#Pairing does not complete with HA's ZHA implementation.

1 messages · Page 1 of 1 (latest)

empty radish
#

Reference:
https://github.com/romasku/tuya-zigbee-switch/issues/28

What does ZHA expect to see from the device during this workflow?

Given the factory firmware pairs just fine, I'm expecting that this is something missing in the custom firmware that causes this to fail.

Trying to hunt for more information as to what / why this happens to give feedback to the developer as to why this happens.

I'm also open to the fact that maybe ZHA is only waiting for a new device on the coordinator - as when I force-added some devices via a router device closer, this was less likely to happen.

GitHub
[BUG] Pairing does not complete with HA's ZHA implementation. · Iss...

Splitting from #14 for ease of tracking. After updating the firmware from factory to this repo, the pairing does not complete successfully with ZHA. I've noticed that the following workflow is ...

fickle mango
#

Oh...I had run into a similar error recently! Trying to pair a tuya smart plug seemed to lead to an incomplete pairing process and some sort of zigbee.db corruption, whereby the coordinator thinks it recognizes the device even if it was removed.

Home Assistant Community
[Help!] Corrupted Zigbee Database

Hi there, I hope someone can help me with problems that recently started cropping up with my Zigbee network. My ~130 device Zigbee network using ZHA and a TubesZB CC2652 coordinator is the backbone of my HA setup and has largely been operating stably for years. About a week ago, I started to get strange behaviour when trying to pair new device...

#

Eventually what I did was reflash the coordinator firmware, and it seemed to resolve whatever was blocking the process. Before that I also tried removing the device IEEE from all tables in zigbee.db and the json that is used to restore backups to the coordinator...to ensure that the device is treated as new. But not sure if that stuff is necessary given the firmware reflash.

empty radish
#

It could be similar - I noticed that ZHA couldn't re-interview the device at all - leading to it not noticing that fuctionality had changed. The only way I could fix this was to remove and readd the device.

fickle mango
#

Removing and re-adding didn't work for me. Perhaps the type of measures that work depend on when the adding process fails?