Why are your switches in the groups with your lights? I create a light group, and then bind the switch to the group for On/Off and LevelCtrl. My switches are NEVER in the group as they do not have the same attributes. Everything works as desired for me. I even bind switches directly to each other and the light groups to create 3/4/5 way switches.
That’s how it is described how to do it on the inovelli video
1 Like
I always put the switch in the group so that I can interact with the group (light) entity only from Home Assistant and all of the bulbs and the switch update together.
1 Like
Ah, get that. I went the custom MQTT light route so that I could have on/off/brightness be commands direct to the switch and prevent additional broadcast for those and then color temp and color be direct to the group when needed, but get why you do it.
1 Like
I wasn’t seeing the errors in my logs so I tried updating. Now I am getting an error that is stopping Z2M from starting. Sigh . . .
[2024-09-09 22:43:46] error: z2m: RangeError: Attempt to access memory outside buffer bounds
at new NodeError (node:internal/errors:405:5)
at boundsError (node:internal/buffer:86:11)
at Buffer.readUInt16BE (node:internal/buffer:334:5)
at /app/node_modules/zigbee-herdsman/src/utils/backup.ts:113:89
at Array.map ()
at Object.fromUnifiedBackup (/app/node_modules/zigbee-herdsman/src/utils/backup.ts:112:33)
at AdapterBackup.getStoredBackup (/app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/adapter-backup.ts:51:32)
at ZnpAdapterManager.determineStrategy (/app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/manager.ts:130:24)
at ZnpAdapterManager.start (/app/node_modules/zigbee-herdsman/src/adapter/z-stack/adapter/manager.ts:57:26)
at Controller.start (/app/node_modules/zigbee-herdsman/src/controller/controller.ts:127:29)
1 Like
I’m seeing this error as well. I’m on Z2M 1.40.1.
I checked my bindings to the group and manuSpecificInovelli is not selected anywhere. All my switches are running 2.18.
Maybe you have tried, but have you removed the switch from the group and then conduct a re-interview with the switch and readd to the group? I wonder if the cluster is tied to the group now, since it was there, and it is trying to read an attribute that was in the group that no longer works.
Tried this without luck
Hmmmmm…so when you add an item to the group, the group is supposed to take in all the cluster attributes for the item so that it can update group members in essence. I wonder if the group has that cluster and the only way to get rid of it is to either remove all members, or complete recreate the group. Can you remove all members, delete the group, then recreate the group with the same ID and name (so automations/etc…still work) and see if the attributes no longer apply?
I did another update today and it cleared up my error that was preventing it from starting. I am not getting the error you guys are seeing though. I have a group like this:
I can control it like normal but don’t get any errors. Can someone try to do a restart or two on Z2M to make sure there aren’t any strange things cached? I could see if it was using some cached converter and then not being able to access the definition from zigbee-herdsman because it has been removed.
I’ve restarted Z2m a few times with continued errors.
Always the set brightness as shown above . I can try what csscott has suggested as well. None of the groups had the manuspecificinovelli cluster bound though so I’m not sure why it would be associated with the group from the get go.
Also functionally things seem fine other than slow downs associated when the system gets spammed with these errors.
When you first joined the switch to the group you bound endpoint 1 to the group which includes manuspecificInovelli in the inbound cluser. The new Z2M layout changes that from manuspecificInovelli to manuspecificAmazonWWAH. This is within Endpoint 1 and the inbound cluster, so in essence the devices expects to get inbound items to manuspecificInovelli, which no longer exists, and so the group breaks. Endpoint 2 and 3 still contain manuspecificInovelli in the outbound clusters. So I am pretty sure the group has the manuspecificInovelli from the inbound cluser to Enpoint 1 in it, just not sure of the easiest way to get rid of that from the inbound clusters for the group.
Is endpoint 1 included in the binding process just by including it in group? I only used endpoint 2 in the binding process otherwise. I’ve individually removed all devices from the groups but never deleted the group out of fear it would mess up automations. Guess I can try it on one of the groups and see if it still pops up
When you bind, you use Endpoint 2 as that is sending data from the Switch to the group/device you bind it to. When you joined the switch to the group, you likely chose Endpoint 1 as that is the items you use to send control commands to the switch. In the same way you bind a switch to a switch in a 3 way, you bind Endpoint 2, from each switch, to Endpoint 1 of the other switch.
So Endpoint 1 was likely, and correctly, the data you placed in the group when you added the switch to the group. Endpoint 2 is what you bind from the switch to any device/group you want to control.
I created a new group and added the lights and switches into the group. Then enabled that in Home Assistant and send on/off/brightness and color_temp/color commands at the group without generating any errors. Currently running Z2M in a docker with latest-dev image.
So I know that it will work without errors in my case if I give it a try. Any luck in deleting/recreating the group and if the errors are generated?
2 Likes
So instead of deleting whole group, I removed the switches from the group and readded and so far so good.
2 Likes
Why does Zigbee2MQTT not show all the updates for all the lights? I have some on firmware 2.11 that show no updates available, while others on 2.11 that do show updates. I bought all switches at the same time.
Edit: it was the issue documented in the yellow box on this page: Blue Series • How to Update Firmware With Home Assistant (Zigbee2MQTT) | Inovelli Help Center. Sorry for the noise.
1 Like
@Daniel15 No worries, glad you got it figured out.
@EricM_Inovelli can you confirm whether the FW has concentrator support enabled?
If you are referring to what I think you are, the feature was disabled in fw 2.08 because it was causing excess traffic on large networks.
1 Like