Home Assistant OS 17.0.rc1

[sairon]

Changes

Home Assistant OS 17.0.rc1 is a major release bringing significant updates to all platforms. This release also drops support for 32-bit armv7 targets (Raspberry Pi 2/3/4 32-bit, ODROID-XU4, and ASUS Tinker Board) as announced in May 2025. The Linux kernel has been updated to the latest version across all targets, and Docker has been upgraded to v29.1.2 with containerd v2.2.0. New installations now use the Docker containerd snapshotter by default. Console keyboard layouts can now be configured through localectl, making it easier to use non-US keyboard layouts when accessing the system console directly. All users on supported platforms are recommended to update.

⚠️ Breaking Changes

This release removes support for the following 32-bit armv7 targets:

• Raspberry Pi 2 (rpi2)
• Raspberry Pi 3 32-bit (rpi3)
• Raspberry Pi 4 32-bit (rpi4)
• ODROID-XU4 (odroid-xu4)
• ASUS Tinker Board (tinker)

Users of these devices will no longer receive updates. If you are using one of these boards, your device will no longer receive updates, please consider migrating to a supported 64-bit platform.

Docker Containerd Snapshotter

Starting with this release, new installations use the Docker containerd snapshotter by default, which provides improved performance and resource efficiency. Existing users can migrate to the new storage driver using the following command:

ha docker migrate-storage-driver

Before migrating, please note:

• Create a full Home Assistant backup before proceeding
• Internet connectivity is required for re-downloading all container images
• At least 50% free storage space is recommended before starting the migration
• The migration will be applied on the next system reboot

Home Assistant Operating System

• Bump OS Agent to v1.8.1 (Bump OS Agent to v1.8.1 #4433) @sairon
• Update to Docker v29.1.2, runC v1.3.4, containerd v2.2.0 (Update to Docker v29.1.2, runC v1.3.4, containerd v2.2.0 #4429) @sairon
• Make console keymaps configurable through localectl (Make console keymaps configurable through localectl #4424) @sairon
• Bump OS Agent to v1.8.0 (Bump OS Agent to v1.8.0 #4420) @sairon
• Bump Buildroot to 2025.02.8 (Bump Buildroot to 2025.02.8 #4415) @sairon
• Add support for QingHeng CH9200 USB ethernet adapters (Add support for QingHeng CH9200 USB ethernet adapters #4400) @jmarshallnz
• Remove Docker patch already included in v28.5.2 (Remove Docker patch already included in v28.5.2 #4398) @sairon
• Use correct per-channel AppArmor profile on clean installs (Use correct per-channel AppArmor profile on clean installs #4396) @sairon
• Update to Docker v28.5.2, runC v1.3.3, containerd v2.1.5 (Update to Docker v28.5.2, runC v1.3.3, containerd v2.1.5 #4395) @sairon
• Update Go to v1.25 (Update Go to v1.25 #4392) @sairon
• Enable iptables comments for all targets (Enable iptables comments for all targets #4391) @monosoul
• Remove support of armv7 targets (Remove support of armv7 targets #4374) @sairon
• Remove handling of Docker key.json (Remove handling of Docker key.json #4361) @sairon
• Use Docker containerd snapshotter for new and wiped installs (Use Docker containerd snapshotter for new and wiped installs #4360) @sairon

Raspberry Pi

• Remove support of armv7 targets (Remove support of armv7 targets #4374) @sairon

Home Assistant Green

• Linux: Update kernel to 6.12.60 (Linux: Update kernel to 6.12.60 #4425) @sairon

Open Virtual Appliance

• Linux: Update kernel to 6.12.60 (Linux: Update kernel to 6.12.60 #4425) @sairon

Generic x86-64

• Linux: Update kernel to 6.12.60 (Linux: Update kernel to 6.12.60 #4425) @sairon

Hardkernel ODROID

• Linux: Update kernel to 6.12.60 (Linux: Update kernel to 6.12.60 #4425) @sairon
• Remove support of armv7 targets (Remove support of armv7 targets #4374) @sairon

Khadas VIM Series

• Linux: Update kernel to 6.12.60 (Linux: Update kernel to 6.12.60 #4425) @sairon

