Zigbee2MQTT grouping/device problems

I have a bunch of groups set up in Z2M, mainly because its way faster than grouping in HA, however, when a device goes offline and Im forced to re-pair it (Ikea globes), despite them having the same hardware address, the group fails to work.

Adding to this , when trying to delete the old devices in the GROUP of Z2M, the deletion fails.

2024-12-14 09:36:43z2m:mqtt: MQTT publish: topic 'zigbee2mqtt/bridge/response/group/members/remove', payload '{"data":{"device":"kitchen03a","group":"kitchen light grp 2"},"error":"Failed to remove from group (DatabaseEntry with ID '38' does not exist)","status":"error","transaction":"xk5gk-3"}'
error 2024-12-14 09:36:43z2m: Failed to remove from group (DatabaseEntry with ID '38' does not exist)

If I try to delete the group, it also fails to delete the group.

 2024-12-14 09:25:27z2m: Request 'zigbee2mqtt/bridge/request/group/remove' failed with error: 'Failed to remove group 'kitchen light grp 2' (force: false) (Error: DatabaseEntry with ID '38' does not exist)'
info 2024-12-14 09:25:27z2m:mqtt: MQTT publish: topic 'zigbee2mqtt/bridge/response/group/remove', payload '{"data":{},"error":"Failed to remove group 'kitchen light grp 2' (force: false) (Error: DatabaseEntry with ID '38' does not exist)","status":"error","transaction":"xk5gk-2"}'

I have also tried editing config.yaml and deleting the device from the group as well as deleting the faulty group in its entirety and restarted Z2M and the group and the devices and the group still persist.

The devices DO work as individual devices in the z2m dashboard.

I sort of read somewhere in a ‘breaking changes’ a long time ago that z2m migrated from config.yaml to a database structure so I would understand that the edits in config.yaml mean nothing.

EDIT: I do not have a data/ directory (its a long term evolved system from back in the HA 0.100 days)

How can I resolve this?

1 Like

Same issue here. Have you reported this as an issue in the Z2M repo?

Same issue here!!

Not yet… wanted to see if there was a solution before I made an issue.