Unified cgroup v2 layout Upgrade warning PVE 6.4 to 7.0

yswery

Well-Known Member
May 6, 2018
78
5
48
54
Before we upgrade I just want to better understand what the following error means and what it could result in after the upgrade:

WARN: Found at least one CT (174) which does not support running in a unified cgroup v2 layout.
Either upgrade the Container distro or set systemd.unified_cgroup_hierarchy=0 in the Proxmox VE hosts' kernel cmdline! Skipping further CT compat checks.

now for context, I dont see anything special in the config for CT 174:


Code:
$ pct config 174

arch: amd64
cores: 2
description: FooBar
features: nesting=1
hostname: z174.foo
memory: 1024
net0: name=eth0,bridge=vmbr0,gw=1.2.3.4,hwaddr=72:B0:22:67:9D:62,ip=3.4.5.6/32,type=veth
onboot: 1
ostype: centos
rootfs: local-lvm:vm-174-disk-0,size=50G
swap: 0


and when I try to search for more global LXC settings, I dont get any cgroup mentions of the terms lxc.cgroup:

Code:
$ grep -r "lxc.cgroup" /etc

So what does this warning mean and how does the pve6to7 tests detect which contains are problematic?

Thanks!
 
is this a centos 7 container?
 
please read the linked documentations - it contains all the info on how to work around the issue for such legacy guests!
 
  • Like
Reactions: Stoiko Ivanov
is this a centos 7 container?
Yes and also happens with Ubunutu 16.X too. Just read the docs and I understand what is happen.

I am curious how the pve6to7 tests detects this since the pct conf doesnt state which version of centos or ubunutu they are
 
I am curious how the pve6to7 tests detects this since the pct conf doesnt state which version of centos or ubunutu they are
The code is around here:

https://git.proxmox.com/?p=pve-cont...f82747ea558d0398919414945b133dc0;hb=HEAD#l523

put shortly - it mounts the container (or enters it if running) and checks if /sbin/init looks like systemd - if not we assume the container will run, if yes it checks the systemd-version (through the existence of libsystemd-shared-<systemdversion>.so)

I hope this explains it
 
  • Like
Reactions: Phyrene and yswery
Code:
GRUB_CMDLINE_LINUX_DEFAULT="systemd.unified_cgroup_hierarchy=0 quiet"

- Will LXC (centos7) be able to boot successfully after upgraded to PVE7 without modifying grub CMDLINE?
If yes, what will happen on this LXC? Its functions still working as before or any function will be lost/strict?

Or modifying grub is the only way to keep it booting?
 
How can I fix that after I upgraded to PVE 7.0? Can I add the boot paramter to my Ubuntu 16.10 container without the container running?
 
So I edited, /etc/default/grub

I changed this line from,
GRUB_CMDLINE_LINUX_DEFAULT="quiet"
to
GRUB_CMDLINE_LINUX_DEFAULT="systemd.unified_cgroup_hierarchy=0 quiet"

I then ran, update-grub. And rebooted the server. However, I am still unable to start CENTOS7 LXC.
 

Attachments

  • Screenshot_2.png
    Screenshot_2.png
    9 KB · Views: 134
So I edited, /etc/default/grub

I changed this line from,
GRUB_CMDLINE_LINUX_DEFAULT="quiet"
to
GRUB_CMDLINE_LINUX_DEFAULT="systemd.unified_cgroup_hierarchy=0 quiet"

I then ran, update-grub. And rebooted the server. However, I am still unable to start CENTOS7 LXC.
Try set/add the following to your LXC conf (/etc/pve/lxc/1234.conf for example)


Code:
unprivileged: 1


and then try to boot the CT
 

About

The Proxmox community has been around for many years and offers help and support for Proxmox VE, Proxmox Backup Server, and Proxmox Mail Gateway.
We think our community is one of the best thanks to people like you!

Get your subscription!

The Proxmox team works very hard to make sure you are running the best software and getting stable updates and security enhancements, as well as quick enterprise support. Tens of thousands of happy customers have a Proxmox subscription. Get yours easily in our online shop.

Buy now!