OpenZWave - Devices need to be updated after Homeassistant reboot

I switched to openzwave plus beta component.
All is well and more detailed then with the legacy component.
I have however one MAJOR problem.
Everytime I reboot Homeassistant I need to send an update command to the nodes to have them bask online in Homeassistant.
After I update them everything works fine until next reboot.

Clearly this defeats completely the plus of uncoupling zwave from Homeassistant.

What “update” command are you referring to?

Update Node in the Homeassistant ZWAVE Component Interface (button “aggiorna” in screenshot)

Oh refresh node, got it. You shouldn’t have to do this, the devices are supposed to be cached.
How are you running ozwdaemon? Addon? Standalone docker container?

Docker container on a separate host.
Since the Server is far from “operating area” I have a raspberry running openzwave allinone with dongle plugged in.
Could it be this?

Are you running the “latest” docker container or build-150?

I think so. Installed just 2 days ago.
e3ee10c475b8

I’m using the openzwave/ozwdaemon:allinone-build-180 image currently, working fine.

You connect by MQTT to external server?
Your maybe you need to set MQTT to retain messages else I’d expect you must refresh so your remote server send the initial device state.

Maybe use automation to refresh after HA reboot

Have you considered runnimg HA in docker on second host and using eventstream or statestream to connect them?

did a couple of reboots and it looks fine now.
Will monitor and open an issue + update here if I see it again

Happens again.
I have to then restart the ozwave container.
Not ideal

Sounds like you’re persistent storage is fubar.

I am not sure I understand.
Can You exolain what you mean?

When you restart the container the ozwcache XML is reloaded and all your devices pop back. If this cache file is corrupted somehow, then you’d need to refresh the node to get the information properly again.

Where is this file in case of a container based installation on a separate host? I suppose at the host in the container
.
Is it possible that the ha insurance is reading at the server and does not find anything so it assumes corruption. ?

It would be where ever you mapped the volumes to be.

If you didn’t specify then you have no persistent storage and that’d be the first place to start looking as the cause of your problem.

Review the README:

1 Like