Official raspberry Pi 4 power adapter (changed within the last year or so due to a power issue)
Argon One case
NVMe drive (from a Steam deck) in a UGREEN enclosure. When this problem started a few days ago, I was running from an Integral SSD
Usual Culprits Tried with No Success
Deleting the home_assistant_v2.db file (using the file editor)
The power supply has been changed relatively recently, but can’t be ruled out
2 different storage devices have been used (SSD and NVMe drive with enclosure). This included restoring an old backup from a few days before (28th Jan). Interestingly the problem wasn’t fixed by this, even though it should have been a few days before the issue started
What it Looks Like
The supervisor logs show the message:
Watchdog found Home Assistant failed, restarting...
So perhaps the supervisor is fine, and I should be looking in home assistant core, but nothing is jumping out.
Apologies for the short logs - I’m not really sure how to get the full logs. Currently grabbing them through the (settings → system → logs) page
Note this is an installation I’ve had for years. Just under a year ago I had a similar issue that was fixed by replacing the power adapter with this, but otherwise everything has run fine since. This problem then started on 31st Jan. First there was also an issue with the add-ons not showing as running (even when the logs showed they were), but one of the subsequent restores/db deletes seems to have fixed that. Now the addons show as running, but don’t run for very long before restarting again.
The power supply could be overloaded if you have anything plugged into the USB ports that draws enough power to exceed its limits (like an SSD). Note: the recommended RP4 PSU is 3.0A. If you are using the SSD with this power supply, it will cause problems - you should have the SSD plugged into an independently powered USB hub and then the hub plugged into your RP4.
Thanks for the suggestion Mark. I’ll try loading a fresh install onto the now spare SSD and running it for a few hours. If it does the same thing, then it’s likely to be a power issue
Ok I’ve tried running a fresh install for the most part of the day, and it hasn’t rebooted itself in the 8+ hours it’s been running. That means it’s unlikely to be a power issue (unless it only occurs under the heavy load of a long running installation).
I cant get in long enough to do anything, I am trying now to restore again but every time I try it comes up with lost connection. I also tried booting to safe mode.
I did manage to do a download of the log, there isnt much in there and I am no expert, this line does look suspect though
2024-08-28 13:20:20.293 WARNING (Recorder) [homeassistant.components.recorder.util] The system could not validate that the sqlite3 database at //config/home-assistant_v2.db was shutdown cleanly
local tuya
Alexa Media Player
Android debug bridge
google nest
Hacs
mqtt
octopus energy
sonos
pi power supply checker
sqy q
Sony Bravia TV
Synology DSM
Tasmota
Wyoming Protocal
WLED
Denon AVR Network Receiver
If only I could, just cant get in long enough to do anything, been trying all afternoon. As soon as I click for a page it goes blank and I get the message lost connection, reconnecting
You should try running it off a new SSD. I had an issue once where the SSD became un-reliable after multiple years causing crash issues. You should also check and see if something is leaking memory. So from the command prompt use top to monitor amount of free space, and look for a processes eating memory (process growing memory). I run a supervised installed and had an integration with memory leak once. Because swap space was enabled it would hit a point where the OS was stuck swapping memory pages and you couldn’t do anything until it finally crashed a while later. I turned off the swap space, which then meant the system would immediately crash once all memory was consumed, followed by a reboot. This was actually a lot better than being stuck in the a page swapping state.
For the time being I have reverted to 2024.7.4 and its working fine. I will leave it for now as I am waiting for a Pi5 with NVME drive so will be reinstalling anyway.
Good point about swap space, I was using that as well, but until 2024.8.0 its been running fine.