Cannot manage vlan connections with nmcli due to ipv4.dns-data member

[Particpant]

Describe the issue you are experiencing

If I try to modify any of my vlan connections, say to rename them, or to mark as 'never-default', I run into an error from nmcli:

nmcli con mod <uuid> ipv4.never-default 'yes'
Error: Failed to modify connection 'Supervisor enp3s0.8': ipv4.dns-data: unknown property

nmcli con mod <uuid> ipv4.dns-data ''
Error: invalid property 'dns-data': 'dns-data' not among [method, dns, dns-search, dns-options, dns-priority, addresses, gateway, routes, route-metric, route-table, routing-rules, replace-local-rule, ignore-auto-routes, ignore-auto-dns, dhcp-client-id, dhcp-iaid, dhcp-timeout, dhcp-send-hostname, dhcp-hostname, dhcp-fqdn, dhcp-hostname-flags, never-default, may-fail, required-timeout, dad-timeout, dhcp-vendor-class-identifier, link-local, dhcp-reject-servers, auto-route-ext-gw].

nmcli similarly refuses to apply modifications via the interactive editor.

What operating system image do you use?

generic-x86-64 (Generic UEFI capable x86-64 systems)

What version of Home Assistant Operating System is installed?

Home Assistant OS 10.3

Did you upgrade the Operating System.

Yes

Steps to reproduce the issue

1. Have vlan connections set up in nmcli (followed tutorial here: https://community.home-assistant.io/t/setup-vlan-and-ha-tutorial/87705/111)
2. Be on current HAOS (unknown if this is newly introduced or longstanding.
3. Attempt to modify any existing connection via nmcli ...

Anything in the Supervisor logs that might be useful for us?

I didn't see anything relevant in these logs

Anything in the Host logs that might be useful for us?

I didn't see anything relevant in these logs

System information

System Information

version	core-2023.7.1
installation_type	Home Assistant OS
dev	false
hassio	true
docker	true
user	root
virtualenv	false
python_version	3.11.4
os_name	Linux
os_version	6.1.34
arch	x86_64
timezone	America/Los_Angeles
config_dir	/config

Home Assistant Community Store

GitHub API	ok
GitHub Content	ok
GitHub Web	ok
GitHub API Calls Remaining	5000
Installed Version	1.32.1
Stage	running
Available Repositories	1272
Downloaded Repositories	3

Home Assistant Cloud

logged_in	true
subscription_expiration	May 18, 2024 at 12:00 AM
relayer_connected	true
relayer_region	us-east-1
remote_enabled	true
remote_connected	true
alexa_enabled	false
google_enabled	false
remote_server	us-east-1-8.ui.nabu.casa
certificate_status	ready
can_reach_cert_server	ok
can_reach_cloud_auth	ok
can_reach_cloud	ok

Home Assistant Supervisor

host_os	Home Assistant OS 10.3
update_channel	stable
supervisor_version	supervisor-2023.07.1
agent_version	1.5.1
docker_version	23.0.6
disk_total	109.3 GB
disk_used	13.5 GB
healthy	true
supported	true
board	generic-x86-64
supervisor_api	ok
version_api	ok
installed_addons	Samba share (10.0.2), File editor (5.6.0), Terminal & SSH (9.7.1), Advanced SSH & Web Terminal (15.0.3), Z-Wave JS UI (1.14.0), ESPHome (2023.6.4), Logitech Media Server (0.0.53)

Dashboards

dashboards	7
resources	0
views	19
mode	storage

Recorder

oldest_recorder_run	June 29, 2023 at 2:06 PM
current_recorder_run	July 9, 2023 at 7:23 PM
estimated_db_size	194.51 MiB
database_engine	sqlite
database_version	3.41.2

Additional information

I've rebooted multiple times (both HA and the hardware itself) to no avail. Additionally, I've seen a similar issue referenced elsewhere, which indicate a possible version mismatch between nmcli and NetworkManager: https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/1307

I wasn't able to run any of the suggested troubleshooting steps, as systemctl and journalctl aren't installed in HAOS:

~ # nmcli --version
nmcli tool, version 1.42.8
~ # systemctl restart NetworkManager.service
-bash: systemctl: command not found
~ # journalctl -b0 -u NetworkManager | grep 'is starting'
-bash: journalctl: command not found

[agners]

Generally, you should not need to use nmcli directly. Instead use the ha network command for your networking needs.

