Installing Home Assistant Supervised using Debian 12

Are you saying this because of the desktop on debian?

Because of this and previous comments. Perhaps I’m misunderstanding.

You’re both right. Op should probably use the other guide. But nevertheless this is a DNS issue. Fix that or nothing works

I’ve seen this before, and it turned out to be a IP6 dns issue, so maybe try again using IP4 only?
:thinking:

Thx all. A couple things. I appreciate the comment about dual booting not affecting anything since when booted to Debian, it can’t see the other partitions. I want to use supervised for several reasons. I’m trying to relearn skills I forgot in the 80s so I greatly appreciate this community!!

So, overnight I DID wipe my entire hard drive yet again and then used the latest Debian installer, selecting the non GUI version. I partitioned the drives successfully and got through the debian issue by using this command: sudo apt install docker.io. That doesn’t need DNS since it was already in the repository.

But several of you are correct that is was a DNS issue. Now the key command to Home Assistant · GitHub … failed with a similar message about unable to resolve host. I have been trying to find a link for fixing DNS issues in plain language. LOL! I’ll keep searching, but if someone can help, I would apprciate it!

apt does use dns because it downloads from a repository.

Actually, i had the same (even in this post :smiley:)

Anyway, i found supervised too strickt, i couldn’t run some other stuff i needed (as the supervisor would uninstall it again)
Still decided to stick to Debian (with gui) and use VMWare workstation (as i was on W10 before)
However I did move some add-on’s to docker on the host ( and who knows, maybe some day i might abandon supervised )

Find your active interface:

sudo nmcli con show --active

The output shows something similar to:

NAME             UUID                                  TYPE      DEVICE 
Supervisor eno1  e874k675-44d4-31a4-8cf8-3b8cc5f3b7cb  ethernet  eno1   
lo               1e5bf4e3-e5c6-40d4-a193-0ffaa1l4fd2b  loopback  lo

where

e874k675-44d4-31a4-8cf8-3b8cc5f3b7cb

is the active interface of my RPI4 used by Supervisor.

Now run the following command but use the UUID of your active interface:

sudo nmcli con mod "e874k675-44d4-31a4-8cf8-3b8cc5f3b7cb" ipv4.dns "8.8.8.8, 8.8.4.4"

With the above command you configure the RPI to resolve TLD’s using Google’s DNS servers (8.8.8.8 and 8.8.4.4). Of course you can change this to whatever valid DNS-servers of your choice.

Restart the NetworkManager service:

sudo systemctl restart NetworkManager
3 Likes

Hi everyone!

A clean supervised installation of HA on a new installed Debian 12 (bookworm) with LXDE. No other applications were installed. The installation was done accordingly to this guide.

First of all, the host with HA in unreachable.

The second issue:
journalctl -f shows that installation is stucked at “Get “https://ghcr.io/v2/””

Here is the log:

Apr 07 12:58:25 debian-ha systemd-resolved[4322]: Using degraded feature set TCP instead of UDP for DNS server 1.1.1.1.
Apr 07 12:58:28 debian-ha systemd-resolved[4322]: Using degraded feature set UDP instead of TCP for DNS server 8.8.8.8.
Apr 07 12:58:31 debian-ha systemd-resolved[4322]: Using degraded feature set TCP instead of UDP for DNS server 8.8.8.8.
Apr 07 12:58:34 debian-ha systemd-resolved[4322]: Using degraded feature set UDP instead of TCP for DNS server 1.0.0.1.
Apr 07 12:58:40 debian-ha dockerd[2219]: time="2024-04-07T12:58:40.570032578+03:00" level=warning msg="Error getting v2 registry: Get \"https://ghcr.io/v2/\": dial tcp: lookup ghcr.io: no such host" spanID=752c9dc820d58d66 traceID=ddb6f20b8123eb80b3f73c77e4e1b7ce
Apr 07 12:58:40 debian-ha dockerd[2219]: time="2024-04-07T12:58:40.570210901+03:00" level=info msg="Attempting next endpoint for pull after error: Get \"https://ghcr.io/v2/\": dial tcp: lookup ghcr.io: no such host" spanID=752c9dc820d58d66 traceID=ddb6f20b8123eb80b3f73c77e4e1b7ce
Apr 07 12:58:40 debian-ha dockerd[2219]: time="2024-04-07T12:58:40.712409403+03:00" level=error msg="Handler for POST /v1.45/images/create returned error: Get \"https://ghcr.io/v2/\": dial tcp: lookup ghcr.io: no such host" spanID=752c9dc820d58d66 traceID=ddb6f20b8123eb80b3f73c77e4e1b7ce
Apr 07 12:58:40 debian-ha hassio_supervisor[2219]: 2024-04-07 12:58:40.714 ERROR (MainThread) [supervisor.docker.interface] Can't install ghcr.io/home-assistant/amd64-hassio-cli:2024.03.1: 500 Server Error for http+docker://localhost/v1.45/images/create?tag=2024.03.1&fromImage=ghcr.io%2Fhome-assistant%2Famd64-hassio-cli&platform=linux%2Famd64: Internal Server Error ("Get "https://ghcr.io/v2/": dial tcp: lookup ghcr.io: no such host")
Apr 07 12:58:40 debian-ha hassio_supervisor[2219]: 2024-04-07 12:58:40.715 WARNING (MainThread) [supervisor.plugins.cli] Error on install cli plugin. Retrying in 30sec
Apr 07 12:59:10 debian-ha hassio_supervisor[2219]: 2024-04-07 12:59:10.746 INFO (MainThread) [supervisor.docker.interface] Downloading docker image ghcr.io/home-assistant/amd64-hassio-cli with tag 2024.03.1.

