My Home Assistant / Hass.io install was an old ResinOS image from the hass.io website running on an Intel NUC.
Occasionally I would have to reboot the machine, as it would loose network connectivity for an unknown reason (The machine was still connected, and the host OS, but Home Assistant could not connect out to the internet? as in, the logs were full of time outs and refused connections due to ‘max retries exceeded’
I wanted to migrate this over to a Ubuntu Docker install so that I had better control over the hardware and could do other tasks on the machine, and to hopefully repair the issue with connectivity.
I took a snapshot of hass.io through the web UI, copied it off as well as manually backed up the files, and then installed Ubuntu 18.04.3 onto the NUC.
I installed Hass.io following the manual install directions on the website, and then the samba addon. Copied the file back into the correct folder and rebooted.
Once it was back up, I selected the snapshot and hit restore.
It took about half an hour to complete for some reason, but once it had, it was pretty much exactly how it was before I started.
There was a few things that required attention, for example Node-Red was not connecting. I didn’t do anything to it but restarting and it connected once more.
There was also problems with MariaDB that I couldn’t understand, it could not connect, but was listening and running fine? I ended up just removing this addon and reinstalling it, as it was pretty simple.
Once it was all up and running again, It seemed to be working fine. I setup Portainer manually and connected it to the local docker instance, and added my other docker endpoints to it.
All was good.
Until the next day. My alarm didnt go off in the morning, which was the first red flag, I tried opening Home Assistant to adjust my Air Conditioner and it would not load, I checked Portainer and the container was still running, so it should be working? Checking the logs it looks like its running but cannot connect to the internet again. This is similar but not exactly the same messages as previously. I restarted the Home Assistant container and the logs indicated it was having connectivity problems out to the internet, but this should not stop the Local UI from working on https://IP:8123 … yet it was? I had to leave for work at this point so I figured no biggie, ill fix it on lunch over VPN.
(update: the alarm didn’t go off because the MP3 it plays is hosted on the internet, I need to switch it over to a local MP3 to remove this internet dependency)
But heres where the plot thickens.
Just yesterday I was toying with settings on my OpenVPN Server with my buddy Tom while we were trying to figure out why DNS doesnt work for local clients when connected to VPN, and something we touched must have been wrong as now my VPN is playing up and apparently he is having internet troubles too.
The next day:
So it turns out I didn’t break anything, Telstra had messed up my account for some reason and after an hour talking to them they were able to revert their changes and fix it. Now everything’s back to normal!
2 thoughts on “Upgrading from ResinOS to Ubuntu”
Thanks for sharing your troubleshooting steps. Glad you found the culprit. Have you had to set up internal DNS for your local clients or are you simply using public dns such as google (184.108.40.206)?
My pfsense does internal DNS and passes 220.127.116.11 and 18.104.22.168 as the DNS via dhcp to clients !