Daily Z-wave crashes

I use the zooz 700 stick myself

2 Likes

Yep, sorry about that 46 and 53. I had the same issues. Those probably didn't include correctly when you added them. Even though you "woke them up", that will likely be temporary and you will have the problem again. I would remove them from your mesh completely then reinclude non-secure. Getting rid of the S0 on other services will also help.

As a follow up, i started having issues,and low and behold, I had 2 bulbs paired as S0 that were causing issues (not sure why it took so long to be an issue). Removed and re-included as non secure, issues began clearing up within a few hours.

3 Likes

I had an issue with the same thing except it would wipe out all my sensors with my old ATT Digital Life system. Turned out there was a neighbor broadcasting with some ham radio on a close frequency that was not authorized by the FCC. The only way I found out is they had a tech come out with a spectrum analyzer at the time it happened and they went and located the signal. The guy that was doing it needed some filter on his equipment that wasn't in place but they worked with the FCC to fix it. I was going to suggest looking at the RF environment but it seems you figured it out.

Still on the default S0 inclusion of 500 devices, it's quite weird that all my Qubinos Flush Relays and Flush Shutters (all 500 series, Zwave Plus) have included as non-secure. In fact, their parameter 250 Unsecure/Secure Inclusion is by default 0, meaning Unsecure Inclusion so it seems there is a way of circumventing that requirement, i.e, that all Plus devices must be included as S0 minimum.

Those devices will include at S0 unless you use the workaround and include them using the Z stick as a secondary controller - that’s the only way. All my Fibaro Dimmer 2s had to be included with the stick, I wouldn’t waste any time trying to reason why some devices manage to include without security and some pick S0. Ultimately if you want to reliably include ZW500 devices with no security you need to use the workaround.

2 Likes

Not wasting time, just wondering why that is with some devices and others not which just proves that it's possible with ZW500. Anyway, I'll try the workaround, if it doesn't work for some reason, I will wait for ZW700 switches to be released and replace all Fibaros.

The Philio PAN3x appear to be reaching the market although I haven't seen a single shop to carry any yet.

1 Like

The workaround definitely works - I excluded/re-included all my Fibaro Dimmer 2s and Switch 2s with an Aeotec stick. I have had the odd Fibaro one that included directly without security for reasons unknown but I couldn't replicate it however hard I tried - I tried doing the inclusion both at the HE and the installed location. Using the Z stick now has become very straight forward and I don't need to refer to the instructions anymore, it's very easy.

1 Like

Just to share my experience, I got a Aeotec Z-Stick Gen5 to get rid of my S0 deices and it was a bag of hurt. First time I included the stick it was itself a ghost quite difficult to get rid of. Once I sorted that, I proceeded to exclude/include devices and it was horrible as they wouldn't go from one controller to other and when they did, they went as ghosts. This occasionally worked if I had the stick close to the device and to the C-7. Terrible. In some cases, if after semi-inclusion with the Aeotec the device passed as a ghost to the C-7, a second inclusion in the C-7 worked but not always.

I decided to give up and get a UZB-7 to try it out. Guess what, it works perfectly, wherever I am in the house with the stick, excluding/including devices. Am returning the Aeotec and keeping the UZB-7. In the next few days I'll clean up my mesh from the S0 devices.

4 Likes

And my daily crashes continue. I have just one S0 device left, which is operating my gate strike but I will also get rid of that once it stops raining.

This is quite annoying and have zero clue in the logs of why this may be happening. I have zero ghosts. I have also bought 4 Ikea Zigbee outlets and they work fine, they are not affected by the Zwave crashes.

Initially I found that only a full shutdown would bring my Zwave back to life but now I found that triggering a backup, either to the cloud or locally will also do the same. A reboot does not.

What does a backup do to the Zwave mesh that a reboot does not?

Really all it does is clean the database (You did a backup, soft reset and restore from back up you saved right?) Something is overwhelming your mesh. Have you turned off all power reporting? Maybe turn on debug logging (not just description text logging) on a few devices at a time and let that run a day or so to see if too much is coming through. @gopher.ny anything you can see?

No, just backup, no soft reset. Zwave is unresponsive and to get it back, I have two options: click backup (and a file is downloaded) or I do do a full shutdown and power the C-7 back up.

What does cleaning the database entails exactly? Any memory buffers get flushed?

Can you PM me your hub's id? I can take a look at the engineering logs on the hub.