Generic aarch64

• Linux: Update kernel to 6.12.60 (Linux: Update kernel to 6.12.60 #4425) @sairon

Build System

• Remove deprecated boards from build matrix and .github templates (Remove deprecated boards from build matrix and .github templates #4399) @sairon
• Clean up hassio build scripts (Clean up hassio build scripts #4394) @sairon
• Use OCI archive in hassio import to fix Containerd snapshotter issues (Use OCI archive in hassio import to fix Containerd snapshotter issues #4393) @sairon
• Fix tests to accommodate wait for container re-download on downgrade (Fix tests to accommodate wait for container re-download on downgrade #4368) @sairon

Dependencies

• Bump actions/checkout from 6.0.0 to 6.0.1 (Bump actions/checkout from 6.0.0 to 6.0.1 #4438) @dependabot[bot]
• Bump actions/setup-python from 6.0.0 to 6.1.0 (Bump actions/setup-python from 6.0.0 to 6.1.0 #4437) @dependabot[bot]
• Bump actions/stale from 10.1.0 to 10.1.1 (Bump actions/stale from 10.1.0 to 10.1.1 #4439) @dependabot[bot]
• Bump actions/checkout from 5.0.0 to 6.0.0 (Bump actions/checkout from 5.0.0 to 6.0.0 #4413) @dependabot[bot]
• Bump mikepenz/action-junit-report from 6.0.0 to 6.0.1 (Bump mikepenz/action-junit-report from 6.0.0 to 6.0.1 #4379) @dependabot[bot]
• Bump actions/upload-artifact from 4.6.2 to 5.0.0 (Bump actions/upload-artifact from 4.6.2 to 5.0.0 #4364) @dependabot[bot]
• Bump actions/download-artifact from 5.0.0 to 6.0.0 (Bump actions/download-artifact from 5.0.0 to 6.0.0 #4365) @dependabot[bot]
• Bump mikepenz/action-junit-report from 5.6.2 to 6.0.0 (Bump mikepenz/action-junit-report from 5.6.2 to 6.0.0 #4352) @dependabot[bot]

---

This discussion was created from the release Home Assistant OS 17.0.rc1.

[Anto79-ops]

ok i updated and my system came back and no issues on installation and bootup on my NUC 12 (x86-64, bare-metal) with USB ZBT-2 and POE/USB Zwave ZWA-2.

The system is running normally.

I'm confused about this:

Starting with this release, new installations use the Docker containerd snapshotter by default, which provides improved performance and resource efficiency. Existing users can migrate to the new storage driver using the following command:

ha docker migrate-storage-driver

Do I need to run this or do I need to reboot the syste again or do nothing?

thanks

[Baxxy13]

Do the command in the cli then you get a repair from supervisor to reboot.
Don't forget to backup first!

[Anto79-ops]

ok, followed those instructions. System came back with some repairs (Ignore the Jeep repair):

ZwaveJs-UI and Grafana addons failed to start, so I simply restarted them manually and now they have restarted.

There is one repair here:

Not sure how to fix this. Attached are supervisor logs:

supervisor_2025-12-09T17-41-48.688Z.log

Is this a false positive? My system appears to running normally.

maybe I just need to do a ha supervisor restart, but before I do, I'll wait for more direction.

[Anto79-ops]

ok, after ha supervisor restart , the repair is gone and the system appears to be healthly again...

.....
2025-12-09 10:59:32.045 INFO (MainThread) [supervisor.resolution.checks.base] Run check for docker_config/system
2025-12-09 10:59:32.045 INFO (MainThread) [supervisor.resolution.checks.base] Run check for ipv4_connection_problem/system
2025-12-09 10:59:32.045 INFO (MainThread) [supervisor.resolution.checks.base] Run check for no_current_backup/system
2025-12-09 10:59:32.045 INFO (MainThread) [supervisor.resolution.module] Create new suggestion create_full_backup - system / None
2025-12-09 10:59:32.045 INFO (MainThread) [supervisor.resolution.module] Create new issue no_current_backup - system / None
2025-12-09 10:59:32.046 INFO (MainThread) [supervisor.resolution.checks.base] Run check for dns_server_ipv6_error/dns_server
2025-12-09 10:59:32.049 INFO (MainThread) [supervisor.resolution.check] System checks complete
2025-12-09 10:59:32.049 INFO (MainThread) [supervisor.resolution.evaluate] Starting system evaluation with state running
2025-12-09 10:59:32.098 INFO (MainThread) [supervisor.resolution.evaluate] System evaluation complete
2025-12-09 10:59:32.098 INFO (MainThread) [supervisor.resolution.fixup] Starting system autofix at state running
2025-12-09 10:59:32.098 INFO (MainThread) [supervisor.resolution.fixup] System autofix complete