What are you trying to achieve?

It sounds like there is a unsupported property in the connection settings, can you check if that is indeed the case in cat /etc/NetworkManager/system-connections/Supervisor\ enp3s0.8.nmconnection?

[Particpant]

There's no way to assign a 'default' connection when setting up multiple vlans, or set the priority of any one network path over another, except via nmcli, as far as I can tell. A non-desired network is being selected as the default, and I'm trying to override it. FWIW, I'm not aware of any way to set up VLANs in HAOS other than this, either, happy to learn if there's a better/right way.

That command doesn't work, I've got an /etc/NetworkManager/system-connections/Supervisor folder, but there's no Supervisor subfolder.

the output of nmcli -p con show for the connection doesn't have an ipv4.dns-data field.

-------------------------------------------------------------------------------
ipv4.method:                            manual
ipv4.dns:                               <ip address>
ipv4.dns-search:                        --
ipv4.dns-options:                       --
ipv4.dns-priority:                      0
ipv4.addresses:                        <ip address>/24
ipv4.gateway:                           <ip address>
ipv4.routes:                            --
ipv4.route-metric:                      -1
ipv4.route-table:                       0 (unspec)
ipv4.routing-rules:                     --
ipv4.replace-local-rule:                -1 (default)
ipv4.ignore-auto-routes:                no
ipv4.ignore-auto-dns:                   no
ipv4.dhcp-client-id:                    --
ipv4.dhcp-iaid:                         --
ipv4.dhcp-timeout:                      0 (default)
ipv4.dhcp-send-hostname:                yes
ipv4.dhcp-hostname:                     --
ipv4.dhcp-fqdn:                         --
ipv4.dhcp-hostname-flags:               0x0 (none)
ipv4.never-default:                     no
ipv4.may-fail:                          yes
ipv4.required-timeout:                  -1 (default)
ipv4.dad-timeout:                       -1 (default)
ipv4.dhcp-vendor-class-identifier:      --
ipv4.link-local:                        0 (default)
ipv4.dhcp-reject-servers:               --
ipv4.auto-route-ext-gw:                 -1 (default)

[ys-li]

I am having this problem as well. Was trying to set static IPs for multiple interfaces. I think the configuration steps are actually documented in the network.md in this repo. Following the instruction results in the error message mentioned. https://github.com/home-assistant/operating-system/blob/dev/Documentation/network.md#using-nmcli-to-set-a-static-ipv4-address

[dgarciadonado]

I think I have a similar issue. Trying to fully disable IPv6 get me an ipv4.dns-data error:

[dgarciadonado]

@Particpant did you get a solution for this? Thanks!

[Particpant]

No, I have not.

[ryan-haver]

This is the same behavior seen when using the interactive editor....it will let me manually remove DNS and add DNS servers with the add and remove commands. With that said when saving, it validates the properties and rattles off the "ipv4.dns-data: unknown property" error if the DNS property has a value set.

I can, however, remove the DNS entry and save without issues, but now I don't have a DNS server configured....

[MaxVonEvil]

Seeing exactly the same here on OS 10.4 - Thought I might be able to work around this by deleting the dns object and re-adding it. Nope. Same problem. dns-data gets added when you add a static route.. why?

To compound the problem; when re-adding the DNS server with set ipv4.dns 1.1.1.1, it refuses to save as the dns entry is invalid? print ipv4.dns revealed a comma is added in front as ,1.1.1.1 - thought it might be expecting two DNS entries, so I tried set ipv4.dns 1.1.1.1, 1.0.0.1

No dice either. I had to restore a snapshot to clean out the mess I made

[fsiek]

Same here. Always get ipv4.dns-data: unknown property on saving a connection. Is there a config file to edit the IP manually?

