Openzwave beta: all nodes unavailable after HA restart

I setup Openzwave and setup all my devices. It seems to work well. However, when I reboot HA all the zwave entities are unavailable (many are hardwired inovelli switches), even after 10 minutes. The Openzwave addon is running, and the integration says the network available.

The only way to get the zwave nodes back is to restart the OZW addon. I’m on HA 0.115, OZW addon 0.5.2, and a HUSBZB-1 stick.

If you issue a mqtt.dump service call, do they come back?

mine does the same. I was disappointed as I thought moving to mqtt would solve this.

Having same trouble here… it was all working perfectly. I was able to restart hassio from configuration menu and the openzwave would continue running.

However after trying to pair/include a new motion sensor it started getting hickups. And after a reboot all devices became unavailable. At first i thought it might be updating to latest hassio and OpenZWave but rolling back on my snapshot didn’t solve the problem.

I tried doing mqtt.dump without solving it. I tried restarted the openzwave add-on. I tried restarting hassio from configuration. Right now nothing is working still :frowning:

Post your logs.

This end result is being experienced by many people. Some examples:

Hard to tell whether the root causes are the same or not. In my case, a HASS restart will often render all Z-Wave entities unavailable until (a) I manually interact with them (and that only works some of the time) or (b) an ozwdaemon restart.

Common thread here seems to be the mqtt addon.
The standalone docker container and standalone mqtt containers don’t seem to have this issue, as far as I can tell.

I’m using the standalone ozwdaemon Docker image and a standalone MQTT broker (VerneMQ in Docker) – nothing hass.io-related – and I’m experiencing it every time I restart HASS.

eclipse-mosquitto seems to work fine for me, have you tried using another broker?

Thanks for the links. The next time I have this happen to me I will try their procedure:

Deverloper Tools -> Service and call mqtt.dump with the following settings

topic: 'OpenZWave/#'
duration: 5

“Then all the OZW components instantly become available.”

3 Likes

I have not, but that’s a great test – will give it a try and find out.

1 Like

Tried Mosquitto-via-Docker, but same thing occurs. Odd! I should add a third possible outcome that I’ve seen:

  1. All Z-Wave entities are unavailable and a HASS restart fixes things.
  2. All Z-Wave entities are unavailable and an ozwdaemon restart fixes things.
  3. All Z-Wave entities are unavailable, but become available after an indeterminate amount of time.

Sounds like a real fun bug to diagnose and fix. :joy:

I have experienced all three scenarios. I’m on Proxmox with HA OS virtual machine / supervised and the official Mosquitto add-on.

Are you running Home Assistant Supervised/Home Assistant OS by chance?

No – I’m running the straight Home Assistant Core Docker image.

1 Like

About how many nodes do you guys have?

I have a controller + 16 nodes (does a controller count as a node?):

I was expecting more :stuck_out_tongue:
This is surely a stumper for sure. :frowning:

1 Like

I have 65 nodes including the controller

I tried this today (again, with HASS Container and a separate MQTT container) and it brought my entities back instantly.

1 Like