Intel NUC device goes offline after balenaOS update

After updating a remote Intel NUC device from balenaOS 2.48.0+rev3 to balenaOS 2.50.1+rev1, the device went offline. Sometimes the device gives the api heartbeat true status after visiting the device diagnostics - supervisor state page, but the api heartbeat goes back to false afterwards (seems like a little bug).

The device api still reports:

  "provisioning_progress": 50,
  "provisioning_state": "Running OS update",

Any suggestions? We have had some other Intel NUC devices going offline after an update in the past, having to execute a power cycle to solve the problem. But that is not always possible without manual intervention.

Thanks in advance.

Hi @thiver,

Have you run device diagnostics on the device from the balenaCloud dashboard? That might offer some clues you could share with us. Are your NUC devices on wifi or wired?

Thanks,
John

1 Like

Hi @jtonello,

I can’t run the device diagnostics because the device is offline. The device goes to “online - heartbeat only” state every time I go to the supervisor state diagnostics page. So I don’t know this is a glitch or the device is still connecting to the api. The NUC has a wired ethernet connection.

If the “online - heartbeat only” state is showing that means the device is connected to the API, but it means that the VPN is not starting up correctly. I think the best way to debug this would be to set up a lab device (local to you) provisioned into your app with the “development” version of balenaOS 2.48 set up on it, so that you have kernel logs on the HDMI out and you have simple SSH access to the device locally. Once we have that set up we can trigger the hostOS update and see what is going on. If the device is continually rebooting we will see that in the kernel logs on the hdmi, but if it is booting correctly but its just the vpn failing, then we should still have SSH access and we can get in a pull out a bunch of useful logs to help us debug. Do you think it will be possible to set up a test like that?

Does anyone have experience getting piHole to load through balena? I can’t seem to get it to work, doesn’t seem to do anything. If I go to it’s ip it just fails, but I can ping the pi. I’m running a pi Zero W if that’s makes a difference.
https://nox.tips/ https://xender.vip/ https://testmyspeed.onl/

Hey there, this thread is not piHole related, can you please open a new forum thread if you are having issues? For reference, here is the repo you should be using: https://github.com/klutchell/balena-pihole and you can go through the issues there and report if something is not working for you. https://github.com/klutchell/balena-pihole/issues/17 this issue looks relevant in this case.

Just following up here that we’ve released an update that addresses the issue referred to above of the device status changing after visiting the device diagnostics - supervisor state page.

Does anyone have experience getting piHole to load through balena? I can’t seem to get it to work, doesn’t seem to do anything. If I go to its IP it just fails, but I can ping the pi. I’m running a pi Zero W if that makes a difference.
https://alightmotionapks.com/ | https://www.kinemasterapks.com/kinemaster-mod | https://www.pandahelpers.com/

Hello, as mentioned by a colleague earlier in this thread, this thread is not piHole related. Can you please open a new forum thread if you are having issues? You can also go through issues or report an issue at the following repo: GitHub - klutchell/balena-pihole: Pi-hole is a Linux network-level advertisement and Internet tracker blocking application.

Per the following issue from that repository: Does not work for rpizero · Issue #17 · klutchell/balena-pihole · GitHub , it appears that Support for Raspberry Pi Zero is in Dockerfile.rpi so by renaming it to Dockerfile it should work as expected.