have over 50 nodes in the Zwave2MQTT network. Each time I restart the network, a number of nodes are reported dead.
Is there anything you recommend, that I need to do to increase the reliability of the ZWave network?
Thank you
Need a LOT more info.
What kind of nodes. Any patterns? Are they all in one location?
What ZWave coordinator do you have, what firmware is it running? How is it physically connected to the machine running ZWaveJSUI?
Nathan, the nodes coming up dead are light switches (UltraPro, Jasco or Zooz) on all three floors, and it looks pretty random. Switches that are good this time, may show dead next time.
The ZWave Stick (ZST10, FW v2.3, SDK: v6.81.3) is in the basement, connected through a 4ft extension cable to the NUC, running the latest Homeassistant version .
The switches are on the latest firmware as well.
The network has been updated already with one hop between nodes (see attached network graph as overview)
High traffic can cause all sorts of issues, Whether that is your problem or not I do not know.
Look at this post for a way to analyze the traffic on your network.
Thanks. I have completed all my changes to the ZWave network and will wait, until the system is “calming down”, before restarting again eventually. As shown in the screenshot, traffic may be high and there were a lot of dropped packages when re-interviewing failed items.
Since I am not too familiar with Linux, I may need to learn, how to setup and load a script in HAOS.
I appreciate your help and hope that things will settle.
Not sure what is the root cause yet, but this morning, one other node showed up dead. When I re-interviewed it, I had to turn on and off the switch to make sure it was set alive again. Then two other nodes showed up dead and finally that first one again. And that all while there should not have been a lot of traffic on the network.
I am still not sure, whether I did something wrong with the network setup
Just another update: I did shut down my entire system, waited for an hour, restarted. 5 nodes were reported dead. I then restarted the ZWave JS UI addon, and after the restart these 5 nodes were good, but another node was reported dead. This is pretty awkward
Awkward indeed. Instead of re-interviewing them - does a simple ping bring the node back?
I tried it, but no success. So I had re re-interview these nodes
Just to add to the frustration: some of the nodes even forget their entire configuration on a random basis. see the screenshot below.
I feel like I do something fundamentally wrong and do not know what
I realized that some nodes randomly continue to go dead. In most cases the ping works. In some cases I pinged another 1-2 times. SOme nodes report back as unknown and I had to remove and reconnect them in the last few days
Quick update after several weeks of operation with the new config:
One of the switches that frequently went dead was a Zooz Zwave switch. When I swapped it out for an UltraPro, the entire network started to stabilize. I am not sure, whether it was a coincidence, or whether this switch caused so much traffic, that the entire network was impacted.
After swapping out the switch, the entire network stabilized. I have not had a Zwave device going offline since then.
This is not that uncommon it seems. A single misbehaving node can destabilize the entire network, to the point of bringing it down entirely. As you experienced. Good you found the culprit !
From a technical point of view, I find this quite strange though. I’m having a hard time imagining a repeatable hardware malfunction that would either continuously spam the network or start replying to messages not intended for the node, etc, that isn’t completely obscure and a once in a lifetime freak thing. In fact, since people have reported these incidents before (often with Zooz), it seems to be more or less common, relatively speaking. It sounds more like a firmware bug that only manifests in certain situations. Maybe there’s also a problem deep down in zwavejs too ? Would be interesting to know how these degraded networks would behave with different controller software.
Anyway, good you solved it.
Agreed.
To be honest, I was about to dump all zooz devices and replace them With GE/Honeywell. I generally like the Zooz devices. more capable and based on the 700 series. But maybe that is also the issue.
Another problem I have with Zooz: when the problem became that difficult, the blocked me, ghosted me even on Facebook and did not respond to any issues.
Glad that I found the founder and owner of Zooz and The smartest house. I will reach out to him and Lorenz Engineering, the mother ship and see if this is appreciated even by their management
I’m having similar problems after Zwave JS UI update 1.13. Every Zwave device on my system stopped working. The nodes are reported dead. I use a Zooz ZST10 700 controller. Subsequent updates to 1.13 have not corrected the problem. I’ve also upgraded the firmware of the Zooz stick.
My HA runs on a mini PC. It’s been rock solid until this. Reinstalling from a backup did not fix the problem.
I have precisely the same situation as @Lonestar10: the entire network is dead since the May HA release (which may have required an update to zwavejsui?). I’ve tried pinging, re-interviewing, healing, praying… now I’m thinking its reset and re-build apparently.
F’n z-wave. Picked the wrong horse a few years ago and it might be time to abandon that sunk cost now. Shelly wifi seems to be the go for powered device, and Zigbee (or Thread one of these days?) for battery devices.