Just upgraded to 2022.2 and I’ve started to get a warning in my notifications panel indicating:
“ cannot reach None, falling back to polling, functionality may be limited.”
where winds up being one of my Sonos speakers (note it always says “None”). When I download the diagnostic log from the sonos integration it says:
“subscriptions_failed”: true
for one of my speakers but not for any others. The speaker involved in the message seems to be random and it will change on an HA restart.
I have no idea what could be causing this as my Sonos and Home Assistant installs are on the same VLAN / Subnet and there nothing blocking the communication between them.
I’m running Home Assistant OS so I don’t think there’s anything else that needs to be configured. Anyone have an idea what might be causing this?
Hi, did you get figure out this one. I have the same problem all of a sudden after upgrading. Although I have opened that port 1400 to my home assistant.
+1 here. Set up a firewall rule to explicitly allow port 1400 from my sonos devices to my server and it still falls back to polling mode immediately.
Also tried manually specifying the sonos IPs and also using my server IP as the advertise_addr in configuration.yaml, as suggested in the Sonos Integration article.
I’m using the homeassistant/home-assistant docker container in Unraid.
Any chance you’re running multiple HA instances on the same box (e.g., for testing or development)? It will listen on port 1400 by default, but if that’s already in use it will increment the port number until a free one is found.
The log message will also tell you this port, along with the callback IP that’s being communicated to the Sonos devices. That’s what advertise_addr changes. Make sure that’s routable and unblocked from the Sonos devices.
Is it possible for you to sniff the traffic hitting the firewall from the Sonos devices? Some firewalls make that task easier than others.
In my case, only one HA instance and I have all Unifi networking gear - not exactly sure how to use that to tell if the firewall is letting the Sonos packets through (maybe something with Wireshark?).
Would it be relevant that my server IP and my HA IP are totally different due to HA running in docker? To test, I added an environment variable to map through port 1400 from the host to the HA container, and that didn’t fix it.
Also my logs are referencing the advertise_addr and showing the manual change that I made to put server IP as the advertise address. It was defaulting to the docker network address (a 172.x.x.x address) so I changed it. It fails either way.
GarageSonos cannot reach None (advertising as [serverIP]:1400), falling back to polling, functionality may be limited, see https://www.home-assistant.io/integrations/sonos/#network-requirements for more details
OfficeSonos cannot reach None (advertising as [serverIP]:1400), falling back to polling, functionality may be limited, see https://www.home-assistant.io/integrations/sonos/#network-requirements for more details
and
Creating subscription failed for OfficeSonos: 412, message='Precondition Failed', url=URL('http://[deviceIP]:1400/MediaRenderer/AVTransport/Event')
Creating subscription failed for OfficeSonos: 412, message='Precondition Failed', url=URL('http://[deviceIP]:1400/MediaServer/ContentDirectory/Event')
Creating subscription failed for OfficeSonos: 412, message='Precondition Failed', url=URL('http://[deviceIP]:1400/DeviceProperties/Event')
Creating subscription failed for OfficeSonos: 412, message='Precondition Failed', url=URL('http://[deviceIP]:1400/MediaRenderer/RenderingControl/Event')
Creating subscription failed for OfficeSonos: 412, message='Precondition Failed', url=URL('http://[deviceIP]:1400/ZoneGroupTopology/Event')
Troubleshooting multiple VLANs and internal firewalls is a bit out of scope of what I’d like to offer help with. I currently do that in my own environment, but it’s definitely not a trivial configuration. I’d suggest to first try everything on one subnet and ensure the devices work in a simple setup.
Secondly, the Docker container should be running in host networking mode as per the HA installation instructions. Once this is done, the HA instance should be able to detect the host’s IP and set it properly without needing to use advertise_addr. I myself run my prod HA instance as a container (with host networking) and have no issues.
Finally, the “cannot reach None” errors (instead of logging an actual IP:port pair) imply that you may have some other networking issues where the Sonos devices are sometimes (temporarily) not reachable from your HA host during initial setup. The usual suspect is flaky wifi connectivity. I’ve recently put in a PR to better handle this scenario: https://github.com/home-assistant/core/pull/73456. Hopefully it will be included in the 2022.7 release.
