Hassio won't start - "Failed to start Docker application container engine"

Installation Home Assistant OS
21

My Home Assistant is completely dead as well.
I have been noticing the the green LED (microSD read/write LED) pretty much staying on continuously for about three weeks.
I was getting worried the card was going to die any day. I now can’t access HA. Samba share is not accessible. So I started all over with a new SanDisk 64GB High Endurance card. Now, I still can’t get HA to start with the Pi4 32b image. This is the worse HA issue I’ve ever had over the last two years. I have not made any HA or router changes for more then two months… I don’t know where to even begin…:frowning:

22

Found the cause and documented a fix here:

23

I’m having the same problem since the day before yesterday. I’m running HassOs on Virtualbox on a NUC. I made a backup everytime before running HomeAssistant updates, though none of the backups work. I’m getting the “failed to start Docker application container engine” every single time, with every single backup I made. It’s really weird.

I tried to run the “sudo nautilus” command in the shell (option 4 of the boot menu). It says that it can’t find such a directory. I’m really stuck here. @JesseWebDotCom Do you have any other suggestion I can try?

24

Same problem here. Normal reboot work always. Power out causes the problem also always. Running on a Pi4 with 128GB sd card. Reflashing the disks solves the issue but is a pain

15867211089233135877106423209171
158672110892331358771064232091711824×1368 1.02 MB

25

Hey guys. Same issue here, but with fresh installs. I’ve tried version 4.11, 4.12 and 5.1 VDI images with the same error/result

I was running fine in a 8250U NUC using VM.

I made 3 changes in the system before the first crash :
Reduce the volume size of the partition to create a new one.
Change to a different router (same network)
Power off the NUC to move a new place.

Since it was a brand new installation (only one month running), I decided to do another fresh install. But the critical error is still appearing with new VMs and in all the versions I mentioned in the beginning.

VirtualBox_HASSIO_24_08_2020_21_52_54
VirtualBox_HASSIO_24_08_2020_21_52_541024×768 14.9 KB

I have no more ideas what caused it. Now I’m stuck with no HassOS working.

Any clues ?

26

Same problem today after a reboot on my rPI 4

27

Hi, facing same issue here…performing a fresh install today. What is the stable version for Hassio which I can avoid such issues?

28

Just want to pile on here. Exactly the same problem. I did a fresh install of HassOS a couple mouths ago and this suddenly happened. The only container that got started after reboot was the observer container and nothing else. I could manually start up all the other container but ssh still doesn’t work. After starting the the homeassistant container, I can get to the ui but cannot get to the supervisor to do backup/snapshot. I am fairly familiar with linux and docker and I couldn’t find anything wrong. Before I reinstall, can anyone point me to things that I can check. I don’t find any log files that show any obvious error. Because ssh doesn’t work, I cannot cut and paste easily into here. My version is 4.12 from /etc/os-release.

I got quite a few of those “A start job is running for …” message but eventually, it said “OK” for started the Docker Application Container Engine

29

What ended up solving my issue:

uninstall Docker, enable Virtualization in Virtualbox Settings(you may have to do it in command line)

I also found it helpful to enable AVX and AVX2, which doesnt seems to be enabled by default.

30

Maybe it is a timing issue. I tried again with a fresh install for the third time with the 5.3 vdi, and this time it worked. At least Docker did not fail.

31

In my case i got /etc/docker/key.json file corrupted after upgrade from 5.10 to 5.13. Strange, but fact.
(I have RaspberryPi 4 installation with USB SSD)
How I found it:

  1. Boot from SD made like discribed here (Raspberry Pi - Home Assistant) - if you have Linux installation or VM with linux you can skip this point & connect SD/SSD from Pi to your computer.
  2. Mount /dev/sda8 (ha-data partition mounted in /mnt/data on regular runs) to some temporary folder like /tmp/a8
  3. Open journal to find problem by running "journalctl -D /tmp/a8/logs —since ‘1 hour ago’ | more» - as i told before my problem was in corrupted file
  4. Replace broken file from fresh SD card - in my case needed file was on device /dev/sda7 (/etc/docker), so i mounted it to /tmp/a7 and so on

Hope it could help someone :slight_smile:

1 Like
32

This fixed it for me.

33

Hey @samspycher , glad it helped! Hope you can help me fix it aswell :smiley: . How do I move the local-kv.db file since i can’t access it through the interface?