template: https://github.com/jokob-sk/NetAlertX/blob/main/install/crontab

executing this script every minute

/app/back/cron_script.sh:

https://github.com/jokob-sk/NetAlertX/blob/main/back/cron_script.sh

[jokob-sk]

I might add more cron dependent scripts in future so that's why its a bit over engineered.

[vladaurosh]

Ok, so cron runs cron_script.sh every minute and that script checks if "cron_restart_backend" is present in the log file. If so it restarts the application?

Btw, python app is supervised by s6, so if it is killed, s6 will start it immediately.
pkill -f "python " && (python ${INSTALL_DIR}/server > /dev/null 2>&1 &) && echo 'done'
So no need to run it again after pkill.

And what would be the reason to have "cron_restart_backend" in log file?

[jokob-sk]

Ok, so cron runs cron_script.sh every minute and that script checks if "cron_restart_backend" is present in the log file. If so it restarts the application?

Yes

Btw, python app is supervised by s6, so if it is killed, s6 will start it immediately.
pkill -f "python " && (python ${INSTALL_DIR}/server > /dev/null 2>&1 &) && echo 'done'
So no need to run it again after pkill.

• oh, great will remove it with the next commit

It's mostly used during development to load up the new version of the code or when a script hangs.

[vladaurosh]

I'll try in next couple of days to see why cron doesn't work properly for you when started via s6.

[jokob-sk]

Thanks a lot. I don't feel as confident with my knowledge with the whole s6 setup (and file permissions as you can tell) , so appreciate the help.

…

On Fri, 10 Jan 2025, 09:29 vladaurosh, ***@***.***> wrote: I'll try in next couple of days to see why cron doesn't work properly for you when started via s6. — Reply to this email directly, view it on GitHub <#945 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AW5URDEHUETIOVYU6Z6O6MT2J3Z33AVCNFSM6AAAAABU2ZACKOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDKOBRGM3DKNZWGY> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[vladaurosh]

@jokob-sk
PR for this coming in next couple of hours.
That will fix #965

[jokob-sk]

Thanks a lot. I don't even know when the error manifests, all my builds work

…

On Mon, 20 Jan 2025, 04:50 vladaurosh, ***@***.***> wrote: @jokob-sk <https://github.com/jokob-sk> PR for this coming in next couple of hours. That will fix #965 <#965> — Reply to this email directly, view it on GitHub <#945 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AW5URDECEFNWUONCTYOAAQT2LPQVRAVCNFSM6AAAAABU2ZACKOVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDMMBQHE2TSMBUG4> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[vladaurosh]

Ne errors in the image build, or when container starts. It seems to be the warning thrown by dcron, every second.
Looks like it happens when crond is not started as pid 1 (which is the case here).

[jokob-sk]

I assume this is fixed in the dev image - adding the appropriate label - thanks for the help @vladaurosh 🙏