Update to HAOS 11 on odroid M1 fails - need to restore backup

Hi,
so far all upgrades on my Odroid M1 have been fantastic and all functionality of HA has made me very happy. So, today, when I saw the new HAOS 11 update, I thought I’d install it. Unfortunately, that was the last I saw from my HA. I got stuck on " Waiting for the Homeassistant CLI to be ready" and did not proceed for over 2 hours. I could not turn off some of the lights…

I need re-installation from backup, but apparently that is a partial back-up…and I cannot tell from the name of the backup what it is for. The only way to find out what it is is to go through the tar files. I guess the biggest one is the starting point. How can I restore everything (especially my zigbee)?

Hi,

Same issue here - Odroid M-1 and never had any problems till this update. After waiting 9 hours, I did a new install on 11.0 on my SD acrd which loaded fine. Then I did a restore form back-up and that took all night but eventually came back to life. Then I moved my data disk to the NVMe M.2 Internal SSD where itt was originally and its been 8 hours and has not come back to life.

Update: Gave up with 11.0, downloaded 10.5 and did a restore. Took aboit an hour to restore and get evrything back Two days wasted with 11.00.

I dont know how to “fix” this problem but i can tell you how to start homeassistant in this broken state after incomplete upgrade:

  1. Attach keyboard and HDMI monitor to your odroid M1
  2. After turning M1 on and waiting about 5 min you can press Alt-Ctrl-F2 to get to the system login prompt.
  3. Enter “root” in login, no password needed.
  4. see if your data volume is mounted. Mine wasn’t. “mount | grep data”
  5. If is not mounted - enter “mount /disk/by-label/hasoss-data-dis /mnt/data”

After that has will start and will work fine, but if you reboot you would have to repeat this all over again.

I don’t know what is the fix but i saw in the jounalctl that the job to mount disk-by label kept timing out.

EDIT: created a bug

Got it more or less restored on 11.0, but it never came back when I moved the data disk to nvme. So I am now trying to restore 10.5

I’ve unfortunately run into the same problem after upgrading my ODROID M1. Relabelling the data partition to hassos-data (from hassos-data-dis) did move Home Assistant closer to starting, with the observer container successfully starting. Unfortunately, that’s where it seems to stall and I’m not too sure how to move forward from here.

Rebooting the system results in the label being reset to hassos-data-dis, which once again causes HA to fail loading.

Does anyone have an idea how to move forward without starting over and restoring a backup?

Thanks to Victorsueca’s help on the GitHub issue, I was able to get HA booted into 11.0. The issue will come back on reboot but at least there is a way to get it up and running again until a permanent fix is available.

On boot/reboot and once you’ve gained access to the system console (Ctrl + Alt + F2), run the following command to correctly label the disk partition:

e2label /dev/nvme0n1p1 hassos-data

Wait a few seconds and then start the supervisor service:

systemctl start hassos-supervisor

3 Likes

Can confirm, this fix worked for me.

Thanks

Running 10.5 now, that is stable. I will not upgrade anymore until we have a more permanent solution :slight_smile:

11.1 version is out and the problem is fixed.
If you have broken install, remove “-dis” from the label name, mount it, install 11.1, reboot and everything will be fine.