My question is similar to Home Assistant migration - ZHA & Conbee II failure - #4 by Hedda.
After a hardware migration the configuration PAN ID etc on ConBee II has changed. I have the old settings, but I get errors when trying to apply them. Automatic migration/restore didn’t work, so I’m trying the manual migration “Restore your network settings from an uploaded backup JSON file. You can download one from a different ZHA installation from Network Settings, or use a Zigbee2MQTT coordinator_backup.json file.” However, I get unspecific error messages if I try to upload the json.
Could somebody possibly explain how I re-configure the ConBee II so that it has the old settings?
[Edit] I’ve played around with the settings in the deconz add-on, and I think I managed to restore the settings. Enabling the ‘control’ panel in the deconz add-on, it seems that zigbee devices have rejoined. However, even after restarting what I can restart, the ConBeeII doesn’t actually start - the main zha panel shows “ConBee II - Failed setup, will retry: - Check the logs”. However, the logs don’t seem to show anything relevant.
Edit 2. Some things are clearer now. The issue was that - in the hardware change - the Conbee setting was rewritten, and the hass-based procedure for resetting it didn’t work. I then installed the deconz add-on. While the add-on prob meant as an alternative to zha, it did allow me to (1) change the settings on the Conbee II adapter and (2) let devices rejoin. HOWEVER, the add-on would then try to grab the Conbee II on boot, meaning zha could not access it. So I needed to disable the deconz plug-in, for zha to access the Conbee. So my zigbee network is now back.
Edit 3. Above (in Edit 2), I tried to get devices to rejoin the zigbee network. Most devices that are mains powered, and some other battery-based, did rejoin. However, hass seems is particular about missing entities it seems, and so the missing devices generate a configuration error, which seems to stop automations from loading, as well as stops hass from being able to reboot.
Edit 4. In the end, I enabled ssh, and obtained core.entity_registry (/homeassistant/.storage/core.entity_registry), exporting this into a tsv with entries like
54:ef:44:10:00:53:c8:d8-1-1280 54ef44100053c8d8-1-1280 living room Iaszone
Then, when hass complains (with the uuid), I could easily identify devices like that. However, it turns out that I didn’t have to remove all unknown entities - it seems to be only some entities that needed removing. Maybe it was that I mainly needed to remove entities that are used in automations… but I’m pretty sure that there are still unknown entities used in some automations.
It’s still not quite clear