Wanted to bring attention to this issue on GitHub since I'm having the same error and log as the OP: https://github.com/home-assistant/addons/issues/4210
Any way to revert to a prior version of OTBR (if no backup was made)?
1 messages · Page 1 of 1 (latest)
Wanted to bring attention to this issue on GitHub since I'm having the same error and log as the OP: https://github.com/home-assistant/addons/issues/4210
Any way to revert to a prior version of OTBR (if no backup was made)?
@ember moss not sure if you're the same puddly on Github but it seems to be your change (in commit 89da067) that brought this issue in
can you temp. disable the silicon labs multiprotocol in ha and see if device starts?
How is the coordinator connected to ha?
Fails again, same error every time
smlight SLZB-MR1
Uh could you send a screenshot of where that option is
I have the EFR32 chip flashed with the rcp firmware and connected over usb to the HA
Which is a HAOS vm running in proxmox
Coordinator’s other microcontroller is set up as a zigbee coordinator and that works over LAN
i dont have it so i have to rely on AI (sorry 😉 but not the only one it seems )
To disable multiprotocol on an SLZB-MR1,
go to Settings > System > Hardware in Home Assistant, select Configure under your SkyConnect device, and then select Remove 802.15.4 radio multiprotocol support
there have been issues with multiprotocol. is the zigbee part up&running
?
or in your addons
that doesnt really make sense
let me look around though
but also not entirely sure if we're talking abt the same thing -- multiprotocol is using the same radio for 2 protocols right?
the mr1 has two separate radios and two separate microcontrollers so it doesn't have that issue
but yes, the zigbee part works
ok. you did powercycle the slzb already i guess?
Seems like everybody having this issue is using the SLZB-MR devices
What chip is being used for Zigbee and what chip is being used for Thread?
Which one for which?
zigbee is the CC
That's the exact same chip as the ZBT-1 (when used for Thread) and I know the code works fine when communicating with the OpenThread firmware
Dumb question: have you tried power cycling the SLZB device and then restarting the addon? I have a vague theory for why this could possibly be happening.
as i dont have one ( i am only troubleshooting) i dont 🙂 but i did ask OP if he/she did...
@ember moss would you mind keeping me posted if there is something new about this issue? Would be helpful for other users in troubleshoot 😉
I've always been curious - when you connect an SLZB-MR device via USB, do you get serial interfaces for one one of the radio chips? which one? or both of them?
for posterity, the issue I'm running into seems to have the same error output but different cause than everyone else in that thread
as I said before, I'm running the thread radio firmware on the EFR32 over USB, and HAOS on a Proxmox VM
turns out you need to pass the USB port through from Proxmox to the VM (using just the serial port ttyS0, which was automatically passed through for me, doesn't work), this fixed it even on latest add-on version for me
on the older firmware it's locked to one of them over USB
on the newer firmware it can be switched to one of them or the other but not both
so you have to pick one to communicate with over LAN and one over the usb (if you wanted both)
Makes sense, iirc they're using a discrete chip for usb-serial so you only get one port. I guess they must have some switches internally to pick which radio it's wired to.
Would have been nice of them to make both available tho.
For sure, though I guess if you wnated that, it’s about double the price of a SLZB-06 anyway so you could just get 2 of those
And have the extra flexibility to put them in different places for example
Could have done it by including an internal usb hub and two converters, or alternately a software usb-serial running on the esp32 like how the ZWA-2 works would allow multiple ports.
Interesting. The migration tool uses pyserial to communicate with the port while the actual otbr-posix daemon talks to it via raw syscalls. There shouldn't be a difference 🤔
Just to be clear, I never had this setup working before v2.15 so might be unrelated honestly
But it is interesting that downgrading changes the error