Closed taglia closed 7 months ago
I ran into this issue, as well. I have not yet figured out why Home Assistant is constantly advertising new names.
I did find that disabling mdns discovery in nextdns helped out: nextdns config set -mdns disabled
.
If there were anything at fault with the nextdns CLI, perhaps mdns records should expire after some amount of time? At the moment, it appears that nextdns holds on to all mdns hosts forever.
I have a rather large home assistant installation with a UDM SE running the CLI at home and can’t reproduce your issue.
The issue seems to present itself when:
https://github.com/home-assistant/operating-system/issues/1163 seems to explain the reasoning behind the constantly changing hostnames:
homeassistant.local
) on VLAN1homeassistant.local
, and comes to the conclusion there is some other host on the network using that name.homeassistant12345.local
) and repeats the process (and on and on).It seems like this could happen with any multihomed system running systemd-resolved... It's probably more common with home assistant because folks try to isolate IOT endpoints.
As a workaround that would allow me to continue using NextDNS mdns discovery AND Unifi Multicast forwarding, I set up Firewall rules on my UDM pro SE that drop all incoming mdns traffic from Home Assistant network interfaces, preventing unifi from forwarding its mdns queries which caused HA/systemd-resolved to believe there was a conflict.
Since HA is already on multiple VLANs, we don't need to rely upon Unifi to forward it's mdns queries, anyway.
Multicast DNS
with port 5353.Can you please test with this binary and tell me if it foxes your issue?
NEXTDNS_VERSION=pr-903/SNAPSHOT-4554031 sh -c 'sh -c "$(curl -sL https://nextdns.io/install)"'
Can you please test with this binary and tell me if it foxes your issue?
NEXTDNS_VERSION=pr-903/SNAPSHOT-4554031 sh -c 'sh -c "$(curl -sL https://nextdns.io/install)"'
I'm still seeing the issue with that build. I've added a comment onto your PR https://github.com/nextdns/nextdns/pull/903/files#r1451796043
Context
Description I have recently noticed that the CPU usage on my USG would rapidly increase, to the point where I start having network issues unless I restart NextDNS on the USG (the memory usage does not increase noticeably though). The problem appeared when I started running Home Assistant (I am running Core version 2023.11.2). I suspect this is linked to Home Assistant also because, if I run
nextdns discovered
, the output is flooded byhomeassistant???????.local.
records. Roughly one hour after anextdns restart
, I found 17,760 such records:This only happens on the subnet I use for IoT devices, the other IP addresses that Home Assistant uses do not appear, so it could be linked to some particularly chatty integration.
Config
I am using two different NextDNS profiles (which is the only config I customized), but I already tried to remove one, and it did not make any difference about this problem. I have not seen anything abnormal in NextDNS logs.
For now I will uninstall NextDNS from my USG, but I would be happy to do more investigation if that is useful.