Many devices don't work anymore after updating to v.1.39.0-1. Some fatal errors. #23302
Comments
Krzysztonek
changed the title
|
Those fatal errors happen during the startup sequence but there is a retry mechanism in place to bypass them (due to hardware flow control issues, Silabs is aware). In short, it's a problem, but it's bypassed for now, and doesn't prevent Z2M from starting; it should not be related to your troubles here. Seems devices are failing to ping though. Any chance you have a fuller Did you try clearing the NVM3 on your adapter (backup/clear/restore)? Seems some configs are failing to set that really shouldn't, could be the adapter is having some troubles. |
No, I'm not aware of any method clearing the NVM on my SkyConnect adapter. I rebooted my HA server several times. Unfortunately, it doesn't help/change/repair anything. Still the same set of error messages appears: I'm even not able to get into UI: |
|
The failed pings on startup are likely preventing the frontend from showing up in a timely manner. Current dev/edge has some improvements in that domain. For the NVM3, you can try using the new You can DM me from zigbee2mqtt server on Discord if needed. |
|
I tried to do backup of my skyconnect configuration following the linked procedure. I got the following log: So, I've followed again this video to upgrade my SkyConnect dongle to 7.4.3.0 from this repo. And I failed again: I'm stuck again, really don't understand what is going wrong in there. |
|
Try this:
Since you have a proper coordinator backup created by Z2M, it should detect that the adapter was reset, and reform the network using that backup. If all goes well, this should clear the NVM3 and restore your backup from Z2M. |
|
Yesterday I did the backup: I already updated my SkyConnect dongle firmware to EMBER. So, my first concern is, why is this dongle still visible as Today I launched Still EZSP, not EMBER.. Is that OK? Then I tried to run In the next step I executed: At the end I connected my dongle back to HA server and started Z2M again with the following result: Looks, I've not been succeeded :/ |
|
EZSP is the name of the "middle-man" in the EmberZNet stack (EmberZNet Serial Protocol = EZSP) so it's still there in some names, just the Z2M driver itself is named Can you go into the data folder for ember-zli and retrieve the log file with the failed NVM3 clearing? |
|
Here is my ember-zli data folder content: And this is my Z2M log folder content: I'm not able to start Z2M at all, therefore I can't get into UI to enable |
|
Alright, let's try another method:
Same as before, make sure your PS: you can enable advanced:
log_level: debug |
|
I launched Since some errors appeared, I launched the below set of commands: first
As you can see, during first atempt the I don't understand what is wrong with this damn adapter. The only I can say, seems it's not damaged. I would also bring your attention to one more point (please, see the below screen with From what I see, the recommended version of the firmware is 7.4.1.0 while I already updated the adapter to version 7.4.3.0. As per your advise I didn't downgrade it to |
|
Since you flashed latest before, make sure you use these settings: |
|
I did exactly what you've suggested: This time However, seems this method still doesn't bring any success, since the same errors appear again: I also enabled |
|
I seem to remember you had some past troubles with your |
|
Yes, you're right, I had some troubles in the past. It was caused by having two different Z2M instances installed on the same machine with two different Zigbee dongles connected to this machine, to ballance the zigbee network traffic ( those days I was facing huge lags in my zigbee network).. But this is the past, few months ago I installed a fresh HA system from scratch with one Z2M add-on. Since then it worked fine until this huge failure, which I'm facing now. |
See if you can at least get the logs to show as debug. Note: August release will support concurrency, so you will probably want to remove |
|
I managed to enable However, I also noticed, that regardless It means that |
|
@Koenkk You know the structure of the HA add-on better than me. Could something be cached and make a mess of the settings in |
|
What a lovely Friday! I powered my HA server off for 2 hours, then powered on back. And my Z2M ...started working! Currently only the UI is still not reachable. When clicking
Clicking |
|
There are some improvements for the UI being slow to load on start for 1.39 in current dev branch/edge. If you want, you can try the I am seeing a few repeating issues in your latest logs though: This is happening right around the time the UI should become available. |
|
You have to put |
|
Hello, I seemed to have a related issue (?) with the logs spammed of and then Z2M refusing to start. Reflashing + doing the [not sure reflashing was needed but can't have hurt either] |
What happened?
Some devices disappear or they are being disconnected from the network randomly for a long time. Then only part of them appears back again.
One of the most problematic devices are the roller drivers AM43-0.45_40-ES-EB, which are out of the network since last update. After repairing they remain in the zigbee network for a while, then become randomly disconnected.
What did you expect to happen?
No response
How to reproduce it (minimal and precise)
No response
Zigbee2MQTT version
1.39.0-1
Adapter firmware version
7.4.2
Adapter
SkyConnect
Setup
x86
Debug log
And here is the most frequently repeatable list of errors:
Here is my config:
The text was updated successfully, but these errors were encountered: