Closed zeddski closed 3 years ago
@zeddski Please attach logs. Follow: https://zwave-js.github.io/zwavejs2mqtt/#/troubleshooting/bug_report
Thanks Robert, relevant logs attached. zwavejs2mqtt.log
@zeddski I also need zwavejs logs please :)
From zwavejs2mqtt logs i can only see the interviews fails, I don't know why
zwavejs_1.log zwavejs2mqtt.log
I made two new ones, so the timestamps match. Curious to what you can figure out of these...
@AlCalzone Could you give the logs a check? Seems that the Interview process fails on the RequestNodeInfo
From what I'm seeing in the logs, I'm getting the feeling that Node 5 has really bad connectivity. It does respond from time to time, but most commands result in
09:45:45.215 DRIVER « [REQ] [SendData]
callback id: 12
transmit status: NoAck
which means the protocol-level acknowledgement that tells us a message was received didn't come back. My first thought went to "this thing sleeps and didn't tell the controller", but power plugs usually don't do that, since they are permanently powered.
Do you have a chance to move the device closer to the controller to test if that improves the situation? If so, a network heal could solve your issue.
OK, what I did:
I do appreciate all the attention :)
I don't see the node going dead in the new log. So looks to me like it was a connectivity issue. I'd put the node back where it belongs and perform a network heal, so the routes can be regenerated.
Hi, I have similar issue with a device which goes through a plug to the controller.
[Node 005] did not respond after 1/3 attempts. Scheduling next try in 500 ms.
I also see this message and the device is declared dead only to be revived in less than a second with subsequent messages. @AlCalzone do you think increasing the 500ms waiting period can help with remote devices or the ones that go over multiple hops?
That message is a result of the controller callback, which we have no control over. The 500ms is the wait time before we try sending again. Increasing that won't help if the ACK does not come back.
I don't see the node going dead in the new log. So looks to me like it was a connectivity issue. I'd put the node back where it belongs and perform a network heal, so the routes can be regenerated.
So it seems. It even helped "finding" other devices that were gone. I'll put a node between this Node 15 and the zwave-server, to increase network coverage.
Leaves me with node 18, the PIR. It is sitting very close to the server but still, it doesn't show up in the device listing. Walking through the logs, it does show up now and then in the Zwave-logs, but very sporadic, even though I've triggered the PIR a lot. Several other devices (like a door sensor) don't show up either, although they are fairly close to the server. And remember: a week ago, when I was still using Open Zwave, all these devices worked fine. I'll do a network heal anyway, but hints and new insights remain very welcome!
Node 18 does not respond to any node info query, so we don't know what it supports. OZW likely already did, which is why it used to work. It also sends all its reports as broadcast, so it seems like it is set up strangely.
Node 14 for example is registered as a listening (always on) device, but does not acknowledge any messages.
Network heal is probably a good idea.
I've been able to get Node018 to provide information through a "re-interview" command and forcably waking up the device. I needed to re-awake it several times, but FINALLY it shows up in the WebUI, correctly. For anyone who's interested: waking up the device meant putting it in association mode by pressing the button three times. I did this no more than four times with 20 second intervals, which led to my final success. For future reference, I've included the relevant logs.
Thanks for all the help! I'll move the ZWave-part to a different RPI and place that one on a more central location in my house. Under the stairs, probably...
For anyone who's interested: waking up the device meant putting it in association mode by pressing the button three times.
Yep, it is a common misconception that if a device does something it must mean that it is awake. Short - it is not and you have to physically wake it up with a button push or something.
Last week migrated from OpenZwave to Zwave-JS. Network key seems to be correct, since the right amount of devices are listed. But, information about the devices is gone. Sometimes one of the non-battery-devices returns for a little while, then disappears again. I'm using Home Assistant too (both in a container).
I've enabled logging, which isn't helping much yet. I'm concentrating on two devices. Node 5, which is a Zwave power plug and Node 18 which is a battery powered PIR (motion sensor). The PIR shows up in logs, but no information is shown in the device listing. Removing and re-associating gives me a new device with (again) no information.
So, let's concentrate on Node 5, because that node logs every few minutes. As can be seen below, it logs "has returned", just to immediatly be declared dead again. After a few seconds, the node's information rolls down the screen. Every five minutes something lkie this happens again: it gets declared dead, with some updated values directly following the RIP report (limited to 3 log rules, not as extensive as this example).
Does anyone have a clue to what might be going on and how to debug further?