Z-Way v4.0.2 Release

[lanbrown]

No, it isn't that tricky. Home Assistant talks to the core Z-Way instance. The core Z-Way instance talks to five remote Z-Way instances; it talks directly to all five of the remote systems. One remote system does not communicate with another remote system. So the core system is just combining the five remote systems and then HA is only communicating with the core instance. So from a Z-Way perspective it is a star topology.

When the core z-way system process starts to have an issue, yes it is from the core system to the remote systems. It seems to impact one remote system and then starts spreading to the others. To recover I just have to restart the core system either by just the z-way-server or rebooting the Pi. Setting the polling to the remote z-way systems to 5 seconds makes the system far more stable; the default of 1 second it seems to cause an issue. I have ran 5 seconds for many months and rarely have had an issue, set it to a second and it usually takes a few days before things start to fail.

I believe the HA integration uses WebSockets. Given that a polling value is given for the remote z-way app for polling, that leads me to believe that it is not using WebSockets. Is WebSockets on the roadmap for the remote z-way systems app?

[PoltoS]

Yes, Websocket is in our roadmap for the RemoteHA module

When we talk about the hang, would be interesting to see tcpdump logs and gdb trace - it might be a threads deadlock or JS garbage collector issue. Or something else. Interesting to debug it. If you know how to reproduce it in an isolated environment, we can try to attack it.

[lanbrown]

I think this would be best once I get a "dev" Z-Way core instance running. Then it can just sit there and run until it fails and then debug it since it won't actually cause an issue.

I do have a spare Pi I could use as well.

Websockets would be nice for everything; Home Assistant and remote z-way since it would allow realtime updates. I would think the polling can cause issues since it can be quite a few devices and some devices have multiple values/states like a multi-sensor device.

[PoltoS]

HA is already on WebSockets. This is why it is super fast. Our new Lite UI is also using WebSockets (navigate to /lite - you can also select it in the mobile app)

[lanbrown]

I sent an email to support with the debug and z-way-server.log. It didn't crash but the unidirectional communication was.

[lanbrown]

Sent another email to support. The z-way process finally went to the exited state. The debug and z-way-server.log files were attached.

Code: Select all

sudo service z-way-server status
● z-way-server.service - LSB: RaZberry Z-Wave service
     Loaded: loaded (/etc/init.d/z-way-server; generated)
    Drop-In: /etc/systemd/system/z-way-server.service.d
             └─override.conf
     Active: active (exited) since Wed 2022-11-30 16:57:37 CST; 2 days ago
       Docs: man:systemd-sysv-generator(8)
    Process: 18171 ExecStart=/etc/init.d/z-way-server start (code=exited, status=0/SUCCESS)
        CPU: 17ms

Nov 30 16:57:37 systemd[1]: Starting LSB: RaZberry Z-Wave service...
Nov 30 16:57:37 z-way-server[18171]: Starting z-way-server: done.
Nov 30 16:57:37 systemd[1]: Started LSB: RaZberry Z-Wave service.

[PoltoS]

Z-Way v4.0.2 was just released.

This release is fixing some bugs found by the community and our tech support:
- in Security module
- in Security S0/S2 interview
- in SensorBinary and SensorMultilevel interview
- in UserCode interview
- in MultiCmd and Transport Service handling

There is one new feature: Notifications can now include device properties and statuses (add <devId:metric:title> or <dev:metric:level>)

Happy holidays!

[Sujata]

Excellent!

[philippescholtes]

PoltoS writes "Expert UI got a new configuration UI for Z-Wave Plus v2 devices. The Expert Commands page was reworked and got API command helpers for easier integration in your projects and parameters to tune the Z-Wave engine."
I noticed the change in http://YourIP:8083/expert/#/configurati ... urDeviceID. However for those of us not too familiar with API commands, the new interface is less user-friendly. The previous interface simply listed, after completion of the interview process, the parameters exactly as they appeared in the device manufacturer's manual, with the values that could then be easily adjusted to one's needs. The new interface does not tell me how to do that--at least not in a language that I can understand. Is there a way to get back the previous UI for that particular page?

[PoltoS]

Dear Philippe, the new UI extends the old one, so it is just an addition.

The algorithm is (by priority):
1. If the device reports config parameters, draw the UI based on this info (for new devices only, 2020+)
2. If there is a Device Description Record with parameters listed inside, draw the UI based on that info
3. Draw the very basic UI that requires you to look in the manual yourself.

So for your device (if it is not a new one) just select the description file