1 Like

@laalves
Ok this may help you then. You may have a corrupt database. Go to settings>>backup and restore. Click the download button at the bottom and save to PC. This will clean any corruption in your database. go to yourhubip:8081 and do a soft reset. On reboot restore the database you saved on your PC (not the built in ones) This will eliminate corruption as a culprit.

2 Likes

Done, you have PM

1 Like

Thx for the suggestion, just did that, a bit hair raising I confess, but after the restore all is back.

Let's see now whether the Zwave network stays reliable, will keep you posted.

3 Likes

Hopefully @gopher.ny sees something too

1 Like

Hello, I know it's been a while since I posted. Rather than try to figure out the root cause, I just pushed the eject button and reset the Z-Wave network. After about 2 weeks, I can say that the Z-Wave network is finally working properly. This was a journey though and I appreciate all of the knowledge I got from these forums!

This may be helpful for others so I'll post what I did:

  1. Reset the Z-Wave network in Hubitat.

  2. I had two repeaters (Aeotec), so I first plugged them in, excluded them (using the Aeotec USB stick, this is CRITICAL), then added them to the new network.

  3. When adding anything, it is really important to know what's happening in the logs before you do anything. With just two repeaters, I was seeing no Z-Wave events and very little activity (which is good). If you know what was going on before you add anything, you will be able to see if something you added may be causing issues.

  4. I then went and repeated the same process for my light switches (since these are Z-Wave repeaters themselves). Once again, the Aeotec USB stick is critical. Trying to exclude these using the Hubitat almost never worked; having the stick next to the unit when I used it to send an exclude signal worked 100% of the time. If you don't have one of these sticks, you will not be able to exclude devices properly (unless maybe your Z-Wave network doesn't exceed a 10-15' radius from the Hubitat). Also, in the past I added the stick to my Z-Wave network but this time I did not, I just use it to send out an exclude signal.

  5. I kept watching the log. It seems the Z-Wave network cannot handle significant event load; when I was having trouble, I was seeing 5-10 events a second. I think that it was just too much for it to handle perhaps. With 30 Z-Wave devices on now, I'm seeing 5-10 events per 5 minutes.

  6. When adding devices, it is CRITICAL that you are patient. I am not. If something doesn't add properly, do NOT move on. You must get the unit added before you try to add another. If you run into something that just won't work, make sure you remove any ghost Z-Wave nodes ASAP. It may take 5 tries to get it to finally delete. It'll go a lot faster if you can kill power to the node itself. Also, I found that a hard restart (pull power from HE) many times will get the Hubitat to "refresh" its view of the Z-Wave network. I have experienced times when the Hubitat will refuse to add any new nodes; a hard restart and magically it adds the node on the first try. I'd also add that it's not a great idea to try to add one a node, walk over within 10 seconds to add another, etc. I waited at least 3-5 minutes between nodes.

  7. If you have had nodes in your network which are now gone for whatever reason (and are removed from Hubitat), when you add a completely different new node, at times Hubitat will add the new node but will use the same driver NAME as the old node (but will use the correct driver). I experienced this many times and it's very confusing; you're trying to add a light switch and Hubitat will tell you it's adding a shade. It's just the name of the node that is wrong, Hubitat does use the correct driver.

  8. I have not ran a Z-Wave Repair and I will not. I'm not sure what the ramifications are of doing a repair truly are but it never resolved any issues in my experience. I think it's very easy to send out overlapping commands from the Hubitat into the Z-Wave network that the nodes in the network have not yet fully "processed".

  9. Last note: When setting things up, especially when doing excludes, make sure no one else in your house presses anything. I had some light switches get inadvertently excluded because someone in my house tried to turn some lights on in a different room while I was trying to exclude something. Doh.

After following this process, I was able to get every hard-wired node in my house added properly, no dropped events yet, logs showing very light load, reasonable speeds, etc. The final learning I have from this experience is (all IMHO):

Z-Wave is a very fragile technology. It is not well designed to handle some real-world use cases (i.e. orphaned nodes, overwhelming traffic, etc.). It places an emphasis on creating a wide mesh versus a fast mesh (i.e. it creates routes between nodes that make no sense when viewed from a "fastest route possible" perspective). It is a relatively slow network (i.e. a command can take .5 seconds to make it between nodes which is a long time IMHO) and it is not designed to accommodate a high volume of events. Hubitat does its best to interface with the Z-Wave network but at times it does not reflect the actual state of the network. Hard resets are an inconvenient reality when trying to add lots of nodes. And finally, Hubitat is not a reliable method to exclude nodes at all but excluding nodes is an essential way to add (or reset) nodes (i.e. a USB stick is critical).

Thanks!

3 Likes