Please give me any suggestions on this error.

Thank you in advance

DNS issues also?

Thank you for your comment.

It’s kinda network issue, I think

root@debian-ha:~# ha resolution info
Error response from daemon: No such container: hassio_cli

I think because the HA is not fully installed

The solution might be just 3 posts up.

1 Like

In my opinion, DNS errors are better solved by:
edit /etc/systemd/resolved.conf, add DNS addresses (for example: DNS=8.8.8.8 1.1.1.1) and systemctl restart systemd-resolved.
I had to edit it twice - once when installing the necessary packages and twice after installing HA itself.
A much better idea than editing /etc/resolv.conf which gets overwritten.

Right before and at the docker installation you will most likely run into the “Temporary failure in name resolution” message, what most call DNS issues.

This is due to the previous apt command having installed, configured and started the network-manager which is now replacing systemd-resolved.

The correct way to solve the issue is Not to poke around the resolve.conf file nor resolvectl nor restart systemd-resolved.
If you did reboot your system first.

Instead run the NetworkManaget tool:

# nmcli d

which will return something like:

DEVICE          TYPE      STATE                   CONNECTION        
enp4s0          ethernet  connected               Supervisor enp4s0 
lo              loopback  connected (externally)  lo                
docker0         bridge    unmanaged               --                
hassio          bridge    unmanaged               --                
veth3ed17d5     ethernet  unmanaged               --                
veth3ff491e     ethernet  unmanaged               --                

Now look for the connection name and run:

# nmcli conn modify "Supervisor enp4s0" ipv4.dns 1.1.1.1
# systemctl restart NetworkManager

I hope this saves someone the time I wasted figuring this out. :wink:

1 Like

Or more detailed instructions 6 posts up HERE.

2 Likes

Just for the record, I’ve been running this installation as instructed for more than 3 years now. No major issues to report other than some recent issues with network connectivity. I’m running it out of an old laptop. Everything works with Apple Home too. Brilliant.
Cheers!

2 Likes

Hi,

on dns config: just my 2cents, hopefully it can help someone:

I also experienced the dns troubles.
My set-up: a fresh minimal install of debian 12 (no gui/desktop).
Typically I set a static ip via /etc/network/interfaces.

However as soon I installed the dependencies for HA (especially network-manager and systemd-resolved), I got in trouble with DNS.

Configuring dns via network-manager seemed:
a) not to work for some reason (i.e. apt could not resolve the repos but nslookup worked ok)
b) config on network via 2 locations/tools: dns via network-manager and static ip in /etc/network/interfaces

I also noticed that as long /etc/network/interfaces referenced my eth device, network-manager set the device as unmanaged and could not manage it (although it would create a connection).

Much google-ing and reading of posts (also in this thread) I came to following solution:

  1. mv /etc/network/interfaces /etc/network/interface.save (so that network-manager could manage the device)
  2. ensure systemd-networkd is not enabled (SystemdNetworkd - Debian Wiki)
  3. configure the network fully through network-manager
    (e.g. NetworkConfiguration - Debian Wiki)

br,
Ruben

1 Like

Some additional information for those who are struggling with DNS issues:

I had to enable NetWorkmanager to manage my wired network first, by changing

managed=false
to
managed=true

in /etc/NetworkManager/NetworkManager.conf.

After that, do a

nmcli con show --active

which gives you something like:

NAME               UUID                        TYPE      DEVICE
Ifupdown (enp1s0)  <someuid>                   ethernet  enp1s0
lo                 <anotheruid>                loopback  lo

and then do a

nmcli con mod "<someuid>" ipv4.dns <myDNSIP>

and finally

service NetworkManager restart

After that DNS issues should be resolved and you can continue installing docker.

2 Likes

Same for me in Debian 12 Supervised install. Nothing restored the DNS not even disabling IPv6 in grub.

This worked for me:
Uncomment (or append) the line precedence ::ffff:0:0/96 100 in /etc/gai.conf

And them a reboot.

Did not work for me. Not even disabling ipv6 in grub.
Additionly I uncommented the line precedence ::ffff:0:0/96 100 in /etc/gai.conf to priortize ipv4 and rebooted.
That brought back the DNS.