Lots of log messages and devices become unavailable after several hours #55
Comments
|
The message 8035 probably means the pdm of the Zigate is corrupted. |
|
@doudz Does my problem on home-assistant/core#40729 is the same that this bug ? |
|
no I don't think it's related |
|
when it stop working, it is always the same message to every request sent to zigate. Trying to turn on two different lights, which were working previously. |
|
per log it did stop working after scanning neighbours on the coordinator. I have only two routers on the network and 1st scan was successful, but the 2nd one failed and after that all outbound requests are failing with Here's the log of the failure. Scans are done every 20min, HA was started at 17:16:xx Going to disable scanning of the coordinator. |
|
seems like scanning any device on the network causes problem. Still not clear if the problem is with the zigate itself or the library. |
|
@doudz I can confirm, sending ZDO Mgmt_Lqi_req request to routers on a regular basis is enough to trigger this issue. I suspect something is happening with zigate, as I stop receiving any trafic from zigate. Initially I thought that scanning the Zigate was the problem, but turns out it locks up even if I only scan devices on the network. @bckthomas what zigate version are you running? |
|
Could you try something when you get error 166 : Put ZiGate in permit join mode by clicking on "+" button just like you want to pair a new device and see if it's working again ? |
|
it actually does get it working again. |
|
What zigpy stack is really performing when doing topology?
Do you mean it does topology all of the time ?
Envoyé de mon iPhone
… Le 6 oct. 2020 à 00:15, Alexei Chetroi ***@***.***> a écrit :
it actually does get it working again.
Had scanning disabled, it worked without issues for 2 days.
Enabled scanning -> couldn't control lights after approx 1 hour
ran through joining a device, without actually joining anything and could control lights again
—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or unsubscribe.
|
|
Thank you, we're investigating, it could be an internal ZiGate firmware bug |
|
@doudz, @Adminiuga I would be interested, as we are using the the LQI Management Request ( 0x004e) in our plugin and do not saw any issues. That is why I'm interested to understand what the Zigpy lib could be doing different to fall into a firmware issue ! |
|
The difference is that zigpy doesn't use the ZiGate command 0x004E, all is done using the command 0x0530. |
|
Really not sure but maybe we could try to force source endpoint to 1 https://github.com/zigpy/zigpy-zigate/blob/dev/zigpy_zigate/zigbee/application.py#L135 |
|
I could give it a try, but Mgmt_lqi_req is a ZDO command and AFAIK all zdo commands are using endpoint id 0 for both src and destination ep id. |
|
yep, endpoint must be 0 for ZDO command in Zigbee. |
|
I manage to reproduce the problem. It seems to be linked to the raw mode |
|
Hi, |
|
Just updated the PiZigate. It's on HassOS so had to jump through some hoops. |
|
so far so good. about 1hr 45min uptime. Before it would go down in about 50min -1 hour |
|
ok, it's been almost 24 hours and it is still running strong. Can control my two test light bulbs just fine. Thanks again for a quick fix. Any ETA when it's going to be generally available? |
|
I think next monday |
Hi
I have profalux covers and a zigate to connect them to homeassistant.
Since the last version, I have a lots of message in logs and my devices become unavailable after 5-6 hours.
I have to open the zigbee network and click on the "device reconfiguration" for all covers to work again. But after 5-6 hours, they all became unavailable again.
Let me know if I can help you with more logs, tests...
Thanks
Related issues :
home-assistant/core#40707
home-assistant/core#40729
Here are the logs when the devices become unavailable : (below, all the logs messages)
The text was updated successfully, but these errors were encountered: