php-fpm[12033]: [WARNING] [pool www] child 24115 exited on signal 9 (SIGKILL) after 26.911869 seconds from start

i get ton of these messages daily. it seems that stopping any random docker fixes it. why is this happening?

diag.zip

its a pretty frustrating thing to have the webui unresponsive because of this. its worth noting that while im unsure what the cause is, i just decided to select the pairs for all the dockers to use only pairs 2,3,4 and only things like authelia, plex, and reverse proxy to use all. running almost all the usual apps and even some that i use occasionally is giving me just a couple of those lines but nothing compare to a wall of warnings like before. 

Try booting in safe mode and/or closing any browser windows open to the GUI, only open when you need to use it then close again.

i cant really try safemode because i dont have an extra monitor and my monitor is mounted. i tried closing all my connections and use the webui as minimal as i can but after like 24hrs i have no luck. when i start getting these messages, i have to ssh into the server and start turning off dockers. when i stop like 2-3 dockers, it becomes stable again. im pretty sure its docker related. i just learned the command to turn off the docker service so next time ill try that but when i turned off the docker service in the settings, it was stable again too. im now trying to leave most of the dockers off to run as little as possible to see what happens.

 

by docker related, that includes maybe that something docker related is using up all the cpu or something. 

Edited November 11, 20232 yr by droopie

On 11/10/2023 at 3:24 AM, JorgeB said:

closing any browser windows open to the GUI

so this stuck with me. i was monitoring on my phone via ssh the unraid logs and would see my phone say it was logging in via webui but i had no windows open. so after changing the password, and not logging in to the webui in days, i didnt see the warnings anymore. today i logged in the webui and monitored the log via, after an hour, i started seeing the warnings again. signed out of the webui and the warnings stopped. so it might be being logged into the webui for long period of time is causing this. im not looking into a dashboard/homepage that would let me start/stop dockers from a website and use iframes to open dockers webui in an iframe. organizr does great with iframe tabs but i cant control dockers to start/stop them so looking for alternatives. trying to minimize the logging into unraids webui.

Same problem here on the latest unraid, started getting these errors lately all over

 

Also sometime when I press the system log icon in dashboard, I get 502 error, nginx bad gateway. I guess these errors are related.

 

Did not have this issue in 11.5 before updating to 12.6

3 hours ago, zaelnorth said:

Same problem here on the latest unraid, started getting these errors lately all over

 

Also sometime when I press the system log icon in dashboard, I get 502 error, nginx bad gateway. I guess these errors are related.

 

Did not have this issue in 11.5 before updating to 12.6

i get that error 502 too when i open a docker webui. i was never able to find a fix. as soon as i restart ANY docker, everything gets updated to the latest running state like it hangs in the background. the cpu is also 100% during this time. i have tried running the absolute bare i can do not require any uptime like only rss stuff (radarr, sonarr, prowlarr, qbit, sabnzbd exactly) and not even running plex to keep it from doing this. i also was told by others to put all these apps to use only the last cores so in my case pair 4 but the cpu is still 100% for some reason.

by any chance, are you running qbit? if so, how many torrents do you have if you dont mind me asking. 

ive also tried setting up a userscript to restart smaller apps like autoscan, and bazarr in hopes to randomly free up the cpu but i dont see it working.

 

when i do a restart of any docker, my cpu is usually around 30% and everything works perfect so im not sure whats causing it to reach 100% and i strongly think its docker related.

Edited December 21, 20232 yr by droopie

https://forums.unraid.net/topic/86114-nginx-running-out-of-shared-memory/?do=findComment&comment=1417198