HAOS 18.3 regression on generic-x86-64: Realtek RTL8168/r8169 NETDEV WATCHDOG and loss of network

Author: yiskangCreated Sep 21, 2026Updated Sep 21, 2026

Describe the issue you are experiencing

After upgrading a bare-metal generic-x86-64 Home Assistant OS installation from HAOS 18.2 to HAOS 18.3, the Ethernet interface becomes unusable after reboot.

HAOS itself boots successfully, but the machine becomes unreachable from the LAN. The local console continuously reports r8169 NETDEV WATCHDOG transmit queue timeouts on enp2s0.

The affected NIC is:

Vendor ID: 0x10ec (Realtek)
Device ID: 0x8168
Subsystem vendor: 0x10ec
Subsystem device: 0x8168
Driver: r8169
Interface: enp2s0
PCI address: 0000:02:00.0

Rolling back to HAOS 18.2 using os boot-slot other immediately restores Ethernet connectivity.

HAOS 18.2 works normally on the exact same hardware and configuration.

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?

18.3

Did the problem occur after upgrading the Operating System?

Yes

Hardware details

Bare-metal x86-64 installation.

Ethernet controller:

PCI address: 0000:02:00.0
Vendor ID: 0x10ec (Realtek)
Device ID: 0x8168
Subsystem vendor: 0x10ec
Subsystem device: 0x8168
Kernel driver: r8169
Interface: enp2s0

HAOS 18.2:

Kernel: 6.18.39-haos
Ethernet works normally

HAOS 18.3:

System boots, but Ethernet becomes unusable
Repeated NETDEV WATCHDOG transmit queue timeout messages

Steps to reproduce the issue

Start with HAOS 18.2 on a generic-x86-64 bare-metal system using a Realtek 10ec:8168 NIC with the r8169 driver.
Verify Ethernet connectivity works normally.
Upgrade Home Assistant OS from 18.2 to 18.3.
Reboot into HAOS 18.3.
Observe that HAOS boots, but the host becomes unreachable from the LAN.
Check the local console and observe repeated r8169 NETDEV WATCHDOG / transmit queue timeout messages.
From the local Home Assistant CLI, run os boot-slot other.
The system boots back into HAOS 18.2.
Ethernet connectivity immediately works normally again without any configuration changes.

I reproduced the regression by upgrading to 18.3 and confirmed that it disappears when booting the previous 18.2 slot.

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

txt

The Supervisor did not report an explicit OS update error.

The update was fetched normally:

2026-09-21 17:54:28.088 INFO (MainThread) [supervisor.os.manager] Fetch OTA update from https://os-artifacts.home-assistant.io/18.3/haos_generic-x86-64-18.3.raucb

After reboot, HAOS 18.3 successfully booted, but the network interface became unusable. The relevant errors appeared on the local host/kernel console rather than in the Supervisor log.

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

txt

These messages repeat continuously while running HAOS 18.3.

r8169 0000:02:00.0 enp2s0: NETDEV WATCHDOG: CPU: 1: transmit queue 0 timed out 5096 ms
r8169 0000:02:00.0 enp2s0: rtl_rxtx_empty_cond == 0 (loop: 42, delay: 100)
r8169 0000:02:00.0 enp2s0: NETDEV WATCHDOG: CPU: 1: transmit queue 0 timed out 6124 ms
r8169 0000:02:00.0 enp2s0: rtl_rxtx_empty_cond == 0 (loop: 42, delay: 100)

System information

NOTE: The system is currently rolled back to HAOS 18.2 because HAOS 18.3 makes the Ethernet interface unusable. HAOS 18.3 remains installed in the inactive boot slot.

System Information

version	core-2026.9.3
installation_type	Home Assistant OS
dev	false
hassio	true
docker	true
container_arch	amd64
user	root
virtualenv	false
python_version	3.14.6
os_name	Linux
os_version	6.18.39-haos
arch	x86_64
timezone	Asia/Taipei
config_dir	/config

Home Assistant Cloud

logged_in	true
subscription_expiration	September 30, 2026 at 8:00 AM
relayer_connected	true
relayer_region	ap-southeast-1
remote_enabled	false
remote_connected	false
remote_server	ap-southeast-1-0.ui.nabu.casa
alexa_enabled	true
google_enabled	true
cloud_ice_servers_enabled	true
certificate_status	ready
instance_id	...
iot_state	connected
iot_tries	0
certificate_expire_date	December 17, 2026 at 11:10 PM
certificate_fingerprint	....
certificate_alternative_names	undefined
can_reach_cert_server	ok
can_reach_cloud_auth	ok
can_reach_cloud	ok

HACS

GitHub API	ok
GitHub Content	ok
GitHub Web	ok
HACS Data	ok
GitHub API Calls Remaining	5000
Installed Version	2.0.5
Stage	running
Available Repositories	4123
Downloaded Repositories	10

