Hi everyone, long-time user first-time poster.
I’m running a raspberry pi using a aeotech gen 7 stick with latest firmware.
My problem is that the entire network slows to a crawl and some devices fail to being “dead” causing automations to fail. It’s 60 devices and I migrated from hubitat to the Nortek security stick combo then to this gen 7 stick. On the nortek the entire network would fail after about a week or two so I’m at wits end in general. When it gets to this state the web UI freezes even after trying to just turn on a light from the ui so I’m not sure what to do as everything I try to look at for logging is hard to even get to load.
I’m using zwavejsui as I switched to that to hopefully see a more detailed view of my nodes but nothing is showing up at all. I used the built in zwave utility to add in all my switches so the ui did only show each device as a node and not any of the named devices but it did work correctly.
I did have a switch that was at far end of my zwave network that created a similar problem. It slowed HA to a crawl and everything on the host was bad. I run xeon server so I still find it weird that it created this issue but I checked extensively and it was crashing everything. Cant remember if this was when I was using Openzwave or after zwavejsui and not sure if that would have made a difference. I eventually removed that switch and the problems went away.
Take a look at the statistics for your zwave devices. Look for anything showing communication problems like retries in the 50+ or lots of drops.
Linkplay is used for whole home audio but was giving me errors becuase one of the devices is meant for used rarely so ive just disabled that in my configuration.yaml file. The rest are very basic and dont give me any trouble in my other installs but this one which is such a big scale seems to be a pain in the neck.
My main issue is that the zwavejsui isn’t showing anything anymore as i would like that to diagnose more in-depth besides just zwave silly logging level.
Edit: VPNing into the site fixed this issue as nabucasa connection wouldnt show this at all, now looking over logs with toggling lights to see where it comes to issue, may have to wait for lights to fail again and go from there.
Do you use an extension USB cable, or is your Zwave stick connected directly in the Pi?
A lot of users here solved this kind of problems by using an extension cable, separating the stick from the Pi. There is a well described interference problem with radio sticks placed in the Pi USB directly…
I suspect zwave is the cause of your issues, as you’ve had the same problems across 3 different controllers.
Zwave is very low bandwidth, and it doesn’t take much to grind it to a hault. I would check the following:
1- be sure you use s0 security only when absolutely necessary, like only a door lock. S0 security uses a lot of bandwidth and should be avoided
2- look for a “chatty” or misconfigured node. I had a node mis-configured to send power management reports every 0 seconds. It’s continous reports crashed my network. Disable power management reports if you don’t need them, and if you need them, change parameters to make them less frequent. Check your wavejs logs and see if the same device is constantly showing up. You can try posting the zwavejs server logs here if you need assistance finding a problem device.
sadly most of these switches are older ge enbrighten that use s0 only 6 can use s2 authentication and 3 failed to use it even after trying to re-set them up. I’ll try to set those 3 up again with s2 but it seems like they didn’t wanna use s2 (they’re eaton smart plugs from lowe’s and I’m considering changing them out as I’m thinking they’re the cause of my problems).
You really don’t need encrypted security for switches. Zwave still has a degree of security for any included device, even without s0 or s2, and it would take someone physically present with sophisticated tools to hack a zwave network.
S2 is fine for bandwith, including switches, but any device that isn’t a lock, I would include with no security. If security is a concern (lots of neighbors, big city maybe?), I would replace those devices with new ones that support s2 security.
Funny you guys bring this up. My network is doing the exact same thing… I went and checked and all my GE 14xxx stuff I joined when I moved over from SmartThings joined S0… I’ve got a bunch more (20-30) switches joined S0 than I wanted or expected… I saw joined with security and assumed (yes my problem) they were S2…
And now I remember about a year ago something changed in the default join to only use S0 if necessary… I did all of these before that.
(edit: @Dodotech I have no need for security on my light switch, I just plan to exclude and rejoin them without security - what I had planned originally anyway)
I thought all my zwave problems were from the Smartthings cloud when I was on their hub, and when they continued in a cloudless Home Assistant, that’s when I really went into the logs and saw all the issues with “nonce errors” on the S0 devices and the endless power management reports from some of my Inovelli switches.
Turns out I couldn’t blame Smarthings for all my problems
Guess ill have to keep an eye on it as I’ve done an upgrade to this network already going from wemo to z wave. Don’t really wanna upgrade all of them to s2 especially since it used to work flawlessly but I’m thinking it had a bad switch somewhere that overloaded the system. I’m concerned that those eaton plugs are causing issues because of their bad setup. Nothing is reporting power just dimness state/on off.
Is it better to use s0 for switches or use s2? I don’t need the security but you made it seem like s0 uses more bandwidth which could’ve been tanking the network. Right now I’ve got 5 switches with dropped packets in zwavsjsui 2 of which are s2 switches that have dropped RX. Since this isn’t my house I don’t know the specifics other than the regular automations stop. And I was out of town dealing with some family things so now I’m getting back with the client saying it’s broken again.
I’ve also implemented the automation that pings dead z-wave nodes as that was an issue that I ran into and found a thread online about. I’m still having issues now and trying to find a way to sniff about which node is causing the trouble. Right now I need to figure out how to enable silly level logging in the background for z-wave as the regular home assistant logs don’t show it.
S2 isn’t so bad as S0. It rarely has the same effect. I would not expect it to be the cause of those nodes dropping.
If you have a poorly performing node. It can absolutely cause issues on other devices. Debug and silly mode will be set in the settings of ZWaveJS_UI. It’s under the ZWave configuration section of settings. In this case I’d suggest troubleshoot what is happening before you start randomly rejoining nodes. Get the logs.
In the logs of zwavejsui it only saves a small portion of the logs which isn’t helpful unless I know that it’s having issues in real-time, I can’t seem to find an option to enable longer retention of the logs.
whoops was checking settings in the addons section, not the actual ui settings.
I’ve set 30,000 as the log amount, I’m gonna give everything a update and reboot then wait til i get a call with issues from there.
