Maybe someone here can point me in the right direction. Like so many people on here I’ve struggled time and time again to figure out why my thread devices stop working and I’ve learned a ton (if that sounds like you, dear reader, try changing your wifi channels because apparently thread can be touchy if you have certain overlapping wifi channels in your house and nobody seems to be talking about that!)
But my current issue is that I’ve tried repeatedly to move my HA install from a Pi to a VM and everything works great EXCEPT Matter/Thread devices.
I’ve also tried changing my rPI’s VLAN and that also breaks Matter/Thread so I feel like I’m on the verge of understanding and maybe someone here can help me get there. For reference I have multiple Google Home/TV Streamers on my network acting as a border router.
Why when I change my VLAN do my thread devices go offline immediately? I would suspect that the thread network is still communicating but my new HA VLAN just can’t see it, but why?
MOST IMPORTANT: Why when I switch to a virtualized home assistant on the SAME VLAN THAT WORKS all my thread devices go offline?
I’m looking for some technical reasoning so I can understand (and help others in the future) as well as “steps to move your matter devices from one HA install to another” because I can’t seem to find that documented anywhere.
Any help is GREATLY appreciated! I’m currently calm, since I’m back on my rPi and all my devices are online but who knows how long that’ll last…
I think read somewhere that a part of the Matter or Thread configuration is not backed up when doing the normal backup in HA, so it fails on restore.
Playing with VLANs and Matter is big can of worms though, because Matter use IPv6 and that is quite different on certain points, so your IPv4 knowledge will not be transferable to IPv6.
Matter often bind to a fe80 network, but fe80 networks exists on all your VLAN interfaces, al though they are not the same network anyhow.
In IPv4 this would be like having a 192.168.0.0/24 network on all your interfaces.
I’m guessing I MIGHT be able to clone to the VM and then re-add all my matter devices one by one… but I’d love to hear someone who did that before I take all that time just to be disappointed.
Whenever matter/thread stops working it seems like a blackhole. I’ve seen plenty of posts on here as well with people begging “How do I debug” and they go unanswered.
I’m happy to see more people (like you) talking about fe80 and ipv6 because for the longest time there wasn’t even that much knowledge on the posts, so I’m slowly learning.
I just hate being scared to touch my HA Matter install because it feels so fragile. I want to understand how to move to a VM that is also on a different subnet (redoing my IP range at home) so if you have any pointers.
Also, I read somewhere you might be able to export your thread config and restore it manually?
The Matter Server comes up and gets mDNS based discovery packets from Matter devices and this provides the way for the Matter Server to discovery the Matter device’s IPv6 address. For Thread based devices, the TBR provides a “proxy” mechanism to send the advertisements on behalf of the Thread device attached to that TBR’s Thread network.
For Thread based devices, the TBR advertises an IPv6 route over the (v)LAN (the (v)LAN that the Matter Server would also be attached to), and this route gets added to HA’s OS’s route table. The Matter Server would end up using this route to reach the Matter devices on the Thread network.
To explain why your rPi VLAN setup isn’t working, you’ll probably have to provide some more info. Namely, is your HA system, and TBR on the same VLAN?
If I recall the issue correct from the thread I had read earlier, then it was something with the encryption key not getting backed up, so your Matter network could not be accessed after the restore.
TBR and HA (pi) are on my IOT VLAN/Subnet, yeah. I can change the IP all day long but if I change the subnet it dies (still not 100% sure why, since all my VLAN stuff is IPv4? some tricky stuff happening beind the scenes I don’t understand?)
Then if I transfer the HA via backup to a VM (still on the same VLAN/subnet) matter/thread dies… I’m guessing thats due maybe to the encryption key or some other file not getting transferred.
Yeah I have searched around some but haven’t found anything that says applying a Matter Server backup to a new system doesn’t work, but on the other hand, I haven’t found anything that says it does.
I do see lots of encryption keys in the backup, but again, can’t say its sufficient.
Maybe one thing to try is goto the UI->System->Devices and find one of you matter devices that is mains-powered, and click on “Device Info” and see if it has an IPv6 address and if it does, see if you can ping -6 that address from HA.