[Anto79-ops]

and just for good measure, I rebooted the host system again and no (new) errors in supervisor, host or core related to this upgrade.

so, besides the couple of addons not starting by themselves on startup and supervisor needing a manual restart, all is going well now.

[sairon]

Thanks for the feedback - yes, it seems that sometimes Supervisor doesn't cope well with the missing images when it starts up too soon. Once everything is downloaded, the errors are cleared up by a reboot or just a Supervisor restart. We should add some guards to make sure it's handler more gracefully, it's just tedious to reproduce.

[alexruffell]

My instance is running in a ProxMox VM and after running the CLI command and rebooting, I got this error:

The only pertinent error seems to be:

Logger: homeassistant.core
Source: core.py:532
First occurred: 15:51:12 (1 occurrence)
Last logged: 15:51:12

Something is blocking Home Assistant from wrapping up the start up phase. We're going to continue anyway. Please report the following info at https://github.com/home-assistant/core/issues: unifi.light, xiaomi_miot.text, ecobee.notify, systemmonitor, assist_satellite, openweathermap.weather, backup.sensor, hassio.sensor, unifiprotect.number, dlna_dmr, file, unifi.image, playstation_network.image, scene, workday, nut.switch, counter, xiaomi_miot.fan, adguard.update, schedule, zha.select, group.light, alarm_control_panel, mobile_app.sensor, pi_hole_v6.sensor, input_button, camera, music_assistant, nest_protect.switch, onkyo, wyoming, sonos, wyoming.tts, integration, xiaomi_miot.water_heater, system_log, lifx.binary_sensor, calendar, withings, search, google_generative_ai_conversation.ai_task, pi_hole_v6.binary_sensor, moon.sensor, ping.sensor, zha, playstation_network.binary_sensor, wiim.number, select, unifiprotect.media_player, history, persistent_notification, ecobee, unifi.update, input_number, nut.button, unifi.button, homekit_controller, wled.button, remote, wyoming.stt, water_heater, nodered.sensor, application_credentials, hacs.switch, zha.climate, zwave_js.lock, unifi.device_tracker, sonos.media_player, music_assistant.button, zwave_js.fan, google_generative_ai_conversation, automation, recorder, zwave_js.light, rachio.binary_sensor, template.binary_sensor, otbr, nodered.text, humidifier, unifiprotect.switch, wled.light, lifx.select, template, wiim.light, cast.media_player, button, adguard.switch, wake_word, event, xiaomi_miot.media_player, sonos.binary_sensor, holiday, onkyo.media_player, xiaomi_miot.number, wiim.media_player, hassio.binary_sensor, spotcast, envisalink, device_tracker, unifiprotect.light, homekit_controller.button, binary_sensor, zha.siren, cast, watchman, ffmpeg, lovelace, hacs.update, wyoming.wake_word, number, bluetooth_adapters, androidtv_remote.remote, nodered, zwave_js, playstation_network, nodered.binary_sensor, uptime_kuma.sensor, hardware, xiaomi_miot.camera, esphome.media_player, auto_backup.button, envisalink.alarm_control_panel, nodered.time, nut, dlna_dmr.media_player, uptime_kuma, pi_hole_v6.update, auth, climate, roku.media_player, zeroconf, lifx.light, energy, webhook, esphome.event, wled.switch, withings.calendar, openweathermap.sensor, zwave_js.event, go2rtc, trace, roku.remote, default_config, sonos.sensor, zha.device_tracker, xiaomi_miot.vacuum, google_generative_ai_conversation.conversation, image, ssdp, xiaomi_miot, stt, watchman.sensor, xiaomi_miot.sensor, workday.calendar, template.switch, zwave_js.climate, zha.fan, zwave_js.binary_sensor, nodered.select, esphome.cover, zha.update, lifx, cover, unifiprotect.camera, google_generative_ai_conversation.tts, nut.sensor