Master HA instance with multiple slaves

Det er herved gjort. Igjen takk for hjelpen, hilsen Tor i Norge. :grin:

1 Like

All isn’t well, it seems. I can’t get some of my entities to show. This is what parts of my GUI looks like on the master:

I do have other automations and switches locally, so it’s not a lack of domains there. Any idea what it can be? I have also added this code to my configuration.yaml:

lock:
light:

Anything in the log this time?

I have this on the slave. But I didn’t think that was related, I thought it was an error I have in a long automation for a Heat it! thermostate that I have never been able to track down (the thermostat automation works as it should even with the error):

Error rendering data template: UndefinedError: ‘dict object’ has no attribute ‘payload’
16:46 helpers/service.py (ERROR) - meldingen oppstod først ved 10:54 og vist 83 ganger

Or do you mean on the master? The logging doesn’t work there for some reason. Edit: Yes, it does. It just hasn’t logged any errors, sorry. OK, got one now, on reboot:

2020-11-18 16:52:26 WARNING (MainThread) [homeassistant.loader] You are using a custom integration for remote_homeassistant which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant.

I see that I can use the missing devices, they just won’t populate the UI automatically, like everything else does.

That’s weird, should work like everything else. At least that’s what it does for me. Not sure what could be wrong TBH.

That might be a coding error. My understanding is that for domains they should be listed with dashes and not colons as shown in the documentation at GitHub - custom-components/remote_homeassistant: Links multiple home-assistant instances together

So for your instance it would be -

    include:
      domains:
      - lock
      - light
      - automation
      - switch

If you don’t use include or exclude then the master should be able to retrieve all domains, and everything else, from the host. If you just use include then only what is listed under include should show up.

What isn’t clear is why you have a switch that doesn’t appear, switch.garasjevinsj, yet you have other switches that do appear. Not sure how that happened if the coding is wrong unless you included them under the following:

    include:
      entities:

@not_a_coder Either I have misunderstood the instructions or you are mixing different things here. I’m not using include at all, I get all domains. But from what I could read, if a domain from the slave doesn’t excist at all on the master (like not having any lights there) you need to activate that domain by writing it up like that in the configuration file.

Now I see in the special notes in the instructions where it mentions adding a dummy entry on the master if a domain doesn’t exist on the master. I didn’t understand that part previously. However, it states it is to prevent Call service failed and not to prevent entities not being available. In the following post there is an example of Failed to call service switch/turn_off. Service not found due to no local lights on the master. He specifically states that he can see his light on the server (master).

I can see my light on the server but I can’t turn then on or off the light.
I just got this error.

When I go in to the RPI i can turn on and off the lights.

Have I missed something?

You can’t even see your light so the there is something else keeping the light enitities, a lock entity, and a switch entity from even being available. It must also be the reason the switch.garasjevinsj is missing yet the other switches are available. The switches that are available must be local. So there is something blocking access or there is still another setting missing. Have you tried using the include: option in the configuration.yaml on the master just to see what happens?

In order to see what would happen I set my master up to have one entity not included and it gave me the same error that you have without any information in the log.

Weird. I haven’t excluded anything. But that gave me an idea! And I think I have found it. Garasjevinsj is a local Z-Wave switch on the slave, ID Lock is Z-Wave on the slave, and all the lights are Z-Wave on the slave. So all the missing stuff (except for the automations and scripts) is Z-Wave on the slave. I do see their sensors. But none of those are actionable. So everything actionable in Z-Wave is unavailable. The switches that are available, are on a local (IP Address, as opposed to cloud) Telldus Z-Wave integration on the slave. I tried to add the domain zwave: but that only made the startup of Hass extremely slow, it didn’t give me more switches. There is nothing in the log, except for the futile attempt to set up Z-Wave integration and the standard:

2020-11-29 10:29:50 WARNING (MainThread) [homeassistant.loader] You are using a custom integration for remote_homeassistant which has not been tested by Home Assistant. This component might cause stability problems, be sure to disable it if you experience issues with Home Assistant.

At least it narrows it down to what I suspected was happening.

Will you post the configuration.yaml of the master? Of course, make sure your passwords are masked.

Sure, no problem. The only password was that access token, which I deleted. Even if it wouldn’t really help anybody, I have a pfSense firewall between anything accessible from outside and this setup. And if anybody’s close enough to get on my wifi, the mastiff (named Khaleesi - before she went all bad) will have noticed them and let me know. The location is pretty remote. :wink:

# Configure a default setup of Home Assistant (frontend, api, etc)
config:
frontend:
mobile_app:
person:
ssdp:
system_health:
updater:
zeroconf:
lock:
light:
recorder:
logbook:

tellduslive:
  host: 192.168.10.100

sonos:
  media_player:
    hosts:
      - 192.168.0.213
      - 192.168.1.73

# Text to speech
tts:
  - platform: google_translate

remote_homeassistant:
  instances:
  - host: 192.168.1.101
    port: 8123
    secure: false
    entity_prefix: "hytte_pi_"
    access_token: (lots of letters and numbers)

group: !include groups.yaml
automation: !include automations.yaml
script: !include scripts.yaml
scene: !include scenes.yaml

Hi folks,
Just a question: Is this integration work with a remote HA thru Nabu Casa, I mean, using a Nabu Casa link instead of local IP?
Do you have an example of this type of configuration?

Thanks in advance.

Yeah, just use your Nabu casa URL as host with port 443 and it should work just fine.

1 Like

Thanks, @postlund!!

I just started playing with this. I am building a shop on my property to run a business out of, but it’s too far to get my z-wave network to so this seemed perfect. Right now I am just testing it out (as I don’t have the new building up yet) in two VMs which is pretty much how it will be when its implemented.

To test it, I installed this on the master and on a fresh slave instance I just made a dummy input boolean. It showed right up on the master and when I turn it on it instantly turns on on the slave. Then in a few seconds, it turns off on the master but remains on on the slave. Acting as if it doesn’t see the state change. If I make any changes on the slave it is not seen on the master.

I am getting the following error in my log:

Unable to find referenced entities input_boolean.basement_test

Any ideas?

Which version are you running and how did you set it up? The error you see is a known issue but it does not cause any real problems.

I am running Home Assistant version 2020.12.2 and Remote Home-Assistant version 3.1.

I set it up via HACS and integrations then chose input booleans for “load component” and “include domains”, excluded a few of the built-in sensors that I don’t need duplicated over and that’s pretty much it.

When I do implement this, I really only plan to use it for Z-wave, but I do want to be able to see the current states of the devices in my master instance.

Try adding state_changed and service_registered under subscribed events (last page in the options flow). It’s a miss on my end not adding them which I have fixed on master. Once a new release has been made, you don’t have to include those events manually anymore. Alternatively you upgrade to master and hopefully it works out of the box.

Thank you! The subscribed events worked like a charm.

I’m a little confused about upgrading to Master. Is that a different version? I searched around a bit and didn’t find anything. Where can I find that and what is the difference in it?