Boot problem caused by setting HA-OS with Adguard as default DNS in network

So this took me a while to find out what exactly happened since it was running fine for a long time. By sharing this, I hope to save people time (and a lot of frustration).

I sort of created my own ticking problem, waiting for me to show up:
So after the install of Adguard, I wanted to let all my devices make use of this by default in the network. So i reprogrammed the dhcp-server to give the address of my HA as DNS server, so all traffic is filtered on new and existing devices that use dhcp-server.

I was a bit sloppy with updating my OS lately, so with the new year coming up I decided to pick up these tasks again.
But with the newest OS I found my pi4 not booting anymore. Oh no, of course I figured this had to be related to the update.

What i did finally discover by this topic, link, was that the time was not up to date because it could not reach the ntp server. Did not take me long to realise this was caused by the inactive Adguard, since this was run by the Pi itself and the dns-address it got on boot by the router. And then it runs in all kinds of trouble. By changing the dns-address on the router, the problem was fixed.

So when using a similar setup, do not forget to set a static DNS server on the pi itself.

What I did find worrying is that without internet, the HA-Os completely refused to start with a new boot, also locally so would it be an idea to have another look at this? Or did I miss something in my configuration that it refuses to start without internet?

Tags (in sentence): When the pi can not sync its time and the time is out of date, it will not start Home assistant OS. It will boot the kernel, so one can login using the root account when connecting a screen and keyboard. From there it will say the CLI is not running so using emergency console.

1 Like

Please don’t do that; you come across as spammy.

Does anything show up in the logs, or were they erased on a successful boot? Can you reproduce this, and then post any logs?

1 Like

The tags were intentional and not meant as spam, since one would not find this with the search string “emergency console” or “CLI”.
I have reproduced it multiple times by now, so quite sure it was because it could not sync the clock. Also confirmed this with the date, which was 8 days behind the real time.

If you post logs, then maybe this issue could be fixed. Also turn on the share reports setting.

1 Like

Thank you for the comments. I will follow up on this with more tests and hopefully some logs that make sense :slight_smile: