Zombie processes within the Docker deployment

[aspeicher95]

Hello,

when running Pandora with the provided docker-compose.yml (Cofigured to our endvironment) - after a recent Update I get Zombie processes.

root        6134  0.0  0.0      0     0 pts/0    Z+   Jul19   0:00 [workers_manager] <defunct>
root        6135  0.0  0.0      0     0 pts/0    Z+   Jul19   0:00 [workers_manager] <defunct>

# pstree -p -s 6134
systemd(1)───containerd-shim(2602)───docker-init(2679)───workers_manager(4820)───workers_manager(6134)
# pstree -p -s 6135
systemd(1)───containerd-shim(2602)───docker-init(2679)───workers_manager(4820)───workers_manager(6135)

I have tried solving this issue using the init: true flag within the docker-compose.yml for the pandora service, without any success. Likely this isn't working out, because these Zombies are grandchildren.

Any adivce on how to fix this issue? Can I provied additional information?

Thank you and regards
Andreas

[Rafiot]

I do not have the slightest idea. Debugging docker is a nightmare, I don't personally use it.

I'd recommend to shutdown the docker image, restart it, and see what happens.

[aspeicher95]

I am not sure, but I think this is not a docker specific issue. I will see if I can confirm.
Will get back to you as soon as I can confirm (or not 😉)

[Rafiot]

It's not impossible the script is somehow stuck (I'd guess on I/O), but you should still be able to kill/restart the docker image.