HomeAssistant VM broke down without real cause

Installation
1

Hey guys,

I am using a VirtualBox VM to run HomeAssistant and last night it just broke down without a reason.
The VM was still up but HomeAssistant was not reachable.
I looked through the logs for the host and core but did not find anything useful beside this in the host log:

2025-02-19 23:40:00.538 homeassistant NetworkManager[457]: <info>  [1740008400.5375] dhcp6 (enp0s3): state changed new lease, address=<ip-v-6>
2025-02-19 23:40:01.149 homeassistant dockerd[560]: time="2025-02-19T23:40:01.066985357Z" level=error msg="[resolver] failed to query external DNS server" client-addr="udp:<address1>:47082" dns-server="udp:<address-of-dns>:53" error="read udp <address1>:47082-><address-of-dns>:53: i/o timeout" question=";version.home-assistant.io.\tIN\t AAAA"
2025-02-19 23:40:01.155 homeassistant dockerd[560]: time="2025-02-19T23:40:01.152822960Z" level=error msg="[resolver] failed to query external DNS server" client-addr="udp:[<ip-v6-1]:34836" dns-server="udp:[<ipv6-of-dns>]:53" error="read udp [<ip-v6-1>]:34836->[<ip-v6-of-dns>]:53: read: connection refused" question=";version.home-assistant.io.\tIN\t AAAA"

So obviously it couldn’t reach my DNS but I do not know why. I also do not get why it broke everything. There were no other logs from that point on neither for the host nor for the core.
Only a restart of the VM solved the problem.

Did someone have a similiar error?

2

The first line indicates you are running your server with a DHCP address and that it requested a new lease period.
Maybe the IP address changed.

3

Yes the HomeAssistant VM has a fixed IPv4 Adresse but no fixed IPv6 and that address most likely changed but how can this lead to the VM choking so hard that I had to manually restart it?

With restarting it should have gotten either a new IPv6 address or kept the same. If it kept the same I do not understand why it choked in the first place and when it got a new one that would imply to me that there are “good” and “bad” IPv6 addresses which I also do not understand.
Is that even possible?

4

IPv6 do not have good or bad addresses, but they have temporary random addresses made to make tracking harder.

The question is if you connect through an IPv4 or IPv6 address here.

5

It should be IPv4 as my reverse proxy resolves to that. But HomeAssistant was not reachable even when I typed the IPv4 into the browser search bar.

I could connect to the VM via ssh but restarting the core from there did not change anything but only took forever.

But I do have severe stability issues with HomeAssistant since upgrading to 2025.02. Every patch update I do breaks something so I need to manually fix something.

6

Check the logs again and make a notice of the timestamps.
If it starts around the freeze, then look into homeassistant.log.1 in your config folder, because then your HA core have restarted.

7

Okay, I will try that once I gained access to HomeAssistant again. Right now it is not reachable at all after restarting the VM.

The Web frontend told me that core version was 2025.02.04 and the ssh connection to the VM told me that core version was 2025.02.05.
After a restart it is not reachable at all.

Will get back to you :smiley:

EDIT:
I do not have time right now as I need to fix my HomeAssistant in general before being able to look further into this error. As this is likely another problem I will open a different thread but will link it from here in case there is an unexpected connection between those two problems.
New thread here

EDIT2:
Ah fuck now I get why I should’ve saved the log right away. With restarting the VM and so on the log is obviously gone.
Will keep an eye out for the error, maybe it happens again.