EmberStatus.DELIVERY_FAILED and NWK conflicts #85881
Comments
|
Hey there @dmulcahey, @Adminiuga, @puddly, mind taking a look at this issue as it has been labeled with an integration ( Code owner commandsCode owners of
(message by CodeOwnersMention) zha documentation |
|
Welcome to ZHA sadly you was not being converted from deCONZ (Joke) :-))). The second i have also seen and some other user have reporting it. |
|
Hi @MattWestb, Thanks for the detailed answer. :) I've started rollback this morning to Z2M with CC2652RB. I guess it's still too early for EFR32MG21. |
|
All coordinators is working little different and running different code (the Zigbee stack on the chip) and also different settings for the mesh network. The "re joining" devices is very likely not one thing with the coordinator (then you have blocking the EZSP for having direct children) but is very likely one problem with its parent is not liking its child or the child is not liking its parent. I recommending looking if some of the "re joiner" is having problem in Z2M with TI coordinator then they is connected to one router. I think you is finding radio interference or one or more bad routers in you network. One very known issue that is not fixed in current TI firmware is nearly impossible having real Zigbee 3 end devices having the coordinator but is working very well if having one normal router as parent (Danfoss Ally TRV is falling in deep sleep and must being re powered) but i think its being fixed in the dev firmware. In the end you shall doing what you like and what is best for your system !! |
|
@MattWestb... just a few more things to add about ZHA in general. :) But I was surprised that some of the issues that are being presented two years ago are still present today.
Not of those issues are present in Zigbee2MQTT. These two are more like feature requests:
I am happy to help to debug those issues on ZHA including the initial ones with devices leaving and joining the network, but I am not sure about ZHA Developers' availability, as well as how many people are affected by the same issues as me, and I cannot keep my production network with those bugs for a long time. FYI @dmulcahey, @Adminiuga, @puddly |
An issue was fixed regarding this over a year ago. No issues here and not sure with what devices you're experiencing this.
Currently ZHA polls all group members after changing the group. Z2M (and the Hue integration) just assume the state. ZHA will also have this fixed in the next beta:
This is because HA scenes send individual commands. The same thing will happen if you set up an HA scene and use Z2M (some coordinators work better with such an "overload of messages" though). For example, I have a TI coordinator and a scene with 12 lights changing at once and it works quite well. One thing that could solve this for everyone (weaker networks) would be to implement Zigbee scenes, as it would be enough to send one "reactivate scene 1" command to a whole Zigbee group (drastically reducing the amount of message).
Other than the internal conversion from Kelvin to Mired that HA does (which can be ignored basically), there's no other conversions in that regard. I also don't think Z2M does any correction, but I might be wrong. Never noticed any issues for all of my different lights.
I don't think this would/should be implemented into ZHA itself. I'd also like to see it as an option for all HA sensors though.
I think this is known and will be implemented at some point in time. |
|
For my is Zigbee groups importing but many users is using HA groups that is sending unicast commands to all lights = bad. Calibrating sensors i have make issues / request for but its not being made but now its possible making it very nice in ZHA / devices. Also OTA / device in GUI is on the way. With the problem group status is not updated it can being that you have some lights that is not Zigbee complaint and is not reporting status changes like Philips HUE and need being pulled for getting the host system. In the end its about how mush user is helping with coding and testing in our community system and its 2 system in the top and its Z2M and ZHA and other have falling behind. Most importing for one production system is it shall being realibol !!! |
|
@TheJulianJES thanks a lot for the answer. Regarding Scenes. Maybe I will give it a try one more time. Two more questions... I want to go with EFR32MG21 anyway, as "in theory" it will be easier to migrate to SkyConnect once I will be able to get my hands on it. Which firmware is recommended? I currently running the official one. But you can get ncp-uart-sw_v6.10.3_115200.gbl from here: https://github.com/xsp1989/zigbeeFirmware/tree/master/firmware/Zigbee3.0_Dongle-NoSigned/EZSP Also as settings, if I going to switch one again, I want to use those settings: Any suggestions? |
|
EZSP 6.7.8.0 (Sonoff 6.7.9.0) is the stable version and EZSP 6.10.3.0 is candidate and looks working well. Edit: If trying EZSP 7.X you is getting problem going back to 6.X and need flashing one NVM reset file for getting it working xsp1989/zigbeeFirmware#28. |
|
