[SOLVED] LXC Gentoo didn't stop right over Proxmox

[fireon]

Hello,

i have here one LXC Gentoo that didn't stop right after shutdown from the host. If look at the logs in the container no shutdown event was triggered. After the timeout the CT stops. Error Messeage in PVE:

lxc-stop: 124: commands_utils.c: lxc_cmd_sock_rcv_state: 70 Resource temporarily unavailable - Failed to receive message
TASK OK

pve-manager/5.4-3/0a6eaa62 (running kernel: 4.15.18-12-pve)
Here are the config:

arch: amd64
cores: 4
cpuunits: 512
description: Gentoo
hostname: mon.tux.local
memory: 2048
net0: name=eth0,bridge=vmbr0,hwaddr=46:05:FE:83:45:AC,ip=dhcp,ip6=auto,type=veth
onboot: 1
ostype: gentoo
rootfs: local-zfs:subvol-124-disk-0,acl=1,size=45G
swap: 512

What have to be in the CT that a normal shutdown from the host is possible?

 

[mira]

Looks like the signal sent by lxc-stop is not handled correctly. There was this issue back in 2017: https://github.com/lxc/lxc/issues/1928
Is your container up-to-date? Which init system are you using?

 

[mira]

I could not reproduce it here. Neither with gentoo-current-default_20180906_amd64.tar.xz nor after updating it. Could you provide some more information? What's running in there? Any modifications to the init system?
And does it happen on a new container as well?

 

[fireon]

Hello,

my CT build by by an earlier CT-Template. Gentoo run's on openrc. With an new CT it works. My one worked also, but i think after some Update inside the CT the issue was generated. I don't know when, because i realized that to late. The Gentoo is up to date and it runs with vdr and an dvbcard. To disable all the specials did't help.

 

[wbumiller]

Lxc usually sends either a SIGPWR (30) or SIGRTMIN+3 (35), you could see if the container's init (pid 1) reacts to either of those.
It would be weird though if something from inside gentoo changed that behavior of the init system...

 

[fireon]

Strange... then i don't understand it, but i check the pid.

 

[fireon]

Works now in 6.0