Home Assistant patially stopped woring and now won't boot

Installation Home Assistant OS
1

So I came home today and realised that my HA (installed on a RPi3 B+, on a USB stick - this should be as reliable as a ssd, right?!) showed all Zigbee2Mqtt devices as unavailable since a couple of hours- I tried to restart the Addon, but it didn’t do anything. So I looked into the logs and this is what it logged at the time everything stopped working:

Error doing job: Future exception was never retrieved
Traceback (most recent call last):
File „/usr/local/lib/python3.8/concurrent/futures/thread.py“, line 57, in run
Result = self.fn(*self.args, self.kwargs)
File „/config/custom_components/thermal_comfort/sensor.py“, line 125, in humidity_state_listener
Self._humidity = float(new_state.state)
ValueError: could not convert string to float: ‚unavailable‘

So I thought I just restart Home Assistant (through the interface) - now it won’t boot at all, and I have no idea what to do… EVERYTHING in my house is controlled by HA - and the last backup is 2 weeks old …

UPDATE:
Ok, the same thing happened today again … in the middle of the night, everything stopped working and the Zgbee2Mqtt devices became unavailable.

Since then the logs repeate over and over :

20-10-26 05:47:07 INFO (MainThread) [supervisor.plugins.observer] Starting observer plugin
20-10-26 05:47:07 INFO (SyncWorker_7) [supervisor.docker.interface] Cleaning hassio_observer application
20-10-26 05:48:09 ERROR (SyncWorker_7) [supervisor.docker] Can’t start hassio_observer: 403 Client Error: Forbidden (“endpoint with name hassio_observer already exists in network hassio”)
20-10-26 05:48:09 ERROR (MainThread) [supervisor.plugins.observer] Can’t start observer plugin
20-10-26 05:48:09 ERROR (MainThread) [supervisor.misc.tasks] Observer watchdog reanimation failed!
20-10-26 05:49:09 WARNING (MainThread) [supervisor.misc.tasks] Watchdog/Docker found a problem with observer plugin!

What is this observer plugin? how can I fix it?

2

USB stick reliability is very varied. From significantly worse than an SD card, to much the same.

You could try recovering your config from the USB stick, but most likely you’re going to need that backup. That and set up regular automatic backups :wink:

3

Just a bummer, it didn’t even last 2 weeks lol

And I didn’t manage to boot from any of my external sata drives…

I guess I’ll try to get my hands on a Compute Moduel 4 with the board as soon as possible…

how do I do that? I assume it’s not as simple as plugging it in my PC and copying the files ?! (the only partition I can access is boot)

How would I access them in the end? they are pretty useless, if they are szpred locally, i guess

4

That depends on how you take backups. Snapshots aren’t backups until they’re off the computer. AFAIK there’s at least two add-ons that can help here, but I don’t use this install method.

You’ll need to install ext3 drivers on your PC, and then you’ll be able to mount the other partitions. Your files will be on one of those (data probably), assuming they’re not corrupt.

1 Like
5

I believe I managed to restore most of my configuration

I’m just so glad it didn’t stop working while the heaters were on - I have to implement some sort of failsafe

6

Ok, the same thing happened today again … in the middle of the night, everything stopped working and the Zgbee2Mqtt devices became unavailable.

Since then the logs repeat over and over :

20-10-26 05:47:07 INFO (MainThread) [supervisor.plugins.observer] Starting observer plugin
20-10-26 05:47:07 INFO (SyncWorker_7) [supervisor.docker.interface] Cleaning hassio_observer application
20-10-26 05:48:09 ERROR (SyncWorker_7) [supervisor.docker] Can’t start hassio_observer: 403 Client Error: Forbidden (“endpoint with name hassio_observer already exists in network hassio”)
20-10-26 05:48:09 ERROR (MainThread) [supervisor.plugins.observer] Can’t start observer plugin
20-10-26 05:48:09 ERROR (MainThread) [supervisor.misc.tasks] Observer watchdog reanimation failed!
20-10-26 05:49:09 WARNING (MainThread) [supervisor.misc.tasks] Watchdog/Docker found a problem with observer plugin!

What is this observer plugin? how can I fix it?

7

Your first post stated an error by a custom component, you say that HA won’t start ever since. My advice is to start troubleshooting from there first. Remove the custom component related to the error and remove (or comment out) everything related to that component (the log makes me think it is a sensor, so start looking there).

Once you’ve made sure that that component is not the problem continue to the next phase of troubleshooting.

1 Like
8

Why? If you are looking to update your Pi3, the compute module is not the answer.

What have you used to re-install HA? Another USB stick, or have you moved to an SD card?

9

I thought it might be Zigbee2MQTT, because all devices related to it stopped working at exactly the same time. But nothing else seems to work correctly either (everything is incredibly slow - starting from regular mqtt devices, to the web interface).

I tried stopping the addon, but it just wouldn’t stop… the errors always showed an issue with obverver plugin or watchdog… I have no idea what do do with this information.

I thought the Pi was more than powerful enough to run HA and the only shortcomming is the SD card - so why not a Pi4 with an SSD?!

A different, new USB stick. I tried to use three different USB SSD drives (regular SATA to USB enclosures), but none of them boot.

The USB drives are broken now btw - both can’t be re-flashed … even diskpart got stuck a couple of times when I deleted the partitions…

10

Yes - a Pi4, not a compute module.

It is, you need an SSD that will boot though. Some drive enclosure are not compatible, others will need you to provide a better power supply for the Pi to handle the draw.

I have been using one of these for some time on a Pi3, and more recently on a 4gb Pi4, running HA Supervised as a backup to my production machine.

This page has a list that is useful for known working drives and enclosures and info on using quirks. I have purchased this one and can confirm it works well with a WD Green SSD.

Stop using USB sticks, use an SD card until you can get a compatible SSD drive or enclosure.

11

I just changed to USB sticks, because I thought SD cards were the absolute worst thing to run HA on hahah - guess I misunderstood something.

the enclosure I would like to use is a cheap ChiliGreen with an AS2105 Chip inside - guess that’s no good, since it’s not mentioned in the documentation?!

Isn’t it the same thing, but the compute module has a compatible extension board with a PCIe slot?!

12

Why would you want to complicate things by using a product that is untested with HA, when you can pickup a Pi4 and compatible SSD today and know it will work?

If you setup your recorder config well, it can vastly extend the life of the SD card. I have had one running for well over a year, probably closer to 2 years by now, in my parents house on a Pi3. Here is the recorder config I use.

recorder:
  purge_keep_days: 7
  include:
    domains:
      - sensor
      - binary_sensor
      - switch
      - light
      - device_tracker
      - vacuum
      - group
      - input_boolean
      - cover
      - person
      - counter
      - timer
  exclude:
    event_types:
      - service_removed
      - service_executed
      - platform_discovered
      - homeassistant_start
      - homeassistant_stop
      - feedreader
      - service_registered
      - call_service
      - component_loaded
      - logbook_entry
      - system_log_event
      - automation_triggered
      - script_started
      - timer_out_of_sync
13

Good Point!

Thanks - I already excluded a lot of stuff, but will probably add more to de-clutter the DB