Hope this is the right place for this. I’ve been running HA for a few months now annd I’ve always taken the updates as soon as possible with no problems. This time though when updating from 2026.7.4 to 2026.8.0 it seems to be failing. In checking the supervisor logs after trying, I’m seeing:
2026-08-05 17:37:04.755 ERROR (MainThread) [supervisor.homeassistant.core] Home Assistant has crashed!
2026-08-05 17:37:04.755 CRITICAL (MainThread) [supervisor.homeassistant.core] HomeAssistant update failed → rollback!
HA ll restart and when it comes up I see that I’m still at 2026.7.4. Any pointers on where I need to look to find why it’s failing?
Below are some options you could try together or separately.
You could try stopping every app, and turn the autostart off for each app when the system starts. Or do this one by one with each app, untlil you find the offending one that could be doing this.
Do a manual backup, then download it and store it somewhere safe. And then skip the backup before doing the upgrade. Just make sure you have one in place first.
You could also read the release notes, and see if there are any apps, configuration.yaml changes, or something else that could be preventing the upgrade.
Also, is your server installed on a desktop PC, Raspberry Pi, container, VM, that kind of thing. It changes troubleshooting depending on how it is installed.
Good idea, but no joy. I only have the apps for Zigbee2MQTT, Mosquitto broker, and Terminal & SSH. I stopped them one by one and applied the update, but with the same results.
I’ve tried both with backup before and without. Still no joy.
Should have included that info before. I’m running in a VirtualBox VM on a Ubuntu Linux host. I’ve tried restarting the entire VM, but still no joy.
Ran into this a couple of months ago. Couldn’t update via the UI (failed 3 times), but when I tried via CLI it worked immediately.
Fire up your Terminal & SSH app & type in ha core update - it should go through. Obviously make sure you have a backup downloaded off the machine before yo udo that, in case something goes wrong.
Thanks, but still no joy. I’m getting the same result: Home Assistant crash, failed update → rollback. At least it does roll back automatically so that I don’t have to do a manual restore.
I would expand the hard drive space available up to around 50 or so if you can. I do not know if VMs use virtual memory the way OSes do when installed normally. Any errors in the logs when the update is through the CLI? Or is it the same error?
I am sorry, I was really hoping it was a space issue. AYou could try a full reboot of the host OS, unless you have already done that. I have to do that now and then with my Linux box, that has HA OS installed inside of Virtual Box. HA still works, but the host OS just freezes for some reason. MAKE A BACKUP BEFORE RUNNING ANY OF the below commands, and use them with caution.
Also make sure that Virtual box is updated to the latest version. Easier said then done, depending on the host OS.
I got all of the below from a google search.
Identify the Exact Block
Run the diagnostic commands below to check the health of your system and read the error logs directly:ha core check - Scans your current configuration for breaking changes or syntax errors preventing boot. ha supervisor logs - Shows why the supervisor might be blocking the download. ha core logs - Shows runtime errors inside the Core container. ha info - Displays current version metrics, supervisor health status, and free space.
Clear Disk Space
Updates fail silently or drop errors if your storage is full.
Run ha system info to look at data disk usage. If it is near 100%, delete old partial backups by using ha backups list and then ha backups remove .
Update the Supervisor First
If your logs state that execution is blocked because the supervisor needs to be updated first, update the manager layer before touching the Core layer. ha supervisor update
Bypass the CLI Buffer and Force the Install
Sometimes the CLI communication bridge times out while Docker pulls large image layers. You can bypass the standard catch-all command (ha core update) by specifying an incremental or forced version target. ha core update --version=<insert_latest_version_here>
I updated my VirtualBox to the latest (7.2.14 from 7.0.16). No change.
But I opened two terminals, ran ‘ha supervisor logs --follow’ in one and ‘ha core logs --follow’ in the other. I then applied the update through the GUI, and about the time is saw the errors in the supervisor log I saw this in the core log:
s6-rc: info: service legacy-services: stopping
2026-08-06 18:04:14.828 WARNING (MainThread) [habluetooth.wrappers] HaBleakScannerWrapper.register_detection_callback() is deprecated and will be removed in a future release; bleak already removed this method from BleakScanner. Pass detection_callback to the HaBleakScannerWrapper constructor instead.
2026-08-06 18:04:17.476 ERROR (MainThread) [homeassistant.components.wled] Unexpected error fetching wled data
Traceback (most recent call last):
File “/usr/src/homeassistant/homeassistant/helpers/update_coordinator.py”, line 435, in _async_refresh
self.data = await self._async_update_data()
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File “/usr/src/homeassistant/homeassistant/components/wled/coordinator.py”, line 149, in _async_update_data
device = await self.wled.update()
^^^^^^^^^^^^^^^^^^^^^^^^
File “/usr/local/lib/python3.14/site-packages/backoff/_async.py”, line 164, in retry
ret = await target(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File “/usr/local/lib/python3.14/site-packages/wled/wled.py”, line 302, in update
if not (data := await self.request(“/json”)):
^^^^^^^^^^^^^^^^^^^^^^^^^^^
File “/usr/local/lib/python3.14/site-packages/backoff/_async.py”, line 164, in retry
ret = await target(*args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File “/usr/local/lib/python3.14/site-packages/wled/wled.py”, line 213, in request
response = await self.session.request(
^^^^^^^^^^^^^^^^^^^^^^^^^^^
…<6 lines>…
)
^
File “/usr/local/lib/python3.14/site-packages/aiohttp/client.py”, line 581, in _request
raise RuntimeError(“Session is closed”)
RuntimeError: Session is closed
So it looks like it’s having an issue with my WLED controller. It’s been a while since I’ve dug through any python code, but I’ll try to see if I can figure something out. And if anybody else has any insight, I would appreciate hearing about it.
Nope. I disabled the WLED integration, and while the MainThread error from wled went away, I’m still seeing the crash in the supervisor log. So I don’t think that was it.
However I just noticed this one in the core log:
/usr/local/bin/python3: No module named homeassistant
[00:47:51] INFO: Home Assistant Core finish process exit code 1
[00:47:51] INFO: Home Assistant Core service shutdown
This was just a few seconds before the crash error in the supervisor log. So I’m not sure where to go next. Any pointers?
Hopefully someone else will respond. Sorry, but this is now beyond my skill. There are discord chat rooms for home assistant, You could ask for help in there.
Mine upgraded but got stuck in some sort of tight loop that meant the UI wouldn’t respond. Debugging this via SSH and AI I found that the zigbee coordinator was stuck. Not 100% sure this was the root of the issue but is anyone else using a SMLIGHT SLZB-06 via PoE? The issue seemed to be the same as an old issue that involved this device.
Well, I had to put this aside yesterday while I took care of some stuff around the yard. In the meantime the update to 2026.8.1 came out and when I applied that this morning (using the GUI) it updated just fine. I guess I could roll back to my image of 2026.7.4 and try updating to 2026.8.0 through the CLI for debugging, but at this point since the latest is working, I’m just going to run with it. Thanks for all the responses.