Z-Wave Locks Up Needing Full Power Cycle

Our relatively new Hubitat C-5 (Australian edition with Hubitat-supplied USB stick) version is connected with circa 100 Z-Wave devices and very regularly (ie 1-2 days) ceases being able to send commands to the Z-Wave network. During these times the Hubitat log shows either no events occurring at all, or sometimes a flood of events where there are considerable duplicate events with timestamps differing only by a few milliseconds.

I followed the guidelines for building out the Z-Wave network, adding only 20 devices at a time, having no automations at all (we mostly use Alexa at present), avoiding polling (we have about 25 Aeotec MultiSensor 6 and they only communicate meaningful changes, with a 6 hourly report aside from that), locating the Hubitat in the centre of the house, running Z-Wave repair after each change (it takes about 5 minutes per device and always completes) etc. Nearly all devices are mains or USB powered and thus repeating. After each Z-Wave repair I always see this mesh communication issue and have to physically cycle power the Hubitat.

A standard Hubitat "reboot" (ie authenticated POST to http://hostname/hub/reboot) is insufficient to resolve this. Pulling out the USB stick and putting it back in does not resolve this. The resolution is a physical power cycle of the Hubitat. This always works and things are fine for another day or two.

I have ordered some Z-Sticks in the hope they'll offer enhanced disaster recovery options and tooling to investigate the mesh itself, but I was hoping someone could perhaps offer other suggestions here? I have a cold spare Hubitat and Australian USB stick but is better to try the new Z-Sticks with the existing Hubitat or something else (separate the Z-Wave devices onto different networks etc)?

Doing a reboot is not a great idea.
A better idea is doing a shutdown,. wait for 5-10 minutes, and then powering up.

You have possibly corrupted your database.
Look up in the documentation (and in this forum) on how to do a database soft reset.


You should mention all devices you have. Maybe you have a bad repeater.

@jtmpush18 I always shutdown before cycling the power so would be surprised if became corrupt. Nevertheless I did a soft reset and restored a recent backup.

@vjv the devices are (in order of quantity) Fibro Dimmer 2, Aeotec Multi Sensor 6, Fibro Switch 2, Aeotec LED Multi Color 6, Fibro Universal Binary Sensors, a couple of Aeotec WallMotes and one Aeotec Smart Switch 6. Everything is powered except the WallMotes. All devices in the radio list are still active, with one exception (a Fakro window opener we were testing and packed away until install).

How do I find if I have a "bad repeater"?

1 Like

Did you remove the Fakro from the hub? If not you should otherwise it might be the sticking point if other devices are trying to route through it and it's not there.

1 Like

I hope you’ve done it already because it’s pretty standard practice and mentioned all through the forum but have you disabled all ‘community’ Drivers and Apps AND then let the mesh heal ?? (Could be more than 1 hour wait !! )
You should have found that there are some ‘dubious’ Apps that Support@Hubitat would tell you to disable first.

@jeubanks I have now wired back in the Fakro window opener just in case.

@njanda I have now removed all user apps and devices with the exception of a FGS-223 driver for the dual switches as there's no other way of using them.

I've rebooted (after a shutdown) and will run another Z-Wave repair to see if it makes a difference.

I hope you just disabled the Drivers and Apps and not actually removed them cause it'd save you a bucket of work. :slight_smile:

You didn't need to wire it back just remove it from the hub. With it being powered off you would have to force remove it but either way it gets removed from the z-wave db and then you can run a z-wave heal to repair the routing tables on the nodes. You may want to run the z-wave heal a couple of times to ensure the new routes are created and updated on all of the nodes.

@bpa I would contact support@hubitat.com The total lack of messages or the duplicate messages may indicate you have a faulty devices that floods the Zwave network and brings it down. They may be able to help track it down if that is the case.


After 24+ hours without any lock up or erratic log behaviour I am cautiously optimistic the following changes may have resolved it:

  • Soft reset and backup restore (thanks @jtmpush18)
  • Reinstalling the Fakro window opener (thanks @jeubanks)
  • Removing user apps and only using Hubitat drivers (thanks @njanda)
  • Running a further Z-Wave Repair, shutdown and power cycle afterwards

I am still seeing infrequent Aeotec MultiSensor 6 events duplicate in the log, but it's only a single duplication per event.

I'll leave it for a few days to see if it remains stable. If so I'll slowly make configuration changes and update this thread if a cause is found.

Thanks to everyone for your help.