My supervisor also died this morning after I did a configuration adjustment and had to restart HA. I run HA supervised on a Debian 10 VM on Proxmox because I use a multihomed network and need the MQTT server accessible on other VLANs (there is probably a new / better solution, but I’ve been running this setup for several years now and I am reluctant to change it). Rebooting the VM did not help, supervisor kept restarting every 5 seconds with “CGroup” error.
Upgrading docker to 20.10.17 solved it (was on 19.03.8 previously, did apt update / upgrade, but had to manually upgrade the docker-ce package that “had been kept back”).
I looks like I was running supervisor version “supervisor-2022.06.dev3004” somehow - I usually turn all automatic updates off exactly because of situations like this, but I suppose the supervisor uses an independent setting; now it is asking me to update to version “supervisor-2022.06.2”. Did a development version of the supervisor accidentally reach the production channel?
looks like during my docker upgrade something happened because the docker has been updated but after reboot, there are no dockers. As I checked all containers exist with images… but the question is how to start all dockers? Is it any script for that?
Unfortunately restarting the host does not always work to restart all docker containers.
and no pre-existing scripts that i know of.
But if you want you can try this on a linux machine: for x in $(docker ps -a |grep -e 'homeassistant\|home-assistant' |awk '{print $1}'); do docker start $x; done
NOTE: still will start almost all containers related to HA Core/Supervisor and addons
and there is nothing dangerous in the command above.
Thank you. My case was a bit different. I had a custom docker configuration. The root dir was changed in past. After the upgrade, the new version wanted to be based on the default one. That was an issue.
I updated the configuration again and the for x in $(docker ps -a |grep -e 'homeassistant\|home-assistant' |awk '{print $1}'); do docker start $x; done runned all dockers ah… night has been saved.
I am having a similar issue as above, using Home Assistant Operating System on a raspberry pi 3. I cannot update the docker version since the entire / directory is readonly. What is the process for updating docker, or getting the supervisor to run again (I tried manually pulling, and triggering the /usr/sbin/hassos-supervisor to no avail).
I’m running HAOS in vmware ESXi on an Intel NUC. Same issue as others - can’t get the supervisor to start. Looks like i’m running an older version of docker (19.x). How do I update the docker version when running HAOS?