, zwave_js.siren, workday.binary_sensor, cloud.stt, linkplay, wiim, unifiprotect.button, http, homeassistant_alerts, xiaomi_miot.climate, nest_protect.sensor, ping.device_tracker, zha.alarm_control_panel, holiday.calendar, openweathermap, xiaomi_miot.device_tracker, zha.cover, logger, cloud, playstation_network.notify, auto_backup.sensor, auto_backup, usage_prediction, unifiprotect.text, androidtv_remote.media_player, google_assistant, update, logbook, wled.sensor, esphome.button, withings.sensor, openai_conversation.ai_task, esphome.assist_satellite, onboarding, roku.sensor, homekit_controller.binary_sensor, rachio.switch, homekit_controller.sensor, uptime, nest_protect, zha.sensor, tts, victorsmartkill, hassio.update, text, ecobee.sensor, esphome.light, blueprint, xiaomi_miot.humidifier, xiaomi_miot.light, sun, zha.light, zwave_js.humidifier, zwave_js.button, google_generative_ai_conversation.stt, cloud.binary_sensor, zwave_js.number, ecobee.switch, systemmonitor.binary_sensor, my, cloud.tts, harmony.select, group, usb, sonos.select, rachio, mobile_app.notify, image_upload, hassio.switch, timer, hacs, mobile_app.binary_sensor, sensor, integration.sensor, file_upload, samsungtv, input_boolean, ping.binary_sensor, nodered.switch, unifiprotect.binary_sensor, zha.binary_sensor, radio_browser, ecobee.humidifier, unifiprotect.event, nodered.button, ipp, linkplay.button, envisalink.binary_sensor, linkplay.select, sun.sensor, device_automation, homeassistant.scene, esphome.switch, wiim.sensor, pi_hole_v6.button, nws_alerts.sensor, victorsmartkill.binary_sensor, rest.sensor, withings.binary_sensor, weather, backup, unifiprotect.select, repairs, xiaomi_miot.cover, ecobee.number, uptime_kuma.update, uptime.sensor, zwave_js.select, group.binary_sensor, siren, adguard, nws_alerts, zha_toolkit, zwave_js.update, notify, person, frontend, favicon, sonos.switch, linkplay.media_player, rest, input_select, wled.number, homeassistant_hardware, xiaomi_miot.alarm_control_panel, lock, zone, local_calendar.calendar, spotify.media_player, spotify, min_max, input_text, unifi.sensor, zwave_js.switch, envisalink.sensor, alexa, network, esphome, min_max.sensor, config, script, assist_pipeline, vacuum, analytics, lifx.button, fan, sun.binary_sensor, esphome.number, roku.select, unifi.switch, esphome.select, hassio, harmony.remote, zha.number, matter, nest_protect.binary_sensor, samsungtv.media_player, unifi, ecobee.binary_sensor, adguard.sensor, playstation_network.media_player, esphome.binary_sensor, group.lock, wled, dhcp, xiaomi_miot.switch, unifiprotect.lock, wiim.select, ping, esphome.sensor, systemmonitor.sensor, openai_conversation.conversation, ipp.sensor, rachio.calendar, music_assistant.media_player, zwave_js.sensor, mobile_app, system_health, roku, template.sensor, unifiprotect, ecobee.weather, zha.lock, backup.event, wiim.button, zha.button, api, time_date.sensor, esphome.lock, xiaomi_miot.binary_sensor, influxdb, stream, switch, zwave_js.cover, nodered.number, nest_protect.select, roku.binary_sensor, victorsmartkill.sensor, pi_hole_v6.switch, energy.sensor, xiaomi_miot.remote, auto_backup.binary_sensor, time, homeassistant, wled.select, ai_task, websocket_api, samsungtv.remote, labs, openai_conversation, esphome.update, sonos.number, pi_hole_v6, intent, bluetooth, mobile_app.device_tracker, harmony, zha.switch, time_date, conversation, local_calendar, esphome.climate, media_source, ecobee.climate, tag, thread, media_player, light, androidtv_remote, unifiprotect.sensor, playstation_network.sensor, moon, xiaomi_miot.button, diagnostics, lifx.sensor, xiaomi_miot.select, wled.update, input_datetime, group.switch

