HA loses network when router changed

[hawkeye100]

Describe the issue you are experiencing

Changed router on LAN (so DHCP changed). HA (on x86 PC) acquired an IP address. Appeared to reboot OK - BUT could not be seen on network by other devices, nor reconnect to external data sources e.g. for backup or media. Showed CIFS error but no network error.

Swapped cables, switch etc - no change.
Found this previous report from 2022 and solution which eventually worked, but looks like it has not been addressed so bringing it to attention again (as that thread is now closed)

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?

6.6.20

Did the problem occur after upgrading the Operating System?

No

Hardware details

No USB devices.8GB RAM. 300GB hard disc. CPU celeron

Steps to reproduce the issue

1. remove router (and DHCP)
2. replace router and DHCP

...NOTE have not repeated as there is too much work involved in sorting the IP addresses for all other devices - but it is confirmed in the link to previous issue in the description

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

-

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

-

System information

System Information

version	core-2024.8.1
installation_type	Home Assistant OS
dev	false
hassio	true
docker	true
user	root
virtualenv	false
python_version	3.12.4
os_name	Linux
os_version	6.6.20-haos
arch	x86_64
timezone	Europe/London
config_dir	/config

Home Assistant Community Store

GitHub API	ok
GitHub Content	ok
GitHub Web	ok
GitHub API Calls Remaining	4861
Installed Version	1.34.0
Stage	running
Available Repositories	1383
Downloaded Repositories	26

Home Assistant Cloud

logged_in	false
can_reach_cert_server	ok
can_reach_cloud_auth	ok
can_reach_cloud	ok

Home Assistant Supervisor

host_os	Home Assistant OS 12.1
update_channel	stable
supervisor_version	supervisor-2024.08.0
agent_version	1.6.0
docker_version	24.0.7
disk_total	292.7 GB
disk_used	10.3 GB
healthy	true
supported	true
host_connectivity	true
supervisor_connectivity	true
ntp_synchronized	true
virtualization
board	generic-x86-64
supervisor_api	ok
version_api	ok
installed_addons	File editor (5.8.0), Samba share (12.3.2), Mosquitto broker (6.4.1), Advanced SSH & Web Terminal (18.0.0), SQLite Web (4.2.0), Log Viewer (0.17.0), Syncthing (1.19.0)

Dashboards

dashboards	6
resources	19
views	19
mode	storage

Recorder

oldest_recorder_run	22 August 2024 at 16:01
current_recorder_run	28 August 2024 at 10:30
estimated_db_size	652.61 MiB
database_engine	sqlite
database_version	3.45.3

Additional information

Workaround/solution is:
ha network update eno1 --ipv4-method auto
from HA server CLI

Desired solution is to avoid this happening if network/DHCP becomes unavailable, though a minimum might be a warning on reboot (similar to the CIFS warnings) to at least point in the right direction.
Thanks

[Anto79-ops]

I wanted to mention something similar happened to me with my NUC 12 running HAOS 13.1. There was a bios FW update for my NUC (oringinal dated for sometime in in 2022, and the new FW dated June 2024). I shutdown the system and moved my NUC to a computer screen via HDMI so that I can update the bios via USB, it updated sucessfully, and moved it back to its permanent (headless) spot. When I rebooted it, HA did not come back, nor could I connect to it. When I connected a monitor to it again and logged in via CLI/HDMI, everything was working, core started up but it was complaining there was no netwrok access.

Long story shory, I did a backup and restore and seems the network adapter changed from enspXX something to enspXX something else, not sure if this was because of a FW update or some issue in HAOS.

[agners]

Swapped cables, switch etc - no change.

Just to be clear, you did also try to restart your Home Assistant instance correct?

The issue you linked has been resolved with home-assistant/supervisor#3676, so this really should not happen anylonger. Maybe your change somehow still made it happen again 🤔

[agners]

I wanted to mention something similar happened to me with my NUC 12 running HAOS 13.1. There was a bios FW update for my NUC (oringinal dated for sometime in in 2022, and the new FW dated June 2024). I shutdown the system and moved my NUC to a computer screen via HDMI so that I can update the bios via USB, it updated sucessfully, and moved it back to its permanent (headless) spot. When I rebooted it, HA did not come back, nor could I connect to it. When I connected a monitor to it again and logged in via CLI/HDMI, everything was working, core started up but it was complaining there was no netwrok access.

The network settings are strictly bound to the network interface. On very initial start, we setup the first network interface (as determined by NetworkManager) as primary interface and with DHCP. From then onwards, we assume that this interface is always "there" (meaning, at the same udev path, so essentially just there, at the same place if it was PCIe). So this means if anything changes in the way the NIC is detected by the OS (e.g. I can imagine that a BIOS update changes the order of enumeration of PCI devices/buses), the system will consider this as a new interface and initially not touch it.

What Supervisor probably should do in the specific case where the previous primary interface disappears and a new primary interface appears, just simply migrate the configuration. But this isn't implemented yet at this point.

That said, I think OP did not change anything on the HA side, so I don't think that this is what happened here.

[hawkeye100]

Swapped cables, switch etc - no change.

Just to be clear, you did also try to restart your Home Assistant instance correct?

The issue you linked has been resolved with home-assistant/supervisor#3676, so this really should not happen anylonger. Maybe your change somehow still made it happen again 🤔

Thanks for comment. It is now a while ago but as far as I recall ....

Initially no HA reboot, then a reboot to see if it fixed it - but it did not. Hence frustration at an obscure issue given everything else 'appeared' to be ok. The only clue was the CIFS error, but the network share resource was confirmed as available suggesting a network issue on the HA box. Superficial investigation suggested the network interface was OK (IP address etc).

No change on HA itself - correct.

supervisor#3676 looks like it should have fixed it (and seems to be a similar description of the issue), but may not have done perhaps given my experience was more recent ?

Your suggestion of migrating the settings might work though I do not profess to understand the more intricate workings of HA and Linux interfaces.

My request is to try and identify the issue at a higher level (at HA prompt) or even fix/bypass it automatically so we do not have to SSH to the HA machine and read/change underlying Linux configs - especially if we have no idea where to look. Progress elsewhere in HA has been significant in bringing more technical things into the HA level of interface or even the GUI though the latter would not work for this presumably.