#Thread recommissioning problem -- one device incapable of recommissioning after factory reset

1 messages · Page 1 of 1 (latest)

harsh apex
#

I have about 70 REEDs in my Thread network. Adding each one as a Matter device directly to HA with a SkyConnect (no other hub involved) using my Android phone to kickstart the process proceeded more or less flawlessly.

A day or so later I was doing some crazy testing trying to turn off 40-50 devices at once via an automation, which caused the network to go berserk, likely due to overloading with status updates as the devices reported a lot of state changes over a short period of time. After a handful of Matter add-on reboots and full HA system reboots, I realized that one of the 70 devices would not become available. Investigating further, the Matter server was clearly logging out that this node was not being found any longer.

So I did a factory reset at the device and manually removed the node from the Matter server using the web UI. Now, attempting to kick off the Thread commissioning again, I get nothing -- Android doesn't seem to find the device at all. I think this is a device problem, not a Matter server/HA problem, because my understanding is that I should at least get Android saying to me that it's found the device and is beginning the process of sending credentials, etc. before HA/Matter server even get involved -- but I'm not perfectly clear on that. Is there anything that either HA or Matter server would be involved in from the commissioning point of view with a Thread device on Android, up to the point that the Android UX reports the status message regarding "finding the device on the {WiFi name} network" (or similar)? Could this be anything other than a device issue if Android's status message never progresses beyond the "searching for device" message?

#

OR -- is it possible that this is an Android/Google issue, in that there is something cached on that device which is causing it to not properly attempt to find the device via Bluetooth when kicking off the commissioning process now... 🤔

hoary summit
#

Yes, we need Matter over Thread groups (IPv6 Multicasts) to stop overloading our Thread networks.

I am interested in understanding your setup. You have 70 MoT devices, nice. Welcome to the club of more than 50 MoT devices. 😃 Can you list your devices brand and model, please.

Great that you got all these devices working with a SkyConnect only as the Thread Border Router. I have 62 MoT devices and 7 Apple Thread Border Routers. With that many devices you will see all kind of issues… 😉

#

Which Web Ui did you use to remove the device? HA‘s integration Web UI or the Web UI of the Matter server? If you did not already do it, check the Web UI of the Matter server. Do you know the node id of the deleted device?

#

Do you have another MoT device to check if the commissioning process has a general problem? Maybe this device has a defect. If I understand you right, all devices are relatively new.

harsh apex
#

I'm using all Inovelli White dimmers, no other devices (yet!). These were my first devices to establish a strong network with many REEDs.

I'm curious what you mean by "you will see all kinds of issues" -- what issues have you seen?

I used the web UI of the HA Matter Add-On (Settings, Add-Ons, Matter Server, Open Web UI). The node ID I do know, node 59. I removed it a few days ago and have not seen it in the server log output any longer (i.e., before I was seeing a lot of failures to reestablish the CASE session with node 59, and now those are gone... so I think it was actually removed from the server).

I don't currently have another device to test and don't want to remove one of my existing devices for fear that it will be impossible to add it back afterwards. 🙂 But... I could test another Android device in case the problem is on that side...

violet bramble
#

Thread (just like zigbee and z-wave) is a low bandwidth network. Its not a very good idea to send a command to 40 devices at the same time, you just flood the network.

harsh apex
#

@violet bramble Yeah that part I understand at this point (I actually started another thread in this channel on that exact topic! Would love your input on that thread... https://discord.com/channels/330944238910963714/1287479529513685092)

But on this specific topic -- do you have any advice for things I can check to ensure that this non-commissioning device is fully "cleared" as far as the Matter server and/or Thread network is concerned?

violet bramble
#
  1. Make sure the node does no longer show up in the Matter Server's own UI.
  2. make sure you really factory reset it - some devices have a bit of weird procedure for that or even their manual is wrong.
hoary summit
# harsh apex I'm using all Inovelli White dimmers, no other devices (yet!). These were my fir...

Ok, Inovelli White only. If I remember correctly @nimble roost hast a lot of them, too. Maybe he has an idea. All I wanted to say is your setup is pretty unique at the moment. I didn’t read that anybody here has 70 MoT devices with a single Thread Border Router (SkyConnect with Ha OTBR in your case) in operation. Even though the Matter (over Thread) protocol got more stable over the last year, it’s still bleeding edge technology. 😉

harsh apex
hoary summit
#

I have 7 Apple Thread Border Routers (AppleTVs and HomePods) connected to HA. This works mostly fine. Before the latest update to Apple OS 18, I had some issues. But in the past it worked fine and now with OS 18 on all my devices it looks good again. Here are users with all kind of TBRs connected to HA, but mainly Apple and/or Google TBRs.

nimble roost
#

I have 55 MoT devices with two Apple TV4K via Ethernet. Single admin mode to HA. No skyconnect.

hoary summit
nimble roost
#

The tab fixes the issue of dropping off the network.

hoary summit
nimble roost
#

There’s a tap sequence to hard reset.

#

One of my two ATVs is on 18.1 beta without issue. Other is 18.0 GA.