[tismofied]

😿 came here was hoping for a fix. I have HA running on baremetal and I am getting the same error

homeassistant: 2023.9.2
hostname: homeassistant
machine: generic-x86-64
operating_system: Home Assistant OS 10.5
supervisor: 2023.09.2
supported: true
supported_arch:
- amd64
- i386

[nberardi]

I am having the same issue on 2023.9.3

[dgarciadonado]

Hi all,

It seems that on this link similar issue has been solved: https://almalinux.discourse.group/t/unable-to-set-dns-with-nmcli-ipv4-dns/2442/4

1. sudo dnf install NetworkManager-initscripts-updown
2. sudo ifdown enp1s0
3. sudo mv /etc/sysconfig/network-scripts/ifcfg-enp1s0 /root
4. sudo systemctl restart NetworkManager
5. edit connection with nmcli

My linux/network knowledge is so poor and barely know what these commands are doing so I am afraid of break something. Maybe someone what more experience knows if these commands are harmless to try.

[supermopf]

I have the same issue, but you can save the config if you remove the ipv4.dns setting:

nmcli connection edit "Your connection"
remove ipv4.dns
save
quit

[github-actions[bot]]

There hasn't been any activity on this issue recently. To keep our backlog manageable we have to clean old issues, as many of them have already been resolved with the latest updates. Please make sure to update to the latest Home Assistant OS version and check if that solves the issue. Let us know if that works for you by adding a comment 👍 This issue has now been marked as stale and will be closed if no further activity occurs. Thank you for your contributions.

[Particpant]

don't close.

[PeterBorfai]

Thanks for the workaround! My problem was ethernet negotiation was turned off around 6th Jan, probably following the OS update. So the ethernet interface defaulted to 100Mbps. This caused my speedtest.net integration to measure only 90+ Mbps whereas my ISP is way faster than 100Mbps. nmcli con edit "Supervisor enp0s31f6" # to select the interface nmcli> set 802-3-ethernet.auto-negotiate on # fixes the problem nmcli> save # failed as ipv4.dns-data unknown property nmcli> remove ipv4.dns # as suggested by @supermopf above nmcli> save # now works nmcli> quit Then I had to host shutdown as the interface had a "no carrier" problem Finally, in HA UI I restored the DNS server address. Now the interface runs at 1Gbps and speedtest measures downloads around 850Mbps

This workaround should not be necessary. Can it not be fixed?

[mpieru]

Hello, same issue here, please don't close the issue.

Verified/reproduced on HaOS 11.1 The other thing I can observe is the following warning message when enter in the NMCLI:

Warning: nmcli (1.44.2) and NetworkManager (1.40.18) versions don't match. Restarting NetworkManager is advised.

I used nmcli on HaOS in the past, but I don't remember this message in previous HaOS versions.

PS: workaround as eplained is ok (after deletion of ipv4.dns is sufficient to re-add DNS in the HomeAssistant user interface (settings->system->network)

[github-actions[bot]]

There hasn't been any activity on this issue recently. To keep our backlog manageable we have to clean old issues, as many of them have already been resolved with the latest updates. Please make sure to update to the latest Home Assistant OS version and check if that solves the issue. Let us know if that works for you by adding a comment 👍 This issue has now been marked as stale and will be closed if no further activity occurs. Thank you for your contributions.

[Particpant]

still present

[github-actions[bot]]

There hasn't been any activity on this issue recently. To keep our backlog manageable we have to clean old issues, as many of them have already been resolved with the latest updates. Please make sure to update to the latest Home Assistant OS version and check if that solves the issue. Let us know if that works for you by adding a comment 👍 This issue has now been marked as stale and will be closed if no further activity occurs. Thank you for your contributions.

[agners]

Unfortunately, that mismatch of nmcli version is not easily fixable as the add-on is Alpine based and nmcli is installed from the Alpine package system. Also, it is possible to run a newer version of the add-on on older versions of HAOS, so it would be possible to cause a missmatch always.

The way to manage network is to use the ha network CLI command. That way configuration is made through Supervisor, and we can handle API changes on NM side gracefully. If features are missing, we should add them to the ha network command.

See also: https://github.com/hassio-addons/addon-ssh/pull/761