Title: Home Assistant Green – containerd keeps crashing, Core installation fails (HAOS 18.3)

Hello everyone,

I’m hoping someone can help me with a persistent problem with my Home Assistant Green.

I’m running Home Assistant OS 18.3, but Home Assistant Core won’t start or install properly.

The system keeps showing “Preparing Home Assistant”, and I cannot access my normal dashboard.

Main errors:

  • containerd.sock: connect: connection refused
  • containerd repeatedly crashes with exit status 2.
  • The logs contain go.etcd.io/bbolt errors.
  • The containerd restart counter reached 186.
  • Home Assistant Core remains on version landingpage instead of installing 2026.10.4.
  • Kernel messages include BUG: Bad rss-counter state.

What I have already tried:

  • Checked my Ethernet connection and internet connectivity.
  • Reinstalled/reset Home Assistant OS.
  • Recovered the Supervisor after an AppArmor-related startup problem.
  • Attempted ha docker reset-storage, but this failed because containerd was unavailable.
  • Created /mnt/data/docker/.wipe-scheduled and rebooted the device.
  • The same containerd errors continued after rebooting.

The Supervisor sometimes reports Connected, Supported and Healthy, but Home Assistant Core still cannot install.

I have external backups available.

My questions:

  1. Does this indicate corrupted Docker/containerd storage or possibly faulty eMMC hardware?
  2. How can I determine whether the scheduled Docker storage reset actually ran?
  3. What would be the safest way to repair this without losing my existing configuration?

I have attached screenshots showing the errors.

Any help would be greatly appreciated. Thank you!

Can you share the output of journalctl -fu containerd.service? Also check journalctl -rp warning.

  1. Maybe. Too little information information
  2. systemctl status haos-docker-reset.service might say
  3. Since you have backups you shouldn’t need to worry. You can try to reset via ha os datadisk wipe and if that doesn’t work via SD Card.