Zwave mesh grinds to a halt + HA nodes start growing new (incorrect) capabilities #6295
-
I've held off on posting this because I have been trying to better-understand the problem, first. But I've been fighting this for awhile and am stumped so I'm posting in hopes of a pointer: The following issue occurs every 12-72 hours or so:
The solution to the Zwave network being jammed up is simply to restart ZwaveJS and I'll be back to normal for 12-72 hours or so. To clear the nonsense notifications (ie, motion detected from a smart switch) I need to re-interview the nodes with the impossible capabilities. The issue is happening this morning as I am typing this (ie, right now) and I happened to get up early enough that I was the only person awake -- so rather than fixing it, I was able to let a Zniffer run for a bit. It appears to me that the controller (node 1) isn't transmitting at all, which I guess explains the flood of messages received as nodes keep retransmitting over and over. I'm going to today's zwavejs log and the Zniffer capture, would love any insight or suggestions on additional data points to look at. I always struggle to pinpoint the "start" of the issue, but it looks like retransmits started overnight sometime shortly after 2023-09-21T07:00:00'ish. I restarted the ZwaveJS instance at 2023-09-21T11:25 so the Zniffer will show signs of life from node 1 after that. Thanks for any help. |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 3 replies
-
Saw this in 11.14.3 notes: The logs I attached to this one don't show this behavior, but I have seen in past events where ZwaveJS logs were complaining about "node not found" -- and the node ID(s) it was complaining about wasn't even possible in Zwave (node IDs above 232). I upgraded to 11.14.3 just now in case this is part of my issue. |
Beta Was this translation helpful? Give feedback.
-
2., 4., 5. are this issue: #6260
Correct! The controller gets stuck in the middle of a transmission and Z-Wave JS's recovery mechanism is not enough to get it back to work normally. I'm not sure which firmware version your stick is on - all reports so far have been on 7.19.x. For 1., I need a bit more help when this happens. There are a few instances where queries take more than a second, and they are appearing in combination with re-transmitted responses, so likely a temporary connectivity issue. For 3., I also need more specific information, at least node IDs. Could also just be this: https://zwave-js.github.io/node-zwave-js/#/troubleshooting/nonsensical-values?id=nonsensical-values-appear-randomly |
Beta Was this translation helpful? Give feedback.
2., 4., 5. are this issue: #6260
Correct! The controller gets stuck in the middle of a transmission and Z-Wave JS's recovery mechanism is not enough to get it back to work normally. I'm not sure which firmware version your stick is on - all reports so far have been on 7.19.x.
For 1., I need a bit more help when this happens. There are a few instances where queries take more than a second, and they are appearing in combination with re-transmitted responses, so likely a temporary connectivity issue.
For 3., I also need m…