Hello,
I have had Home Assistant running on an Odroid-N2+ for 2 years now.
For the past week I have had a strange phenomenon that I cannot solve. Out of nowhere, Home Assistant suddenly freezes. Home Assistant will then no longer be available. However, I can still ‘ping’ his IP address.
I do note that the blue LED on the Odroid quickly flashes continuously in that case.
Any idea what’s wrong?
Let’s start with some more info please:
Hello Mark,
Meanwhile, I have an even bigger problem…
Previously I could simply unplug the power cable and plug it back in and HA would restart. Now nothing happens anymore…the blue LED displays the typical heartbeat (flashing twice followed by a pause, sometimes slow, sometimes fast). When I ping the device I get no answer. I do not have an orange/green LED on the connection of the network cable.
So I am indeed working with an SD card, but even when I plug in another one (which also has HA on it and worked perfectly last week) it doesn’t work.
Of course I can’t look at the logs anymore, but I did this last week and I didn’t see anything special there.
Why use a different power adapter, because this one seems to work anyway (sorry, I don’t know anything about it)?
Power supplies can fail. Try another power supply. Note: cycling the power without doing a proper shutdown can corrupt the data on the SD card.
Yes, I know but…
After unplugging and plugging several it all of a sudden started again ! I can’t see anything special in the log.
Problem is I don’t have another power supply, so will have to buy one.
Hope it helps…
After 2 days it was down again.
I plugged out and in, the system restarted…but when starting home assistant I see this :
Home assistant
To the Home Assistant command line.
for supervisor to startup…
Information
Returned from Supervisor: System is not ready with state: setup
systemd(1): systemd-timesyncd.service: Watchdog timeout (limit 3min)!
systemd(1): systemd-udevd.services: Watchdog timeout (limit 3min)!
systemd(1): systemd-tilesyncd.service: Watchdog timeout (lilit 3min)!
systemd (1): Failed to start Journal Service
systemd (1): Failed to start Rule-based Manager for Device Events and Files
systemd(1): systemd-tilesyncd.service: Watchdog timeout (lilit 3min)!
systemd (1): Failed to start Journal Service
…
I’ve attached another power supply and had to try several times to get Home Assistant alive :
I still can’t imagine the power supply is the problem, but we will see…
The HA files on your current SD card may be corrupted. Have you tried installing HA on a new SD card and restoring from backup?
Also, reviewing this may give some additional help
Edit:
I think the problem could fall into one of two categories: 1) hardware (power, SD Card, Odroid); or 2) Software (corruption, misbehaving integration, incorrect configuration, etc.).
I don’t think hardware can be ruled out yet. So, even though you have tried another SD card and power supply, there may be something else. There have been issues with additional devices plugged into the USB ports that can exceed the the limits of the standard power supply. What devices are plugged into the Odroid USB ports?
Has there been any change to the hardware? Is it placed in a good environment - proper cooling, no local RF interference (microwave, strong wifi access point, etc.), ethernet cable has been checked, etc.
The software troubleshooting link above starts with examining the home-assistant.log files in the configuration directory. Note that there are multiple files: home-assistant.log, home-assistant.log.1, and home-assistant.log.fault. Check all three.
Finding the source of software errors can be difficult. You said that you tried a previous version of HA on another SD card. This was a good test but the same problem exists (configuration, corruption, etc.). Installing a fresh copy of HA and using a backup to restore the configuration may give us some more info. But if that does not work, you may have to start removing integrations one at a time to see if an integration is the problem or rebuilding your system from scratch (which I know is very painful). And this would be very aggravating if it did not solve the problem.
Configuration problems external to your HA environment may also be the issue. Changes in the network, either local or by your ISP, may impact how HA works. The devices being monitored and/or controlled by HA may have an impact (hardware / firmware failure, increased or corrupted communication, etc.). This could cause an integration to fail.
I think the systemd-timesyncd.service error indicates that you cannot get a valid time from the NTP server. Have there been any changes in your network configuration?
I have decided (I was planning to anyway) to switch to eMMc (it has been ordered and will hopefully arrive soon). A new Home Assistant installation and restoring my last backup will hopefully solve the problem. Currently, Home Assistant remains up and running, but we are only day one…
I will keep you informed of further developments.
Thanks for helping me figuring out what the problem may be and… Merry Christmas !
