Seemingly random duplication of MultilevelSwitchCCReport

[kelchm]

Describe the bug

While trying to chase down some performance problems on my Z-Wave network, I observed a large number of duplicated MultilevelSwitchCCReport messages from specific devices. This isn't 100% reproducible and seemingly happens at random.

Device information

Which device(s) is/are affected (manufacturer/model)?

• Inovelli LVW31-SN (firmware v1.56)
• Zooz ZEN27 (firmware v3.3)

What is/are the node IDs?

Node Id	Model
6	Zooz ZEN27
8	Zooz ZEN27
28	Inovelli LZW31-SN
29	Inovelli LZW31-SN
42	Inovelli LZW31-SN
49	Zooz ZEN27

How are you using node-zwave-js

• zwavejs2mqtt (Supervisor Add-On v0.18.1)
  • node-zwave-js v7.7.1
  • zwavejs2mqtt v4.5.1
• HomeAssistant core-2021.6.0

To Reproduce

The best mechanism I've come up with to reproduce this issue is to repeatedly toggle a group of devices. Sometimes it works absolutely fine, other times a series of ~10 MultilevelSwitchCCReport messages show up in a period of less than a second.

Additional context

In the attached log file, I'm repeatedly toggling a group of 6 lights or setting this group off when they are already off. If needed, I'm happy to try to capture more logs.

Logfile:

• zwavejs_2021-06-07-7.log

[blhoward2]

Beyond my pay grade. @AlCalzone ? Odd that it happens with two different devices but there isn't much we can do about messages originating from devices.

[AlCalzone]

This is not only the MultilevelSwitchCCReports, I'm seeing duplicated SupervisionCCReports too. Usually this is an indication of a suboptimal mesh. I don't know exactly what happens, but I think one or more repeaters (and/or the nodes themselves) try a bit too hard to get the message to their destination because they didn't get the ACK from their target.

Unfortunately, this isn't anything we have control over - you could try the following:

• put the controller on an extension cord and move it a few cm
• heal the network
• add and/or move repeaters (permanently powered devices)

[kelchm]

Unfortunately, this isn't anything we have control over

Thanks for the response @AlCalzone -- I was afraid that was going to be the case. I've tried pretty much everything I can think of at this point (including all of your suggestions).

One thing I will note is that I'm also experiencing intermittent failures when doing a network heal that I've been meaning to open a separate issue for. Much like the issue reported here, sometimes it works 100% fine, sometimes a subset of the nodes on my network end up with failed to update the neighbor list after 5 attempts, healing failed (see attached log for an example).

All the devices on my network are permanently powered, though the majority of them are dimmers/switches which are (unfortunately) installed in metal gang boxes. Interestingly, the devices I seem to be having the most consistent issues with (nodes 2, 6, 8, 49) are the only dimmers/switches which are installed in a plastic gang box.

With the addition of some dedicated wall plug style repeaters (3x Aeotec Range Extender 7), the mesh at least appears to be fairly strong, but I suppose there's currently no way to know how borderline the links actually are between certain nodes.

Attachments:

• nodes_dump-2021-06-08.json.txt
• zwavejs_2021-06-08.log

[AlCalzone]

failed to update the neighbor list after 5 attempts, healing failed

I think this is the stick having a problem, but then I'm not very familiar with what happens under the hood there. Maybe it just needs to be power cycled. 🤷🏻‍♂️

You might also have some luck by listening with a Zniffer in different locations (near the controller and near the problematic nodes) and figuring out what happens by looking at those traces.

[kelchm]

Thanks for sharing your thoughts on this @AlCalzone -- closing this 👍