My openhab particularly Zwave has been slowing down (response time). I’ve been using an sd card on my build since 2021 and have had multiple power outages. I also had a nortek HUSBZB-1.
I bought an NVME with external USB enclosure and a Zooz ZST39 zwave stick to try to freshen things up. I know all the features aren’t available on the ZST39, but what I read in the forums is it still works. When I first set it up it worked, but all the devices go to ERROR:COMM. When I pull the zstick and put it back in the devices work for a while then stop again. I had chatgpt analyze the debug log files and it couldn’t see anything wrong.
I cloned the SD card to the SSD and it mostly worked. It deleted the log folder, a restart would delete log folders I created and remove persistence. I was able to get the log folder and persistence fixed, but I wonder if other folders were also deleted. I’m assuming I have an errant zwave device. Can anyone help? Thanks!
Log file: openhab.log - Google Drive
Raspberry Pi 4 8GB Model B
Openhabian 4.2.2 - Release Build
Assuming you started with openHABian, there is a feature to preserve the SD card by limitiing writes called zram. If you are now running on an SSD you should turn off zram. Folders with lots of writes (persistence, logs) are saved to zram instead of the SD card and only written out to the SD on a normal shutdown. If you pull the power there will be no write.
Note that running openHABian on an SSD, while it works, is outside a supportable configuration.
But all of that is probably a red herring. None of this is going to impact the performance of your Zwave network.
I’m not an expert in Zwave so am limited in what I can provide help for on that front. But I can say that the move to the SDD isn’t going to impact the Zwave in any meaningful way.
Thanks. I verified that zram was not being used.
I’d run the file through the zwave debug log viewer here. I think you will see the problems. On a quick scan some devices seem to be sending reports every few milliseconds. Also disable the network heal until messages are under control, it is just adding to the congestion.
Also fyi do you know what SDK is on the zooz ZST39? Possibly the zstick is not acking the messages received from the device, so the device keeps sending
I’m not sure which SDK was on the ZST39, but I downloaded and flashed the latest from the Zooz website. Everything shows online, but nothing works. I get
2024-11-11 17:57:14.564 [DEBUG] [ding.zwave.handler.ZWaveThingHandler] - NODE 3: Polling…
2024-11-11 17:57:14.565 [DEBUG] [ding.zwave.handler.ZWaveThingHandler] - NODE 3: Polling deferred until initialisation complete
It has been several hours since I flashed the ZST39 with a few restarts inbetween.
I can’t pick any problems out of the debug log viewer.
I use a rough guideline that the max ZWave capacity is 10 messages a second. I recall the log above far exceeded that, so I think congestion is the problem, but not sure why the devices are sending repeating messages.
Anyway the two messages mean that node 3 isn’t initialized, possibly due to congestion or something else.
Another option would be to just turn some devices off and let the network stabilize, check the debug logs for messages and slowly add devices back again until you see weird behavior.
After upgrading firmware on the ZST39 I am unable to get a single device to work.I rolled back to my Nortek controller and am adding devices in small groups and leaving them for at least a day. I’m at 35 devices and so far everything is working quickly.