Home Assistant Supervisor

host_os	Home Assistant OS 18.2
update_channel	stable
supervisor_version	supervisor-2026.09.2
agent_version	1.10.0
docker_version	29.6.2
disk_total	234.0 GB
disk_used	13.8 GB
nameservers	192.168.8.1
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	Matter Server (9.2.0), Zigbee2MQTT (2.14.1-1), Mosquitto broker (7.1.1), chrony (6.1.0), File editor (6.1.0), MariaDB (3.0.1), phpMyAdmin (0.14.0), Network UPS Tools (0.18.1), Samba share (12.10.0), Terminal & SSH (10.5.0), Get HACS (1.3.1), Home Assistant Google Drive Backup (0.112.1), AirCast (5.2.1), Nginx Proxy Manager (4.0.3)

Dashboards

dashboards	1
resources	1
mode	auto-gen

Network Configuration

adapters	lo (disabled), enp2s0 (enabled, default, auto), enp2s0.10 (disabled), hassio (disabled), docker0 (disabled), veth92d9551 (disabled), vethced3a45 (disabled), vethf8d869f (disabled), veth5f51ba0 (disabled), vethb945406 (disabled), veth8e9a1dc (disabled), veth8bcf4a5 (disabled), veth98c0ec7 (disabled), veth7361ecf (disabled), vethe103fed (disabled), veth246b08e (disabled), veth0a11fe8 (disabled)
ipv4_addresses	lo (127.0.0.1/8), enp2s0 (192.168.8.19/24), enp2s0.10 (192.168.10.19/24), hassio (172.30.32.1/23), docker0 (172.30.232.1/23), veth92d9551 (), vethced3a45 (), vethf8d869f (), veth5f51ba0 (), vethb945406 (), veth8e9a1dc (), veth8bcf4a5 (), veth98c0ec7 (), veth7361ecf (), vethe103fed (), veth246b08e (), veth0a11fe8 ()
ipv6_addresses	lo (::1/128), enp2s0 (fdad:e520:5169:494f:d9fa:c13e:27ce:24f3/64, fe80::11ec:c90c:738e:47e8/64), enp2s0.10 (fe80::abd1:ee27:3606:8c3f/64), hassio (fd0c:ac1e:2100::1/48, fe80::2c51:55ff:fe3c:aa4e/64), docker0 (fdc6:b40c:4d0b::1/64, fe80::4814:96ff:fe44:2037/64), veth92d9551 (fe80::7cba:f2ff:fe22:2074/64), vethced3a45 (fe80::d46c:27ff:fe28:64ea/64), vethf8d869f (fe80::a8de:7dff:fecb:b80/64), veth5f51ba0 (fe80::145e:43ff:fe91:f6b3/64), vethb945406 (fe80::f8b1:e8ff:fe5e:4b6e/64), veth8e9a1dc (fe80::6831:3eff:fe53:71d0/64), veth8bcf4a5 (fe80::cc62:2cff:fe61:b078/64), veth98c0ec7 (fe80::78be:8aff:fecc:ec97/64), veth7361ecf (fe80::249b:3aff:fe31:920d/64), vethe103fed (fe80::54a8:c7ff:fe65:a860/64), veth246b08e (fe80::74d7:b1ff:fe15:edf4/64), veth0a11fe8 (fe80::602f:44ff:fe74:3169/64)
announce_addresses	192.168.8.19, fdad:e520:5169:494f:d9fa:c13e:27ce:24f3, fe80::11ec:c90c:738e:47e8

Recorder

oldest_recorder_run	August 23, 2026 at 10:58 AM
current_recorder_run	September 21, 2026 at 7:16 PM
estimated_db_size	1236.45 MiB
database_engine	mysql
database_version	11.4.10

Xiaomi Miot

component_version	1.1.5
can_reach_server	ok
can_reach_spec	ok
logged_accounts	1
total_devices	22

Additional information

board: generic-x86-64 boot: A boot_slots: A: state: booted status: good version: "18.2" B: state: inactive status: good version: "18.3" update_available: true version: "18.2" version_latest: "18.3" version_pending: null

After the failure, ha os info showed HAOS 18.3 as the booted OS.

I used the local console because the machine was no longer reachable over the network and ran:

os boot-slot other

The machine rebooted into HAOS 18.2 and Ethernet immediately started working again.

After rollback:

Slot A: HAOS 18.2, booted, status good
Slot B: HAOS 18.3, inactive, status good
HAOS 18.2 kernel: 6.18.39-haos

No network, switch, cabling, or Home Assistant configuration changes were made between the failing 18.3 boot and the working 18.2 boot.

Source: home-assistant/operating-system

View original on GitHub View discussion on GitHub