If you read what I wrote in the feature enhancement, checking the “Host Network” option does not work for Homebridge. Several of us have tried it and it fails. When I check the Host network option, Homebridge goes into a restart loop where it keeps trying to initialize, but can’t bind to the network and restarts, then repeats over and over. If I leave the Host Network option unchecked, then it properly starts up, but binds to the 172.16 network, which lets it phone home for updates and such, but can’t talk to any devices on my network. I also can’t access the Homebridge GUI from the browser as you need to be able to do. Under Dragonfish and prior, selecting the “Add external interface” resulted in an app where I had both the 172.16 and my internal IP (192.168.68.152) as available network devices. In that scenario, I can go into the Homebridge settings inside the GUI , and then tell it to advertise on my internal IP, which then results in everything working.
Maybe it is a lack of knowledge, but hopefully if so someone can point those of us having issues into the correct direction.
Hey everyone,
I cannot take credit for finding the solution, but I can confirm that I was able to get my old Homebridge install working using this configuration.
The homebridge team has updated the Docker instruction for post TrueNAS 24.10.0 installs and following those instructions worked. This was posted 4 hours ago. I took their barebones docker compose and combined it with the earlier version posted in this thread. This included passing through my Intel ARC GPU. Hope this helps everyone get their configs working. I have posted my compose file for reference below.
One more tip: If you already have docker network that you used to use for your homebridge as a bridge then you will want to remove it. You can check what networks are created by going to the Shell or SSH if enabled and typing:
docker network ls
You can then remove it using
docker network rm {network name}
This was causing me issues initially as I was getting errors that there was an overlaying network.
networks:
homebridge_net:
driver: macvlan
driver_opts:
parent: br01 # I have a bridge interface set up due to VM limitations. This should either be your NIC or the bridge using your NIC.
ipam:
config:
- gateway: ###.###.###.### #gateway IP
subnet: ###.###.###.###/24 #subnet using CIDR notation
services:
homebridge:
container_name: homebridge
deploy:
resources:
limits:
cpus: '2.5'
memory: 3072M
devices:
- /dev/dri:/dev/dri #passing through Intel QSV Card
environment:
- PUID=568
- PGID=568
- TZ=America/Los_Angeles
image: homebridge/homebridge:latest
logging:
driver: json-file
options:
max-file: '1'
max-size: 10m
networks:
homebridge_net:
ipv4_address: ###.###.###.### #IP Address of your homebridge server
restart: unless-stopped
volumes:
- /mnt/pool/dataset/homebridge:/homebridge
Here is a reference for the likely passthrough parameters needed for NVidia as well
This thread is the closest I’ve come to seeing the issues I’m having with homekit bridge on home assistant app (official install. I can’t turn off mDNS because it’s used by TN for time machine. I just want to pass through port 5353 to the host, or set the driver to host on the official docker container, but I can no idea where that is. I would have to imagine a very large portion of Home Assistant users on TrueNAS Scale (EE) will be needing homekit bridge, and the fact that my home network can’t see it running behind the docker 172.x.x.x network has caused me to read through dozens of threads without any realy solution. Can’t mDNS and Homekit Bridge play nice on port 5353 together? I even have a 2nd NIC that I would bind the docker container to if I knew how, which hopefully would free up port 5353, but it seems that TrusNAS scale just runs everything x2 on both NICs anyways.