I’m using zwave2mqtt in a docker container (v6.15.2) and home assistant core (v2022.8.4) using HassOS in a VM.
I had an unexpected power cut, and after coming back online, the only thing not working is Z-wave in HA. The zwave2mqtt control panel (local_ip:8091/control-panel) works fine, I can ping devices and get updates about power consumption, etc. But nothing is available in HA.
In the logs for the zwave2mqtt docker container, I see:
2022-08-14 15:44:12.024 WARN STORE: scenes.json not found
Logging to file:
/usr/src/app/store/zwavejs_2022-08-14.log
2022-08-14 15:44:14.065 INFO ZWAVE-SERVER: ZwaveJS server listening on 0.0.0.0:3000
2022-08-14 15:44:15.047 INFO ZWAVE-SERVER: DNS Service Discovery enabled
2022-08-14 15:44:36.032 INFO ZWAVE-SERVER: Client disconnected
But the zwavejs_2022-08-14.log is empty. I found this topic, and tried, as described, to disable soft-reset and restart (both HA and zwave2mqtt), but this didn’t do anything.
I have had this happen with mine after power outage. For some reason the VM starts with the USB on a different port and i have to go in and change the passthrough on proxmox and then go in and tell zwave where the usb is. I just added a UPS to my computer just for this reason.
with the level set to debug. And nothing in the HassOS system logs.
@JonathJon I don’t think this is the issue, since HA is not aware of the zwave stick (it’s not passed to the VM, just used directly in the zwavejsmqtt docker container, and it works directly in that).
I do occasionally see this issue for my ZigBee stick on system reboot though, because this is passed to the HA VM. Then I have to do virsh detach ... followed by virsh attach ..., and it works again.
I tried re-configuring server in HA (I only input ws://host_ip:3000, same as it was initially set up with), but no change.
This should be what’s exposed, since I can connect just fine to http://host_ip:8091/control-panel from a browser.
Do you think it’s related to the:
2022-08-14 15:44:12.024 WARN STORE: scenes.json not found
I’ve never looked at the docker logs before, so don’t really know if this is new, but could it be that this file was deleted/corrupted due to the unexpected power outage?
If both HA and ZWaveJS2MQTT are running within manual Docker I don’t think the host_ip is available by default (opposed to the container IP) from within Docker itself. Are you managing your own Docker containers or are you using HassOS as mentioned in your initial post? In case of manual management, ensure the container is mapped to the host networking, if using HassOS and thus the ZWaveJS2MQTT addon then use the container name instead of IP.
I’m using HassOS, but have most things in manually managed docker containers (didn’t want to move everything when I moved HA from core in docker to hassOS).
I don’t think the host_ip is the issue, since (1) I’m using the exact same setup as before the power cut, where everything worked fine, (2) I can access the zwave2mqtt server control panel using host_ip.
Also, after trying a few things in the control panel, e.g. healing the network, it does seem like there is an issue in that, not just the connection to HA. Healing fails, and I can’t see certain “capabilities” (e.g. the switch.XX and most sensor.XX of switches are gone in HA, but I can see the button.XX_ping and sensor.XX_node_status for each node. So HA can see the server (on host_ip), but something likely got messed up in the server. Well, that’s my current best guess…