Did anyone realize that after you are running Debian 11 some of the data is showing 0% on the container? Both Core RAM Usage, Supervisor CPU Usage, and Supervisor RAM Usage is basically 0%?
Possibly the same issue you noticed on docker monitor custom component?
CPU usage works for me though.
Glances can see memory though.
Its because Debian 11 kernel implementing cgroup v2 that cause this problem. But dont have any clue on fixing it.
BTW portainer is also experience the same issue
Have you tried the Debian 11 release notes? There’s specific guidance if you require cgroup v1 functionality.
Chapter 2. What's new in Debian 11 (amd64 but same notes are under arm64 as well)
I added systemd.unified_cgroup_hierarchy=false and systemd.legacy_systemd_cgroup_controller=false to my kernel command line and all working fine after upgrading to Bullseye. At least to give the supervisor and addon teams time to adapt to cgroups v2.
3 Likes
Aha! Another problem solved by reading the instructions. Wonders never cease. Well done.
What do you need to do to implement does command line? Can you help me?
On a Pi it is /boot/cmdline.txt
need one more for proxmox user like myself.
Please try a search engine. I searched for “proxmox kernel command line” and this was the first result.
2 Likes
tested and doesn’t seem to work.
@speedst3r
Thank you for this and the link on how to change the kernel command line.
I have updated my system and restarted Proxmox and also HA (and rechecked that I have added correctly).
I am also still not seeing any ram usage information in Supervisor>system panel.
Not posting with any expectation that you will resolve. I thought I would just post my experience for others.
I’m guessing your HA isn’t installed directly into Proxmox; but rather into a virtual machine. You’re going to have to look into how whatever operating system runs on that VM manages its kernel command line.
1 Like
I run the right command coming from this pages
https://github.com/systemd/systemd/issues/13477
However nothing changes, so I guess this does not work.
The debian warning seems to allow some aplication on openstack to be working on Debian 11 but does not replace cgroup v2 so our apps need to adapt in order to get the informations.
Monitor Docker already running perfectly with minor modification on the code
Thank you for this.
Removed the commands from Proxmox and added them into the HA VM Grub file and all information now available.
Cheers
I’ve got the variables and will install the OS Agent shortly, but I’ll wait a month or two before updating to Debian 11, but this thread will be useful if I face any issues.
Good work guys, posting issues and solutions!
So, i’ve just updated the Supervisor to today’s release, on Debian 11. The unsupported message is gone. Nice!
Now I only have some warnings:
21-09-01 18:26:05 WARNING (MainThread) [supervisor.addons.validate] Add-on config 'snapshot_post' is deprecated, 'backup_post' should be used instead. Please report this to the maintainer of MariaDB
21-09-01 18:26:05 WARNING (MainThread) [supervisor.addons.validate] Add-on config 'snapshot_pre' is deprecated, 'backup_pre' should be used instead. Please report this to the maintainer of MariaDB
21-09-01 18:26:05 WARNING (MainThread) [supervisor.addons.validate] Add-on config 'snapshot_exclude' is deprecated, 'backup_exclude' should be used instead. Please report this to the maintainer of deCONZ
21-09-01 18:26:06 WARNING (MainThread) [supervisor.addons.validate] Add-on config 'snapshot' is deprecated, 'backup' should be used instead. Please report this to the maintainer of UniFi Controller
21-09-01 18:26:06 WARNING (MainThread) [supervisor.addons.validate] Add-on config 'snapshot_exclude' is deprecated, 'backup_exclude' should be used instead. Please report this to the maintainer of ESPHome
21-09-01 18:26:06 WARNING (MainThread) [supervisor.addons.validate] Add-on config 'snapshot_exclude' is deprecated, 'backup_exclude' should be used instead. Please report this to the maintainer of AdGuard Home
Well, I don’t use any of these addons. So, I think, this will sort it self out with future update’s of HA, em I right?
1 Like
Those will sort themselves out when the maintainers of those add-ons update their configuration
1 Like
Do I need to be worried about this update as I’m running Supervisor in Docker on a raspberrey pi 3+
I know this is an unsupported system, but till now it was working
Same question as @deluxestyle (slightly different configuration). Curious (not being a power user) if my installation method is being sunset and I’m now facing an extinction event (for my current configuration).
FWIW, I used this install method quite awhile ago (from JuanMTech): Home Assistant on Docker and Ubuntu - Setup Guide - YouTube
1 Like
FWIW, I just finished installing os-agent on two systems, both running Home Assistant Supervised on Debian 10 except one is ARM architecture (RPi3B+) and the other Intel (laptop).
- The RPI has server version of Debian. udisks2 was not present so had to install that first.
- The laptop has desktop version of Debian. udisks2 was present.
Ran gdbus example (shown in the documentation) and it completed successfully on both systems.
Updated to latest version of Supervisor (2021.08.01) and so far everyone appears to be running smoothly on both systems. 
No plans to upgrade to Debian 11 just yet.
NOTE
I also see the warnings about Add-ons using deprecated terminology despite the fact none are actually installed (on either system). Will ignore for now as instructed.