The system is waiting for tasks: {<Task pending name='Task-11128' coro=<AbstractConfig.async_initialize..sync_google() running at /usr/src/homeassistant/homeassistant/components/google_assistant/helpers.py:114> wait_for=<_GatheringFuture pending cb=[Task.task_wakeup()]> cb=[set.remove()]>, <Task pending name='Task-11682' coro=<async_enable_report_state..initial_report() running at /usr/src/homeassistant/homeassistant/components/google_assistant/report_state.py:187> wait_for=<_GatheringFuture pending cb=[Task.task_wakeup()]> cb=[set.remove()]>}

Edit: a reboot seems to have fixed it.

[Anto79-ops]

the error shown above is actually an old but (unrelated) error related to Google core integration (waiting for a depency fix).

Do you have supervisor logs? Thats where the logs are and you may need to restart it like it with ha supervisor restart in cli

[alexruffell]

@Anto79-ops - Thanks for the tip. I attached the log.
supervisor_2025-12-09T22-05-30.991Z.log

It appears HA was complaining that Google Assistant integration was causing the issue... but I don't have one. The closest is Google Gemini.

Edit: Maybe it was referring to the default "Google Assistant" under Voice Assistants...

[Anto79-ops]

@alexruffell your logs looks very similar to mine. Probably a race condition somewhere.

[def1149]

Updated to 17.rc1 and performed the Docker conversion
Other than some bogus error messages, everything seems to be OK.
I'd add an advisory to the release notes that the Docker conversions after the reboot take a long time

[Baxxy13]

Just updated my HA-OS Proxmox VM to OS 17.0.rc1 without issues.

The reboot after ha docker migrate-storage-driver takes a very long time, possibly due to slow downloads of the containers from GitHub.
But in the end the system comes back up without issues.

The only thing I've noticed so far is that the system now takes up about 1.6 GB more storage space.

Before:

After:

[Anto79-ops]

i can confirm the same (but about 5 Gb more space), red dot was installation of 17.0rc1 and green dot was when I ran the ha docker migrate-storage-driver command

[sairon]

Unfortunately, this is a fact we need to live with, the Containerd overlayfs storage driver has slighly higher overhead because of the way how it stores downloaded image layers. In long run, since Docker is slowly moving away from the legacy graph driver, we would have no other choice anyway - at least it has some other benefits that should outweigh it.

[def1149]

Space used increased 3GB after the 17.rc1 update.
After 17 running, no observable issues.

[WoBBeLnl]

Note: if you are using AdGuard add-on and your HA is using that as its DNS...upgrading is uuhmmmm...challenging 🥲 it could not resolve DNS to download the images.

My logs were spammed with issues getting the images. Hope everything is starting again...

[edit]
How to check if migration is successful?

[Baxxy13]

How to check if migration is successful?

Good question, my primary (not converted) HA (HA-OS 16.3) shows:

# ha docker info
enable_ipv6: null
logging: journald
mtu: null
registries: {}
storage: overlay2
version: 28.3.3

And the converted one:

# ha docker info
enable_ipv6: null
logging: journald
mtu: null
registries: {}
storage: overlayfs
version: 29.1.2

[def1149]

[edit] How to check if migration is successful?

[WoBBeLnl]

How to check if migration is successful?

Good question, my primary (not converted) HA (HA-OS 16.3) shows:

# ha docker info
enable_ipv6: null
logging: journald
mtu: null
registries: {}
storage: overlay2
version: 28.3.3

And the converted one:

# ha docker info
enable_ipv6: null
logging: journald
mtu: null
registries: {}
storage: overlayfs
version: 29.1.2

Great, thanks!

[sairon]

The main clue is storage: overlayfs line - this comes directly from Docker API, so overlayfs means the new backend is indeed used, overlay2 is the legacy graph driver.

[bdelima]

Question: Does 17.0 kernel have support for Intel Core Ultra NPUs? Would be very useful if it did.

[AllardKatan]

Update to 17.0rc1 was smooth. Storage upgrade went well too, no error messages. All is good :-)

My instance is OVA, Virtualbox 7 under Linux.