Maybe you can find a bug and help others? What’s your timezone? I’ll tell you what time to look.
leviton is the issue i think I have the same problem for some leviton device, you need to un-enroled then enroled again useally it help but sometime you need to replace it (warranty replaced one already after about 1year and half, and like 6 month after I again have problem with that brand, switch)
every other device work like a charm
I have 4 leviton devices, they all go offline now and then. I’ve found that ‘ping’ usually brings them back, but maybe not always. Which is also frustrating - if that’s all it takes, couldn’t HA do that for me somehow?
Write an automation to make HA do it for you. That’s why ping and refresh_value came into existence. Gives the user the ability to fix hardware that doesn’t follow zwave specifications or if the hardware has bugs. Also, you might want to try zwave_js.refresh_value instead of button.press with the ping button.
Didn’t know I could ping from an automation. Yes, I could create one to ping the Levitons every now and then and that might avoid most of the problems. Thanks.
I can all but guarantee you that you’re barking up the wrong tree looking any specific device or brand as the root cause of your issue (nodes going unresponsive).
In that vein, creating an automation to ping nodes or including pings in existing automations may alleviate the issue, but it’s really like putting a band-aid on a severed hand.
The issue is that you’re using a Z-Wave 700 stick that hasn’t been updated to the newest/correct firmware.
How do I know this? Because I see the same thing with a Zooz 700 stick running the next-to-latest firmware. The same devices all running on a Gen5 (Z-Wave 500) stick have no such issues.
In my case, the devices affected are typically ones that are triggered solely by automations and never have manual switch operation. And that’s because manually operating an affected switch, like a ping, will bring it back online instantly, thereby fixing them before I ever notice anything has gone awry.
That’s a reasonable theory; but actually, I hardly ever actuate any of my devices manually, it’s all on automations. And the Levitons are, seriously, the only ones that drop off.
I know I should update the controller. That’s a big reason I made the switch to the MQTT version. Then I found out it could only update the devices, not the controller, so I’d have to do it on a PC. And then I saw posts from people saying they’d updated the controller and lost everything - all the devices had to be re-added and renamed. And that, unfortunately, is a possibility I cannot allow, as some of these devices were very uncooperative in the initial inclusion process, requiring multiple attempts.
Z-wave is a “hobby farm”.
I don’t know what else to tell you. You can replace any switches you want - that won’t solve the root cause. If you find it more fun to randomly re-wire switches, who am I to stop you? 
I can do electrical work in my sleep, yet it’s not something I’d choose to do unnecessarily.
The Z-Wave 700 stick firmware has a critical flaw which can affect one or more devices. I don’t see much of a rhyme nor reason even after looking at my own patterns and those of others who reported the issues last year. On my setup in testing it happens to a couple of Zooz Zen26 switches via automations and on a Zen21 (if I’m not mistaken) if I hammer it with commands (repeated calls to activate or toggle the switch).
If you don’t have firmware 7.17 on a series 700 stick, you’re going to have a bad time with Z-Wave, and it doesn’t really matter if someone has only few devices.
I would advise anyone who is not interested in jumping through the hoops to not buy a series 700 stick unless it is confirmed to already contain the 7.17 firmware released this year. If you get anything without that firmware return it immediately or make the update. For less headaches and “work” use a Gen5 stick.
I just wanted to also mention I’ve now done the update / downgrade cycle 3 times while testing, using the PC-based software without any issues. No nodes or configuration lost. The software is absolutely horrible, but following the poorly written instructions wasn’t difficult and worked as described. You must create an account (but can use all fake information along with a real email address) AND you must log into that account from within the software otherwise the necessary modules/ass-ons for Z-Wave that are needed for the firmware update won’t appear.
Well that’s good to hear. The whole thing, though, leaves me asking myself why I would buy more z-wave devices. It seems like the z-wave industry has some growing up to do.
IMO, if you’re not willing to see things break at least a few times per year, Home Assistant is the wrong platform. It’s a growing and evolving platform, changing at a fast pace with code touched by dozens of developers.
It’s generally very good, but with this kind of development and release cycle, it’s what companies in many other industries would call a perpetual public beta test.
When I ran a Gen5 Z-Wave stick with Indigo Domotics, which has by comparison a snail’s pace development cycle, I had no such issues (in fact no issues whatsoever) with Z-Wave. Ever. On the flip side, it was a right pain to create the kinds of automations I can with HA and there was no support for Zigbee and countless other integrations we take for granted with HA.
HA is great. I’ve basically been able to do wanted to with it, and it’s 95% reliable. The 5% is pretty much confined to ZWave. I’m a (former) developer, I expected problems and confusion and worked through them. It was only when I found out about the hoops I’d have to jump through just to update one device (the 700 controller) that I said - wait a minute, I’m drawing a line here.
Wifi is far from perfect, you have the maddening proliferation of cr@ppy, buggy apps to configure the devices - but at least you’re guaranteed of that. HA is all in on Linux - that’s great, but the ZWave makers are not, so updating ZWave devices on Linux is problematic. In retrospect I’d have had an easer time running HA on a tiny Windows box instead of the RPi4, but I guess wanted to learn something about Linux.
That’s not true at all.
There are instructions linked way up above for flashing the 700 controllers in Linux. I used an existing Windows VM installation instead.
As I run HAOS in a VM on Unraid, I just shut it down temporarily and fired up a Windows VM, attaching the same USB port the controller was already plugged into. So I was able to do it all from my Mac, remotely without having to go down to the basement and touch my server or Z-stick at all. Win-win.
I am NOT a Z-Wave expert. But I recall that before updating the firmware on the dongle you have to back up your device data because the upgrade erases everything.
Backups are always a good idea. But the rest of the info is incorrect. The firmware update doesn’t erase anything.
If the dead nodes (ping to revive) issue is (still) caused by the 700-series sticks, then the latest/current firmware (7.17.2) doesn’t fix it.
To determine with certainty this is caused by this series stick, I’ll need to copy/migrate my network to my older 500-series stick and then go another round of testing. In multiple years of running Z-Wave on 500-series stick with Indigo Domo, I never saw such an issue, so this testing will also help figure out if this is something that’s Home Assistant/Z-Wave-JS specific.
Thank you for this. I think it’s the best explanation about what are the main differences between them.
At this point. Do you think that zwaveJS2MQTT is as stable as zwaveJS add-on to be considered as a fully valid option?
They both run the same version of Zwave JS
How do I find the address for the community add-on?
Thank you!