@MattWestb, i think the official firmware is actually v6.10.3.0 build 297 (ncp-uart-sw_EZNet6.10.3_V1.0.1.gbl) from https://github.com/itead/Sonoff_Zigbee_Dongle_Firmware/tree/master/Dongle-E/NCP. My question was mostly if the official version is any different from the xsp1989 version. |
|
The xsp1989 version is having patches for the hardware and shall working well. The bad thing is xsp1989 have moving on so we is not getting any support at all now :-(( I think both shall working OK but you can trying both but we have not getting feedback form users if some is working better then the other. PS: Thanks for the link i was looking for the router firmware one user is using :-)) |
|
I received my SkyConnect a couple of days ago and migrated now everything to ZHA (once again). :) |
|
Note: Color XY/color temp attribute reports aren't parsed in ZHA (as the Create a Zigbee group with the devices using the ZHA UI. Remember the group ID. In the example, I'll use 0x000a (hex) (or 10 in decimal). Note: We cannot use the modern You can find the commands when using the "Manage Zigbee device" -> clusters UI. To create a new scene: service: zha.issue_zigbee_group_command
data:
cluster_type: in
cluster_id: 5
group: 10 # group id
command: 0 # add new scene
args:
- 10 # group id
- 1 # scene id
- 10 # transition time (in 1/10th seconds)
- Scene1 # scene nameThen, set all devices to the states you want to store in the scene. To store the current settings into the scene: service: zha.issue_zigbee_group_command
data:
cluster_type: in
cluster_id: 5
group: 10 # group id
command: 4 # store state in scene
args:
- 10 # group id
- 1 # scene idTo recall the scene (again, this will not set the correct state in HA): service: zha.issue_zigbee_group_command
data:
cluster_type: in
cluster_id: 5
group: 10 # group id
command: 5 # recall scene
args:
- 10 # group id
- 1 # scene id
- 10 # optional: override transition time (in 1/10th seconds) |
|
@TheJulianJES thanks a lot! |
|
What i is missing is working scene add commands (data sets for all ZCL clusters) with all parameters so can adding advanced scenes for lights and not only storing scenes then its not flexible. |
|
After experimenting a bit, there are a few things to add. To remove a scene this payload can be used: service: zha.issue_zigbee_group_command
data:
cluster_type: in
cluster_id: 5
group: 3 # group id
command: 2 # remove scene
args:
- 3 # group id
- 1 # scene idI have 13 bulbs from different vendors that are involved in the scene, the transition time not acting the same for all bulbs, so the solution is to set the transition time to 0 during scene creation, for example: service: zha.issue_zigbee_group_command
data:
cluster_type: in
cluster_id: 5
group: 3 # group id
command: 0 # add new scene
args:
- 3 # group id
- 1 # scene id
- 0 # transition time (in 0/10th seconds)
- Warm White # scene nameI've tried to set transition time at scene recall, but it was not working for me. @TheJulianJES, @MattWestb maybe it is worth creating some Wiki page for everyone? I think this information is very useful for the whole HA community. If you can point me to where to create it, I will be happy to do it. |
|
You are very welcome adding scene in the wiki https://github.com/zigpy/zigpy/wiki. |
|
@MattWestb Thanks! I will do the page next week. @TheJulianJES, @MattWestb I think I've found what is wrong with Kelvin's Temperature behavior. I've found two issues related to it: So the solution is to use some converter, for example, this one: https://www.luxalight.eu/en/cie-convertor to convert Kelvin to XY values. |
|
OOO !! Edit: The new CWS3 is having working CT in ZHA :-)). |
|
As promised, I've created a wiki page explaining the Zigbee Scenes in ZHA: https://github.com/zigpy/zigpy/wiki/Zigbee-Scenes |
Great work done !!Sadly ZHA have not GUI for Zigbee scenes like de(F)CONZ that is adding then in HA GUI. |
The problem
Hi All,
I've migrated my 100+ devices Zigbee network to the new USB Dongle: Sonoff ZBDongle-E with EFR32MG21 chipset.
As part of the migration, I've migrated from Zigbee2MQTT to the ZHA, as ZHA has official Silicon Labs support.
In general, the network is stable but I have a few strange messages in the log that I cannot explain.
When I am restarting Home Assistant, I see in the log those messages:
During regular work, I see strange NWK conflicts and that devices leaving and joining back the network in the middle of nowhere for no reason.
The Sonoff ZBDongle-E running official firmware v6.10.3.0 build 297 (ncp-uart-sw_EZNet6.10.3_V1.0.1.gbl)
My ZHA settings in the configuration.yaml are:
My network looks this way:
Any explanation for this behavior?
What version of Home Assistant Core has the issue?
2023.1.4
What was the last working version of Home Assistant Core?
No response
What type of installation are you running?
Home Assistant OS
Integration causing the issue
ZHA
Link to integration documentation on our website
https://www.home-assistant.io/integrations/zha/
Diagnostics information
config_entry-zha-fb17b05629f882f6722ee497150796b9.json.txt
Example YAML snippet
No response
Anything in the logs that might be useful for us?
No response
Additional information
No response
The text was updated successfully, but these errors were encountered: