Every single z-wave device fails repairs

So, I was having a weird issue when I looked at my z-wave radio where almost all of my devices were running through a node that was pretty far from the base and some of the routes seemed really sub-optimal. So I thought a repair might fix things. When I ran the repair every single node came back as failed and timed out regardless of the group or route it was assigned. If I reboot the hubitat things seem to work ok for a while but eventually I have issues again and every time I run a repair every device fails. I even excluded and removed the first hop device from above thinking it might have been a problem but now I am really screwed because it still can't repair and it doesn't have legacy routes to fall back on for a lot of the devices. Any ideas? I have rebooted the thing, power cycled it manually, tried various config options and can't seem to get anywhere with it. I would rather not have to re-set up everything manually as I have about 40 devices and I don't even know that that will fix things.

Post screen shots of your Z-Wave Details page.

Before you did the repair were things working OK? It's always best to ignore the Z-Wave Details page/routing unless you are actually experiencing a problem w/lagging or failing automations/devices.

What were the issues (if any) before you ran the repairs, and what are the issues you are experiencing now?

1 Like

Basically I was having issues with devices not responding consistently. The issues that I am having now is that devices are not responding at all. Some of the routes that were assigned seemed bizarre, It was going across the house and back to hit devices that were a few feet from the hub and things like that. Now when I do a repair every single node times out. I am sure that all of my devices didn't just go bad at once. don't really know where to start though, so any tips to get me in the right direction would be great, I don't mind doing the legwork on this.

@danabw gave you the first step:

1 Like

It won't let me embed an image in my post, and it also won't let me put in a link? Maybe because I am new? so uh imgurdotcom/tVHY8i6 . Anyway, you can see in the link that there are a bunch of devices showing no route, The top device shows a route that goes through a node that doesn't exist anymore (12)

Yep, that will clear up soon, just keep spending time in the forums, reading & posting, & the forum software will bump your trust level. It’s to discourage trolls and spammers.

Here’s your screenshot so others can see. No ghost nodes, which, I believe, is what @danabw wanted to see:

I’d suggest trying a single node repair (not full Z-Wave repair) on that first device. Just hit Repair. Have log window open in another tab to watch.

Any reason you have those two lights paired S2 rather than no security? The C-7 still has some teething issues with some devices and security.

2 Likes

Exactly...

@permettehome - @672southmain's advice about doing single node repairs is good, there are few cases where you'd want/need to do a full mesh repair.

Also, the screen shot only shows 10 z-wave devices. You said you had about 40. Are the rest all Zigbee, or are we missing some of the Z-Wave details page?

If the single node repair doesn't do anything, then:

  1. Use the Settings>Shutdown Hub option to turn off your hub
  2. Pull power on the hub (at the wall, not at the hub), wait 30s, then plug in again
  3. After boot up see if the Z-Wave details page looks any better
2 Likes

I don’t know about you, but it takes at least 24 hours for all the route to fill in for me. Helps if I go around and cause devices to do something.

2 Likes

DOH...right. :slight_smile: Very good point (that I missed :slight_smile: ), the best thing to do may be to wait a day or two and let the mesh sort itself out. Especially if these are Z-Wave+ devices, which they appear to be - 0x5E.

2 Likes

ok, I ran a repair on the first node and it just failed out. I opened up the logs in a different window and I am not getting anything. is there a setting to specifically enable or disable logging? The reason those 2 lights have the security setting is just that I had set them up recently and they asked for that setting when I turned them on, I have a few others that I didn't give that setting and I was going to try and figure out if there was a reason they needed it or the system requested it. I guess I am dumb I did not realize the devices would "sort themselves out" over time, I thought that It only happened when you manually ran a repair. I guess I am confused as to the purpose of manually running a repair.

It’s to get the routes fixed. The purpose of having logs open in another tab is so you know when the repair is running. If you hit Repair while a repair is running (double click, or impatient), you get Z-Wave busy messages. That’s all. You might try exclude/include on that first node.

And, like @danabw pointed out, we haven’t seen the rest of your nodes. There may be ghost devices.

Before doing exclude/include on that top node, check the bottom of its device page to see if it’s “in use by” any automations. Need to remove from the automations first or they will break. A trick is to make a virtual device of same type, swap that device into the rule/automation for the one you are going to exclude. Repeat swaps until there are no more “in use by”. Then, after exclude/include, swap the device back in for the virtual device. Wouldn’t hurt to factory reset the device while it is excluded, before re-including.

If your device is Z-Wave Plus (0x5E as one of the in clusters), it will repair/re-route on its own, eventually; Node repair just hastens things. Z-Wave non-plus don’t do this, you need to repair them manually.

The fact that the top node failed a node repair indicates it’s got issues.

Ok thank you for all of the help I will try the suggestions and also let it settle for a day or so and report back.

2 Likes

If you have any Z-Wave devices that don't have anything in the "Clusters" column on the Z-Wave Details page that's a clear sign the device is not joined to your hub correctly.

image

If you have any like that after you do the reboot, then you'll need to remove it.

If you have not already done so, it is sometimes helpful to do a hub shutdown, remove power, wait a minute or two, and then power back up. Removing power will give the z-wave radio time to restart, which may not happen with a regular reboot cycle.

2 Likes

Disconnect power at the wall, not at the micro USB connector at the hub. It’s fragile and can break the solder joints to the PC board.

Every single time my Z-wave network has acted even a little funny, or slowed down. A simple hub safe power down, unplug for 30 seconds and power back up has cleared it up 100%.

It's like the age old PC joke when you call tech support "have you tried rebooting it yet?". :stuck_out_tongue:

2 Likes

I know this is old, but I was having similar issues with my z-wave mesh.

I have built my z-wave network up slowly over time, adding devices intermittently over time. Things with my Hubitat Elevation C-7 hub were working snappy for a while, then over about a 6 week period during which I added a few devices and physically moved a couple others, my z-wave network became super slow. Logging analysis showed me that my rules were not the bottleneck, messages were being sent out on the z-wave mesh network and sometimes taking 30 seconds to a minute to change the state of the endpoint switch!

I tried to perform "z-wave repair" in the Hub's settings and was getting over 50% failed nodes. I tried repairing nodes individually with the buttons in the browser and this would fix some nodes but make others fail that hadn't failed previously. Rebooting the hub did not help.

Finally, I went out into my garage and turned off every breaker in the house and then flipped them back on (forcing all except the battery powered z-wave devices to reset and reacquire connection to the mesh as I have a lot of in-wall switches, dimmers, fan-controllers, etc.). This also obviously restarted the Elevation Hub as well.

I then logged back in to the Hub on my browser and ran SETTINGS | Z-WAVE DETAIL | REPAIR Z-WAVE and the repair executed faster than ever and all the devices have valid routes, none failed, and all the devices are quick and responsive.

I wasn't expecting this idea of resetting all of the devices at once and restarting them at once to work as well as it did, it's probably not reproducible, but my z-wave mesh network was utterly broken and now it's fantastic. I didn't have to exlude/remove and include/add a single device which would have been a pain because it would have broken a lot of rules. It seems a global power-outage reset of the house works, devices all just re-established connections and routes and the mesh network "fixed itself!" Thought I'd share.

Sean