Yesterday updated zigbee2mqtt to 2.1.2-1 - after that’s some devices stop working in HA frontend but working well in zigbee2mqtt frontend.
How to find what happened? Why the are not working in HA.
After get back to zigbee2mqtt 2.1.1-1 fro backup everything working well.
Any help please.
The add-on logs will tell you what is going on.
Same thing here, but only after a power failure this morning. All of my lights are unavailable in the front-end, but managable from Z2M. I restarted HA, restarted Z2M, restarted MQTT, nothing helps so far. Logging of Z2M is also not very helpful
|[2025-03-02 11:18:48] [32minfo[39m: |z2m: Currently 44 devices are joined.|
|---|---|
|[2025-03-02 11:18:48] [32minfo[39m: |z2m: Connecting to MQTT server at mqtt://core-mosquitto:1883|
|[2025-03-02 11:18:48] [34mdebug[39m: |z2m: Using MQTT login with username: mqtt_user|
|[2025-03-02 11:18:48] [32minfo[39m: |z2m: Connected to MQTT server|
....
[2025-03-02 11:18:53] [34mdebug[39m: z2m: Discovering entities to Home Assistant
[2025-03-02 11:18:53] [32minfo[39m: z2m:mqtt: MQTT publish: topic 'homeassistant/binary_sensor/1221051039810110150109113116116_0x00124b0024cbd77f/connection_state/config', payload '{"device":{"hw_version":"zStack3x0 20211217","identifiers":["zigbee2mqtt_bridge_0x00124b0024cbd77f"],"manufacturer":"Zigbee2MQTT","model":"Bridge","name":"Zigbee2MQTT Bridge","sw_version":"2.1.2"},"device_class":"connectivity","entity_category":"diagnostic","name":"Connection state","object_id":"zigbee2mqtt_bridge_connection_state","origin":{"name":"Zigbee2MQTT","sw":"2.1.2","url":"https://www.zigbee2mqtt.io"},"payload_off":"offline","payload_on":"online","state_topic":"zigbee2mqtt/bridge/state","unique_id":"bridge_0x00124b0024cbd77f_connection_state_zigbee2mqtt","value_template":"{{ value_json.state }}"}'
[2025-03-02 11:18:54] [34mdebug[39m: z2m: Skipping discovery of 'binary_sensor/1221051039810110150109113116116_0x00124b0024cbd77f/restart_required/config', already discovered
[2025-03-02 11:18:54] [34mdebug[39m: z2m: Skipping discovery of 'button/1221051039810110150109113116116_0x00124b0024cbd77f/restart/config', already discovered
[2025-03-02 11:18:54] [34mdebug[39m: z2m: Skipping discovery of 'select/1221051039810110150109113116116_0x00124b0024cbd77f/log_level/config', already discovered
And from MQTT broker:
[11:23:52] INFO: Starting mosquitto MQTT broker...
2025-03-02 11:23:52: mosquitto version 2.0.20 starting
2025-03-02 11:23:52: Config loaded from /etc/mosquitto/mosquitto.conf.
2025-03-02 11:23:52: Loading plugin: /usr/share/mosquitto/go-auth.so
2025-03-02 11:23:52: ├── Username/password checking enabled.
2025-03-02 11:23:52: ├── TLS-PSK checking enabled.
2025-03-02 11:23:52: └── Extended authentication not enabled.
2025-03-02 11:23:52: Opening ipv4 listen socket on port 1883.
2025-03-02 11:23:52: Opening ipv6 listen socket on port 1883.
2025-03-02 11:23:52: Opening websockets listen socket on port 1884.
2025-03-02 11:23:52: Opening ipv4 listen socket on port 8883.
2025-03-02 11:23:52: Opening ipv6 listen socket on port 8883.
2025-03-02 11:23:52: Opening websockets listen socket on port 8884.
2025-03-02 11:23:52: mosquitto version 2.0.20 running
2025-03-02 11:23:52: New connection from ::1:56348 on port 1883.
2025-03-02 11:23:52: Client <unknown> disconnected due to protocol error.
2025-03-02 11:23:52: New connection from 172.30.33.3:36024 on port 1883.
2025-03-02 11:23:52: New client connected from 172.30.33.3:36024 as mqttjs_48222bf9 (p2, c1, k60, u'mqtt_user').
[11:23:52] INFO: Successfully send discovery information to Home Assistant.
[11:23:53] INFO: Successfully send service information to the Supervisor.
2025-03-02 11:23:56: New connection from 172.30.32.1:33265 on port 1883.
2025-03-02 11:23:56: New client connected from 172.30.32.1:33265 as 4IWHOdYgbdjdHBgZTOg66M (p2, c1, k60, u'homeassistant').
2025-03-02 11:25:28: New connection from 172.30.32.2:43434 on port 1883.
Could this lines from MQTT be the problem?
2025-03-02 11:23:52: Client <unknown> disconnected due to protocol error.
