Failed ZHA radio migration - any idea how to revert?

Configuration Zigbee
1

I have been using ZHA with an Xbee S2 for several years and want to move to a SkyConnect. The migration guide looked easy enough - so yesterday I went for it.

It looks like the underlying capability for doing a seamless migration for my radio doesn’t exist: the Download Backup function produces an “Incomplete Backup”. So when I went through the radio migration wizard, the “Restore from automated backup” option was not visible. I then tried uploading the incomplete backup file, which produced an error. At which point the wizard would only let me X out of the dialog.

I’m expecting that I’ll have to migrate devices individually.

My hope is to get my Xbee radio back working, and run both coordinators with separate networks as it’ll take me some time to find and re-pair all 24 of my devices.

But: I can’t get the Xbee coordinator back up and running. I restored my Home Assistant from a backup to make sure that nothing from the failed migration is causing a problem but it almost looks like the Xbee module is unresponsive.

Does the radio migration disable the old radio? Do I need to something manual to re-enable it?

I tried connecting to the module with XCTU and I can still connect that way but I don’t know what most of the parameters mean.

Debug log snippet:

2024-01-06 15:18:38.309 DEBUG (MainThread) [zigpy.serial] Opening a serial connection to '/dev/ttyUSB0' (57600 baudrate)
2024-01-06 15:18:38.362 DEBUG (MainThread) [zigpy_xbee.uart] Connection made
2024-01-06 15:18:38.362 DEBUG (MainThread) [zigpy_xbee.api] at command: AP (2,)
2024-01-06 15:18:38.362 DEBUG (MainThread) [zigpy_xbee.api] Command at (b'AP', b'\x02')
2024-01-06 15:18:38.362 DEBUG (MainThread) [zigpy_xbee.uart] Sending: b'\x08\x01AP\x02'
2024-01-06 15:18:41.210 ERROR (MainThread) [homeassistant] Error doing job: Task exception was never retrieved
Traceback (most recent call last):
  File "/config/custom_components/ge_home/update_coordinator.py", line 280, in on_device_update
    entity.async_write_ha_state()
  File "/usr/src/homeassistant/homeassistant/helpers/entity.py", line 918, in async_write_ha_state
    raise RuntimeError(f"Attribute hass is None for {self}")
RuntimeError: Attribute hass is None for <entity unknown.unknown=unknown>
2024-01-06 15:18:41.364 WARNING (MainThread) [zigpy_xbee.api] at: No response to AP command
2024-01-06 15:18:41.364 DEBUG (MainThread) [zigpy_xbee.zigbee.application] No response to API frame. Configure API mode
2024-01-06 15:18:42.566 DEBUG (MainThread) [zigpy_xbee.uart] Command mode sending b'+++' to uart
2024-01-06 15:18:44.567 DEBUG (MainThread) [zigpy_xbee.api] Command mode no response to AT '+++' command
2024-01-06 15:18:44.567 DEBUG (MainThread) [zigpy_xbee.api] Failed to enter AT command mode at 57600bps, trying 1200 next

(log continues to try and never works)

Full debug log here.

2

In all my fluster trying to figure out what was going on, I’d managed to barely separate the Xbee board from the explorer board (the part that does the usb-to-serial). So in fact I was getting no response to any AT commands - until I plugged it in all the way!

I’ve all but given up hope though. It appears that somehow the PAN ID has changed, and while I can reset the Extended PAN ID back to what it was, getting the 16 bit PAN ID back to it’s original value doesn’t appear to be something I can configure in XCTU. It just keeps setting it to 0xFFFF, and then zigpy sets it to a new ID on startup.

Given that I’ve also just discovered that ZHA won’t let me run multiple simultaneous networks, I’m just going to connect up the SkyConnect and form a new network and start re-pairing everything…