Go2rtc doesnt start after update to 2026.4.0

Installation Home Assistant OS
1

Im on Raspberry Pi 3. First i couldn’t update to 2026.4.0 from within the WebGUI. The update timed out every time. However from the cli, after logging in via SSH, the update worked. After the update, i restarted HA and now get the following error:

Logger: homeassistant.setup
Quelle: setup.py:278
Erstmals aufgetreten: 14:56:26 (2 Vorkommnisse)
Zuletzt protokolliert: 14:56:26

* Setup failed for 'go2rtc': Integration failed to initialize.
* Setup failed for 'default_config': Could not setup dependencies: go2rtc

It seems there is some timeout:

2026-04-03 14:56:26.193 ERROR (MainThread) [homeassistant.components.go2rtc.server] Go2rtc server didn't start correctly
Traceback (most recent call last):
  File "/usr/src/homeassistant/homeassistant/components/go2rtc/server.py", line 223, in _start
    await self._startup_complete.wait()
  File "/usr/local/lib/python3.14/asyncio/locks.py", line 213, in wait
    await fut
asyncio.exceptions.CancelledError
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
  File "/usr/src/homeassistant/homeassistant/components/go2rtc/server.py", line 222, in _start
    async with asyncio.timeout(_SETUP_TIMEOUT):
               ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.14/asyncio/timeouts.py", line 114, in __aexit__
    raise TimeoutError from exc_val
TimeoutError
2026-04-03 14:56:26.210 WARNING (MainThread) [homeassistant.components.go2rtc] Could not start go2rtc server
Traceback (most recent call last):
  File "/usr/src/homeassistant/homeassistant/components/go2rtc/server.py", line 223, in _start
    await self._startup_complete.wait()
  File "/usr/local/lib/python3.14/asyncio/locks.py", line 213, in wait
    await fut
asyncio.exceptions.CancelledError
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
  File "/usr/src/homeassistant/homeassistant/components/go2rtc/server.py", line 222, in _start
    async with asyncio.timeout(_SETUP_TIMEOUT):
               ~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.14/asyncio/timeouts.py", line 114, in __aexit__
    raise TimeoutError from exc_val
TimeoutError
The above exception was the direct cause of the following exception:
Traceback (most recent call last):
  File "/usr/src/homeassistant/homeassistant/components/go2rtc/__init__.py", line 175, in async_setup
    await server.start()
  File "/usr/src/homeassistant/homeassistant/components/go2rtc/server.py", line 190, in start
    await self._start()
  File "/usr/src/homeassistant/homeassistant/components/go2rtc/server.py", line 229, in _start
    raise Go2RTCServerStartError from err
homeassistant.components.go2rtc.server.Go2RTCServerStartError: Go2rtc server didn't start correctly
2026-04-03 14:56:26.229 ERROR (MainThread) [homeassistant.setup] Setup failed for 'go2rtc': Integration failed to initialize.
2026-04-03 14:56:26.230 ERROR (MainThread) [homeassistant] Error doing job: Unclosed client session (task: None)
2026-04-03 14:56:26.231 ERROR (MainThread) [homeassistant.setup] Unable to set up dependencies of 'default_config'. Setup failed for dependencies: ['go2rtc']
2026-04-03 14:56:26.232 ERROR (MainThread) [homeassistant.setup] Setup failed for 'default_config': Could not setup dependencies: go2rtc
2026-04-03 15:28:26.574 ERROR (MainThread) [homeassistant.components.fritz.coordinator] Service/Action Error: cannot execute service WLANConfiguration with action GetInfo

Is there anything, i can do?

2

If everything else is running, try reloading the integration ( frigate or anything else?)

3

If you are trying to run cameras on a PI3, you will almost certainly run out of memory, forcing your already slow system to use lots of swap space on your drive, which slows stuff down even more. If you are also using an SD card, the swap space usage will wear out the card incredibly fast.
You need a different host computer, bottom line.
(Why I don't recommend SBCs for running servers · GitHub)

4

I understand what you mean. With most of what you have written on GitHub, you are absolutely right. However, i run my (private) servers since more than 8 years on SBC’s without much issues. I never had a corrupted SD card during this time. But, and thats the main point, i modified the Debian on this servers a lot. There is a absolute minimum write processes to the SD card. Most is written to SSD’s or to tmpfs (as long useful). However, this is only a small nginx and PHP with 4GB. MariaDB is on a separate RPi with 8GB and a Mailserver on another RPi. Specially the mailserver was not without issues, specially in regards of SPAM marking and so on.

A Home Assistant is completely new for me. This is not a standard Debian, this is heavily modified and i really hate docker and co. But you’re right—I hadn’t even thought about the RAM requirements for the streams. Sure, with only 1GB, the swap space is going to be dancing around :-). I know that the SD card is not going to be happy with that. The current setup is just for testing for now, to get some initial experience. I just had the RPi 3 lying there unused. If everything works, I’ll switch to an RPi 5 with an SSD. For now, I’ve removed the Ring integration and restarted. But the error message won’t go away, even though the RTC isn’t needed anymore. I suspect that after the reboot, the startup process simply takes too long and exceeds the timeout.

Is there a way to extend the timeouts somewhere? Is that in a config file I’m not familiar with yet?

5

I do not use go2rtc for anything, and my raspberry pi 3 was also almost frozen, i was able to restore a 2026.2 backup and works perfect

Edit: I opened an issue Raspberry pi 3 does not work after 2026.4 (go2rtc) · Issue #167398 · home-assistant/core · GitHub

6

I don’t know if this helps. But today I flashed the same image as on my previous install, updated to 2026.4.1 and restored the backup. This took some while, specially the update to 2026.4.1, but after that, the error was not seen anymore. All apps, devices and automations are fully working.

I don’t know how the data is restored, specially for the sqlite database, but I have the feeling that something was corrupted in the database. If it is restored with inserts and not by copying the file itself, this would "repair " a otherwise corrupted db. But this is only a wild suggestion.

BTW: is it possible to use mariadb as backend on a separate server? I only ask because I use a dedicated RPi only for mariadb, for nextcloud, roundcube and some other things.

7

You appear to have taken cluster computing to extremes. Maybe go the other way and consolidate.

Maybe something with a bit more grunt and lots more memory and throw all your separate devices onto it as virtual machines?