Timeout while contacting DNS servers when forwarding resolver has broken EDNS cookie handling with newer c-ares versions

[RedNo7]

The problem

I have been getting this error in my logs multiple times/day recently and have no clue what’s causing it.
Sometimes it’s an issue with hacs.zyz, sometimes registry.hub.docker.com and sometimes others integration websites.

I haven’t knowingly changed any network settings, but it feels like it started with 2025.5

I’m running g HA Container 2025.5.2 (Docker) on a Pi4.

This may be a problem wit the aiodns library

What version of Home Assistant Core has the issue?

core-2025.5.3 (but may have started with core-2025.5.0

What was the last working version of Home Assistant Core?

core-2025.4.4

What type of installation are you running?

Home Assistant Container

Integration causing the issue

aiodns library

Link to integration documentation on our website

https://www.home-assistant.io/more-info/unsupported/dns_server/

(not sure if the above is correct)

Diagnostics information

No response

Example YAML snippet

Anything in the logs that might be useful for us?

2025-05-27 10:31:19.876 ERROR (MainThread) [plexwebsocket] Websocket connection failed, retrying in 15s: Cannot connect to host 192-168-1-64.3d6760daf0854d4693b06152345e1004.plex.direct:32400 ssl:default [Timeout while contacting DNS servers]
2025-05-27 10:31:36.415 ERROR (MainThread) [plexwebsocket] Websocket connection failed, retrying in 30s: Cannot connect to host 192-168-1-64.3d6760daf0854d4693b06152345e1004.plex.direct:32400 ssl:default [Timeout while contacting DNS servers]
2025-05-27 10:32:07.951 ERROR (MainThread) [plexwebsocket] Websocket connection failed, retrying in 60s: Cannot connect to host 192-168-1-64.3d6760daf0854d4693b06152345e1004.plex.direct:32400 ssl:default [Timeout while contacting DNS servers]
2025-05-27 10:33:09.454 ERROR (MainThread) [plexwebsocket] Websocket connection failed, retrying in 120s: Cannot connect to host 192-168-1-64.3d6760daf0854d4693b06152345e1004.plex.direct:32400 ssl:default [Timeout while contacting DNS servers]
2025-05-27 11:06:21.579 ERROR (MainThread) [homeassistant.components.version] Error fetching version data: Error fetching version information from container Cannot connect to host registry.hub.docker.com:443 ssl:default [Timeout while contacting DNS servers]
2025-05-27 12:16:22.363 ERROR (MainThread) [homeassistant.components.version] Error fetching version data: Error fetching version information from container Cannot connect to host registry.hub.docker.com:443 ssl:default [Timeout while contacting DNS servers]
2025-05-27 12:36:23.286 ERROR (MainThread) [homeassistant.components.version] Error fetching version data: Error fetching version information from container Cannot connect to host registry.hub.docker.com:443 ssl:default [Address not available]
2025-05-27 13:21:24.738 ERROR (MainThread) [homeassistant.components.version] Error fetching version data: Error fetching version information from container Cannot connect to host registry.hub.docker.com:443 ssl:default [Timeout while contacting DNS servers]

Additional information

Here is a post on it on the forums
https://community.home-assistant.io/t/timeout-while-contacting-dns-servers/894493/1

[Paja-git]

The same experience. Maybe @bdraco can help.

[richie510]

I am having the same issue. I noticed because some integrations that require internet for their data are periodically becoming unknown and for each one that becomes unknown I get the same DNS error.

[fevo21]

Same here on a docker installation (linux server). I have multiple healthchecks (rest command) running and it started reporting issues after updating to 2025.5. Typically the following error is logged:

Error fetching data: Cannot connect to host hc-ping.com:443 ssl:default [Timeout while contacting DNS servers] Error fetching data: Cannot connect to host hc-ping.com:443 ssl:default [Network unreachable]

One healthcheck ping is triggered every minute and it fails every 5 minutes on average.

[renaudallard]

I had the issue where my DNS server was not responding to all requests due to some problem too complex to describe here entirely (in short, EDNS cookies were used and aiodns received a bad cookie and thus invalidated the answer). But it causes the same issue. It seems after some time, it breaks everything based on aiodns, effectively killing stuff like "in app notifications", HACS, netatmo and probably others.
It's actually quite great someone also noticed this blocking bug.

[charlescwcooke]

I also have this issue. I'm not using Docker.

[RedNo7]

I'm running HA Container (docker) on a PI.

Just downgraded to 2025.4.4 and can confirm the problem doesn't exist prior to 2025.5.

[Paja-git]

Good to know @RedNo7. The question is, if somebody from code maintainers read this issue section. The issue is not very common.

[jvandijk]

I concur with how @renaudallard experiences the issue. I also commented in the thread on the community forum. Sharing here for visibility as well.

I’m using the docker version of Home Assistant and can definitely say this has something to with the aiodns library. I experienced this live and immediately checked the host and the container. Both were perfectly able to resolve any DNS query that was failing for all kinds of integrations, but HA couldn’t resolve them.
I think this behavior started with version 2025.5, but definitely experience it in .1, .2 and .3

[jvandijk]

I've also tried to run the 2025.6 beta version, but since nothing changed on the aiodns library version there I experience the same problems with that version. I'm also downgrading to 2025.4.4 for now, because the situation becomes unworkable for me.

[RedNo7]

same. I'm sticking with 2025.4.4 but that is also unsatisfactory.

...I'm not sure how to get a dev to look at this...

[fevo21]

I've also tried to run the 2025.6 beta version, but since nothing changed on the aiodns library version there I experience the same problems with that version. I'm also downgrading to 2025.4.4 for now, because the situation becomes unworkable for me.

What makes you think it is related to the aiodns package? I started two new 2025.5.3 test containers, downgraded the aiodns package on one container to 3.2 (comes with 2025.4.4). The automation which triggers the healthcheck ping every minute runs on both containers now, but so far no error logged on either container.

So with a clean container I cannot reproduce the issue I have on my main installation (which is still giving me errors every ~5mins).

How is your docker network configuration looking? The test containers are just bridged networks, but my main home assistant installation has several networks connected to it (macvlan, bridged, ..)

[RedNo7]

Only because someone mentioned it as a possibility in the HA forum here

[jvandijk]

@fevo21 I don't have proof that it's the aiodns library unfortunately. But it was the biggest DNS related change between 2025.4 and 2025.5, hence my line of thought. But by your library downgrade test you might have proven me wrong here.

I'm not doing any fancy docker networking stuff. Just plain setup with home assistant container running in network_mode: host.

Continuing the investigation....

[fevo21]

Well with my test I can't rule out aiodns either, because I am not able to reproduce the issue with just the plain installs. I might have time tomorrow to downgrade the aiodns on my main installation to see if that gives any insights