I switched to zwavejs a few days ago. everything looked good, but today one of my dimmer devices is showing on all the time. when I turn it off through my UI, the toggle button switches back to on.
Does anyone have any recommendations on how to trouble shoot?
I have an automation for this light that will expect to execute based on it being on, but might fail.
I have two automations that I confirmed do work - one turns on the light, one turns it off.
check control panel in zwave js (available when you used the mqtt version of JS). Lot of native info is there. I struggled with motion sensor since JS, but got trhough it because the control panel showed met what happened with state and zwave traffic.
Thanks. I am not using the mqtt version though. just zwavejs. I would hope that is enough. I would had to have to migrate again.
Does zwavejs not have any control panel?
I see this in the log, but do not know how to interpret
10:48:11.526 CNTRLR « [Node 007] Received updated node info
10:48:13.532 CNTRLR « [Node 007] received CentralScene notification {
"nodeId": 7,
"ccId": "Central Scene",
"ccCommand": "0x03",
"payload": "0xd68001"
}
The central scene command, is infact a native zwave command to activate a scene. so if you have that one activated and an automation, it might conflict. (this is comletely guessing here.) I think in openzwave 1.4 the central scene command wasnt default ‘on’. Zo moving to js stack making it availble might do the difference.
Really would suggest the mqtt add in for the control panel, then you can see more of whats going on on zwave level and settings.
(the regular js server does not have control panel yet).
Switching is pretty easy, however they dont share the cache, so it might take some time again that mqtt js server shows all nodes to HA again.
I am not using any ‘scenes’ in my system. None of the devices are programmed to control other devices.
I do not find any literature with a simple internet search.
From a hubitat thread, I believe the CentralScene Notification is the mechanism for the device to tell the hub its state. If so, it is an expected message.
Turning on and off node 7, I get the above message with these payloads:
"payload": "0xd98001" when turning ON
"payload": "0xda8002" ... OFF
A list of payloads would be great - again, I find no literature.
The deduction is that Hass is not picking up the message and updating the UI.
Is there a way to delete the device and re-add, or do I have to exclude it from the network, then include?
Or is there some artifact from the transition from OZW?
Here is the device from core.entity_registry
{
"entity_id": "light.kitchen_recessed",
"config_entry_id": "7f2a2bbd773ac5226f3a6efb50a14f7a",
"device_id": "0afeea16b2f7a2b4699447a0710dab90",
"area_id": null,
"unique_id": "4179798772.7.7-38-00-currentValue-00",
"platform": "zwave_js",
"name": "Kitchen Recessed",
"icon": null,
"disabled_by": null,
"capabilities": {},
"supported_features": 33,
"device_class": null,
"unit_of_measurement": null,
"original_name": "GE Enbrighten Z-Wave Plus Smart Dimmer: Current value",
"original_icon": null
and from the core.device_registry
{
"config_entries": [
"7f2a2bbd773ac5226f3a6efb50a14f7a"
],
"connections": [],
"identifiers": [
[
"zwave_js",
"4179798772-7"
]
],
"manufacturer": "GE/Jasco",
"model": "46203",
"name": "GE Enbrighten Z-Wave Plus Smart Dimmer",
"sw_version": "5.54",
"entry_type": null,
"id": "0afeea16b2f7a2b4699447a0710dab90",
"via_device_id": null,
"area_id": null,
"name_by_user": null,
"disabled_by": null
},
I also have a zwave dump and can find a Node: 7 section that looks like many of the others. I can upload or paste if needed.
there is no value on device class, are other nodes having same ‘no value’ for device class?
What does the zwave dump have for info? (more info , more hints)
What you describe that ha didnt update, i had with some sensors… The items on zwave node and the entities in ha were no very clear. I had to enable a hidden entity and also in some cases restart ha to have working entities. (there is no ‘reload integration’ button yet for that refresh)
correct. all of the devices have the following:
"device_class": null
The js dump is quite lengthy and not formatted. If there is a specific area to look, I can paste that.
Is there a way to refresh the node? If not, I might just exclude and include.
Well, the device is working again, but now two others have died.
[edit] the two devices had to be renamed. Their entity_id after the transition to JS included ‘_2’ suffix. After a restart, the entity_id no longer has the suffix. I went into my UI and removed the suffix.
It would be nice to know why this happened, and why the other device is now working.
Curious: does anyone know if there is a linting tool for the zwaveJS dump?
I got the same error a few days after migrating, first all looked OK, then one specific device was stuck at on. Only in HA though, in reality it was off. Yesterday I tried setting the state to off using developer tools, after that it seems to work again. Will see if it keeps working the coming days.