HAOS random restarts when running on Proxmox

Installation Home Assistant OS
1

Hi All, was using home assistant OS in a virtualbox VM housed on my xigmaNAS previously for 2 years. Albeit slightly slow, it works perfect.

I am trying to migrate over to proxmox VM as I have more computing power on another newer machine N5105. I have successfully install proxmox and used the community guide script to install HAOS.

I am getting random restarts within HAOS, I could see from proxmox interface that it is still rolling the uptime, which implies the restart could be within HAOS. The uptime was indeed reset in HAOS.
EDIT: I found the logs home-assistant.log and home-assistant.log.1
However, there was no indication of anything unusual, a very typical checking of the TV state:

2022-07-19 08:04:34 ERROR (SyncWorker_1) [homeassistant.components.command_line] Command failed: binary_sensor.MiTV_state
2022-07-19 08:04:35 ERROR (SyncWorker_0) [homeassistant.components.command_line] Command failed: binary_sensor.MiTV_state

Then the log ended and I believe it restarted. At that moment I was in bed so it happened during a state where it was running fine undisturbed.

Does anyone encounter this on proxmox and could figure out what’s wrong?

2

I am observing this behavior for days in Proxmox 7.2 and with the last version of homeAssistant(2022.7.5) I do not know why this is happening… but it is exactly as you described, just random restart of HAOS I would say, but if I check the docker containers, actually those restarts are not counted by docker… So I start to think that it can be networking… not sure…

3

i am running Proxmox 7.1-12 with HAS 2022.7.4 without issue. Proxmox has been super reliable. I have a 4 node CEPH cluster running. I now wont be upgrading to 7.2 for a little while me thinks.

Watching with great interest.

4

Hi All, in the proxmox haos community installation guide thread it is shared that the fix is to install the edge kernel. It seems to fix the issue for me as well thus we can close this thread. Will monitor and report back in that thread if I encounter issues.

2 Likes
5

So after changing a bit the network config, assigning a fixed IP address, it seems that it’s not doing those restarts anymore … I will test it more in depth but it looks much better, today no restarts at all…

6

Can you please provide a link on how to do this? thanks

7
3 Likes
8

interesting, mind sharing your changes? I have used fixed IP from the start as well.

9

of course I don’t mind, but unfortunately I did no magic, I have just changed from DHCP to a fixed IP, nothing else… And still the 2nd day stable…

10

Did you set DHCP from within HAOS? Or set the allocation by the router?

11

DHCP of course was configured in the router, but I just have give via HAOS some fixed IP (not a fixed DHCP IP for the MAC on the router).

12

I used the proxmox helper script to install the edge kernel. So far it is acting much much better. Just in case someone else comes along and reads this. I am also using DHCP with a lease. Much easier to make network changes after the fact if using DHCP vs static configuration.

13

Hi,

just to let you know it started to do the same once more, and it seems that the frontend is restarting every 5-10 minutes, it can be really annoying… And anything I did it would solve it so I am going now to Supervised installation, at least I will have more control with the OS, and it seems that for the time being is not doing it.

14

I talked too fast …

Aug 24 13:14:10 supervised kernel: [ 2551.678612] hassio: port 7(veth2c972ed) entered disabled state
Aug 24 13:14:10 supervised kernel: [ 2551.679087] veth02d46b3: renamed from eth0
Aug 24 13:14:10 supervised kernel: [ 2551.726743] hassio: port 7(veth2c972ed) entered disabled state
Aug 24 13:14:10 supervised kernel: [ 2551.727993] device veth2c972ed left promiscuous mode
Aug 24 13:14:10 supervised kernel: [ 2551.727999] hassio: port 7(veth2c972ed) entered disabled state
Aug 24 13:15:59 supervised 92d3bc96b29f[361]: [13:15:59] [172.30.32.2][6bdd6c22][ExtensionHostConnection] <358> Extension Host Process exited with code: 0, signal: null.
Aug 24 13:16:00 supervised 92d3bc96b29f[361]: [13:16:00] [172.30.32.2][894303d0][ManagementConnection] The reconnection short grace time of 5m has expired, so the connection will be disposed.

This is exactly what I got in /var/log/messages before loosing the frontend for some time…

15

IT started to be really unusable, I tried the following with no success:

  • Removing HTTPS, and use just HTTP
  • Remove the telegram
  • Use DHCP or Fixed IP
  • Used HAOS or Supervised installation
  • Changed the network card type from Proxmox
  • Now I am trying not using Postgres for the recorder it didnt happen yet…

Edit: The postgres 14.4 installation is on a LXC running alpine…

16

Have you tried the PVE edge kernel?

17

Hi would like to update again that I’m still having the same issues of restarts. I’m not sure if there is a link that this happens if I utilise 4 cores (all) as by default the script to install uses 2.

I might want to try to migrate to esxi to see if it fixes. There is some discussion on N5105 and PVE.

18

I updated to the Edge kernel yesterday, and so far is going well, at least for now… but I desisted using HAOS, I went to supervised installation, so I could control myself the OS…

19

For total control, run Home Assistant Container. It’s the method I chose. I feel like the Supervised method is more trouble than it’s worth, now and in the future.

1 Like
20

uhmm I thought on that but that would be already too much control for my little time :slight_smile: but I though even in using a LXC for it but… uhmm Add-ons are just comfortable to me… :blush: