wastez
commented
2 weeks ago @Nerivec What do you think about it? The problem was already before 1.37. I don’t think the problem is in the driver.
Open wastez opened 2 weeks ago
wastez
commented
2 weeks ago @Nerivec What do you think about it? The problem was already before 1.37. I don’t think the problem is in the driver.
Nerivec
commented
1 week ago Can you provide some debug-level logs of when this happens?
wastez
commented
1 week ago Ok sorry missed this message
wastez
commented
6 days ago Ok here it is. Network is to busy because of multicasting, you will see it in the logs. debug.log
wastez
commented
6 days ago Can something be done here?
Of course a firmware with higher bitrate could help.
wastez
commented
6 days ago @Nerivec Any thoughts?
Nerivec
commented
6 days ago The Zigbee spec prevents more than 8 broadcasts over a 9-second period. Silabs docs (see paragraph "Broadcast Table").
That's ~160 broadcasts in roughly 1 minute.... (broadcast/multicast, same limit applies) So averaged that's 2.67, when the limit is 0.89. Definitely too much. 😅
I've been wanting to run some tests on this, see how we could maybe "optimize" it (though definitely not to that extent), but haven't had the time yet, plus I don't have enough devices in my testnet to run this effectively at the moment...
wastez
commented
6 days ago @Nerivec
Ok i can understand that this is much but then z2m groups are nothing to change the colors. If i use home assistant groups then it is working without problems. (clear no multicasting) Can i help you with the tests?
diegocjorge
commented
2 days ago I have the same issue. I have a total of 95 devices, with 57 of them organized into 14 groups (each device belongs to only one group).
My log is full of errors such as:
error: zh:ember: Delivery of MULTICAST failed for '65533' [apsFrame={"profileId":260,"clusterId":768,"sourceEndpoint":1,"destinationEndpoint":255,"options":256,"groupId":13,"sequence":145} messageTag=32]
info: zh:ember:ezsp: Received network/route error ROUTE_ERROR_MANY_TO_ONE_ROUTE_FAILURE for '5742'.Rarely, I also encounter this:
info: zh:ember:ezsp: Received network/route error ROUTE_ERROR_SOURCE_ROUTE_FAILURE for '22550'.Is there anything I could do to improve the situation? I already use channel 25 and have my 2.4GHz WiFi on channel 1.
I have a spare Sonoff dongle-P. Do you think it would be better to split my network into two?
Let me know if you would like for me to send some logs.
Nerivec
commented
2 days ago @diegocjorge That would not be the same problem. Yours are failing because it looks like the network isn't very stable, and it can't reach devices properly all the time. Try putting your logs (info level is enough) through this little tool I made https://nerivec.github.io/z2m-ember-helper/ and see what you get (you need several hours worth of logs to get decent feedback, the more the better).
diegocjorge
commented
18 hours ago @diegocjorge That would not be the same problem. Yours are failing because it looks like the network isn't very stable, and it can't reach devices properly all the time. Try putting your logs (
infolevel is enough) through this little tool I made https://nerivec.github.io/z2m-ember-helper/ and see what you get (you need several hours worth of logs to get decent feedback, the more the better).
Thanks @Nerivec, I ran the tool and encountered a few issues:
Broadcast Table Full Detected - This can indicate the network is relying too heavily on broadcasts (messages to the whole network or to groups): Could this be why some of my grouped lights don’t always change color correctly?
ROUTE_ERROR_MANY_TO_ONE_ROUTE_FAILURE: These errors are mostly occurring with Gledopto’s GL-C-001P (5-in-1 Smart LED Controller).
ROUTE_ERROR_ADDRESS_CONFLICT: The tool detected 26 instances of this error, but the device hex does not match any device in my network. Do you have any idea what could be causing this? Could it be an old device still lingering in the database?
CCA Failure Count is High: This is odd because I ran an RF environment scan with my Ubiquiti AP, and it reported that WiFi channel 13 (frequency 2462~2482 MHz) has only 13% utilization and -87 dBm interference. Therefore, Zigbee channel 25 should be mostly clear, and CCA failure shouldn’t be so high.
There might be something broken with my database. I’m considering completely removing Z2M files and starting over from scratch. I'm just pondering whether to keep everything in one network or split it with my old Sonoff Dongle-P. Any suggestions?
Nerivec
commented
17 hours ago NETWORK_BUSY returned for messages. The broadcast table can hold 15 messages (standard size), you need to reach over that limit before it starts actually being a problem. See "Broadcast table" if you want more detailsAre you using the default Z2M network config (i.e. didn't used GENERATE when you created the network originally)?
Any chance there's another Zigbee network around you?
Find me on Discord if you can, when you have some time, we can run some tests, figure out what's best. It should be fine with one network, but you could put "problem routers/devices" in another one for example so they don't mess with the rest.
What happened?
Since I while I’m using z2m group for a group of lights (6 spots). The problem is that if I change colors using the color pallete that it is not working good. Sometimes they are changing the color really slow or some of them do not change the color. If i use Home Assistant groups I don’t have the problem, it is working as it should.
My stick is a sonoff dongle e with ember. I‘m using the z2m 1.37. 0.
What did you expect to happen?
No response
How to reproduce it (minimal and precise)
No response
Zigbee2MQTT version
1.37.0
Adapter firmware version
ezsp ncp 7.4.2
Adapter
Sonoff dongle e
Setup
Addon on Home Assistant
Debug log
No response