Raspberry Pi crashes with 12 container running (24H)

General
1

Every day my RaspberryPi crashes DHCP Failed and sigsev :frowning:

typically im running 12 container on a raspberry pi (10 containers is OK…)

logs :
Jun 22 18:53:11 poseidon systemd[1]: dhcpcd.service: Main process exited, code=killed, status=11/SEGV

Jun 22 18:53:12 poseidon dockerd[558]: time=“2021-06-22T18:53:12.164382075+01:00” level=info msg=“shim docker-containerd-shim started” address=/containerd-shim/moby/8a9a5a98cc7fc4742d804474f9dd0849af2ca37d37d1211dfda245bd3fbadc70/shim.sock debug=false pid=1820

Jun 22 18:53:12 poseidon systemd[1]: dhcpcd.service: Failed with result ‘signal’.

Today I’m trying this workaround :

I add deny-interfaces veth* to dhcp.conf

I have this bug since 1 year and I don’t have a solution right now…

Anyone as a similar bug ?

2

Just out of curriousity: are you setting ressource contraints on your containers?

I wouldn’t be surprised if a system with limited resources - like a pi - would suffer from OOM Kills.
You might want to check system logs for OOM Kills - the kernel will free up as much ram as it needs to breath and kill whatever it finds fit.

Setting ressource constraints is the only way to limit single containers from eating up to much ram. You should consider to apply them on your container.

3

It’s worked !
just add at the end of /etc/dhcpcd.conf the “denyinterfaces veth*”
$sudo echo “denyinterfaces veth*” >> /etc/dhcpcd.conf

No i dont setting ressources constraints but i have a lot of raspberry pi4B.
I don’t see any killed apps in /var/logs event for the pi who used a lot of ressources like big api in springboot.