If so, what does it mean? I didn’t change anything
Zigbee2mqtt show no error but in HA Core I’ve got
2025-03-02 07:06:30.055 WARNING (MainThread) [homeassistant.helpers.service] Referenced entities light.sypialnia_bedled1 are missing or not currently available
2025-03-02 07:06:30.056 WARNING (MainThread) [homeassistant.helpers.service] Referenced entities light.sypialnia_bedled2 are missing or not currently available
Entities are gone but still exists in zigbee2mqtt frontend.
@elstupid the same here. After update works ok - restart and most zigbee lights are gone.
I found now that some Zigbee bulbs are available in the frontend, but specfic models are not.
For me that’s all of my light bulbs from Tuya with model TS0505B_1_1
Manufacturer: _TZ3210_mja6r5ix
And I have many of that 
Maybe here’s a problem in Z2M logging (Range Error):
[2025-03-02 11:19:45] e[34mdebuge[39m: z2m: Received Zigbee message from 'Eethoek', type 'raw', cluster 'lightingColorCtrl', data '{"data":[8,132,10,9,240,72,0],"type":"Buffer"}' from endpoint 1 with groupID 0
[2025-03-02 11:19:45] e[34mdebuge[39m: z2m: No converter available for 'TS0505B_1_1' with cluster 'lightingColorCtrl' and type 'raw' and data '{"data":[8,132,10,9,240,72,0],"type":"Buffer"}'
[2025-03-02 11:19:46] e[34mdebuge[39m: zh:zstack:unpi:parser: <-- [254,27,68,129,0,0,0,3,145,240,1,1,0,51,0,124,18,215,0,0,7,8,133,10,9,240,72,0,145,240,29,123]
[2025-03-02 11:19:46] e[34mdebuge[39m: zh:zstack:unpi:parser: --- parseNext [254,27,68,129,0,0,0,3,145,240,1,1,0,51,0,124,18,215,0,0,7,8,133,10,9,240,72,0,145,240,29,123]
[2025-03-02 11:19:46] e[34mdebuge[39m: zh:zstack:unpi:parser: --> parsed 27 - 2 - 4 - 129 - [0,0,0,3,145,240,1,1,0,51,0,124,18,215,0,0,7,8,133,10,9,240,72,0,145,240,29] - 123
[2025-03-02 11:19:46] e[34mdebuge[39m: zh:zstack:znp: <-- AREQ: AF - incomingMsg - {"groupid":0,"clusterid":768,"srcaddr":61585,"srcendpoint":1,"dstendpoint":1,"wasbroadcast":0,"linkquality":51,"securityuse":0,"timestamp":14094972,"transseqnumber":0,"len":7,"data":{"type":"Buffer","data":[8,133,10,9,240,72,0]}}
[2025-03-02 11:19:46] e[34mdebuge[39m: zh:controller: Failed to parse frame: RangeError [ERR_OUT_OF_RANGE]: The value of "offset" is out of range. It must be >= 0 and <= 5. Received 7
[2025-03-02 11:19:46] e[34mdebuge[39m: zh:controller: Received payload: clusterID=768, address=61585, groupID=0, endpoint=1, destinationEndpoint=1, wasBroadcast=false, linkQuality=51, frame=undefined
Try reverting back to 2.1.1 (not sure if that’s possible with HAOS?), there are a lot of issues with 2.1.2 popping up.
The same here:
IKEA LED1537R6/LED1739R5 gone
but
IKEA LED1537R6/LED1739R5 exists and work well.
I have no idea what is going on.
Only reverting to 2.1.1 help here.
2.1.2 introduced a new feature that “brings a significant decrease in memory usage” by loading converters (“device handlers”) on-demand.
Judging by the relevant PR, it’s quite a big change (lots of files involved) so my guess would be that a bug was introduced that is causing these problems.
It looks like the 2.1.2 update broke lights that support colors, more info here.
Here is bigger thread #26557 of listed issues.
For now I suggest you skip the update.
Thanks Robert. Now I see why the huge memory decrease 
Will try to revert to 2.1.1, no backup available
I did have a backup fortunately. Reverted back to 2.1.1 and all working as expected
Fixed in 2.1.3 see Github
Just wait for it to come to HA
What’s considered good practice for updates? Be 1 update behind to avoid things breaking?
It’s hard to say since Z2M just released 2.1.3 which fixed this issue
Which fixes a breaking change coming up in the HA release on March 5
The HA release on March 5 fixes the SmartThings Integration
I personally don’t update unless things aren’t working in my setup. I will update if Z2M adds support for a new device I have. I will update HA if an integration I need is fixed/released.
These new updates will allow me to finally use some devices which have been sitting around.
I do a backup / snapshot before every upgrade and tend to be on the latest version. I’ve needed to roll back a handful of times the last 7 years.
And this is the one in HA setting > system > backup? Just making sure it’s not the one that’s within the z2m interface.
I’m using hypervisor level backups and snapshots. As long as you ensure you have something to roll back to you are good.
I experienced similar issues yesterday with random Zigbee devices after upgrading Z2M to 2.1.2-1 and to 2.1.3-1.
Simply doing a Z2M restart or even a host reboot didn’t help. Neither did a re-pairing of the affected devices remedied the situation.
But what finally helped was:
All devices went back to normal and are accessible and controlable through HA again after the above procedure.