Instead of using a locally-connected zwave device (usb stick/etc), we can use a serial device mapped over the network with ser2net and then map it to a local zwave serial device with socat.
Start the image with the environemnt variables:
PAUSE_BETWEEN_CHECKS=2
In seconds, how much time to wait between checking running processes.
SOCAT_ZWAVE_TYPE=“tcp”
SOCAT_ZWAVE_HOST=“192.168.5.5”
SOCAT_ZWAVE_PORT=“7676”
Where socat should connect to - will be used as tcp://192.168.5.5:7676
SOCAT_ZWAVE_LINK=“/dev/zwave”
What the zwave device should be mapped to. Use this in your home assistant configuration file.
@quasar66 Just making sure I understand this correctly. This docker file builds an entirely new docker image including home assistant and the socat bits? Meaning if I already have a home assistant docker image I would have to replace it with this one? And when there is a new build I would have to rebuild it manually vs the built in update sequence?
You use this image instead of the homeassistant one (just replace the image name) and add the extra environment variables. If you build the image again, it has homeassistante:latest in it it will grab the latest home assistant.
I’ll try to set an automatic build with the lastest homeassistant build in the next week so you should be able to use the latest image automatically.
Sorry for reviving an old post, but I wanted to see about using this docker image since I am in the same boat (needing remote z-wave connection).
First, is this image based on the full HA implementation, or the HassIO version? I am curious because I want to keep the HassIO functionality (full plug-in support, automatic upgrades, etc).
Second, has the auto build process been completed?
Additionally, what steps will be done in the future to maintain the image (in case you decide to move on to something else)? If not a HassIO image, any plans on creating/supporting one?
This is for normal home assistant, I don’t want to touch HASS.IO.
You can download the git repo and do docker build -t and build the image yourself whenever you want. If you’re afraid of not having the files, fork the git repo or copy the files somewhere.
I had an automated pipeline, don’t know why it does not work anymore, I’ll try to fix it in the next weekend.
I’m not using this image anymore due to some docker issues I’ve had, but I’m still using the scripts that are in the image in a new virtual machine and they work just fine, haven’t touched them in more than half a year…
I could make an image that only mounts the serial device in the docker host but that would be useless without restarting docker… maybe I can also add an https call to restart docker ??
Anyway, if you want to use this in hass.io i would recommend running it as is, with a second home assistant instance, and publish everything to mqtt. Then, in the main home assistant, just import the devices from mqtt. This way zwave will be decoupled from main home assistant (that’s actually how I run things right now and have been for quite a long time).