Working Backwards - Troubleshooting Devices causing Hub Lockup

I've had some challenges, 2.2.4 was my last stable release. Every since then, every night my hub locks up and becomes unresponsive, including using :8081. This means I need to unplug my plug and plug it back in. I'll admit, i'm a little jealous of everyone loving their hub and 2.2.6 releases. I've rolled back, tried removing apps and reducing functionality on my hub to find the problem. I'm just about ready to exclude everything and start from scratch.....

Most recently, i did a soft reset, and restored from an old DB. I'm now running on 2.2.6.140, still same problems.

I guess my next question, is how do i find what is causing this issue? Can it be z-wave, zigbee network issues? Is it a device? What logs should i be looking at? Ghost devices? Other apps that I should use to figure this out? S0/S2 Issues? I have about 33 devices. I know, could be all of the above....

Current Devices:
Inovelli Switches/Dimmers (V1, Red)
Zen Switches/Dimmers (Zen26 V2, Zen27 V2)
SmartThings Leak Detectors (Qty 4)
Hue Motion Sensors (Qty 1)
IRIS Motion Sensors (V3) (Qty 3)
Sengled Element Bulbs (Qty 7) - Zigbee - Not on smart switches

Removed;
Zen25 (Plug)
ZSE40 (4-in-1 Sensor)

Any suggestions on troubleshooting paths, logs to look at, or confirmation to just start from scratch is appreciated.

Running custom code? Disabling that is first place to start. See if it improves the situation. If it does, then slowly, one at a time, re-enable, wait a while to see if the problem comes back. If things remain stable, enable the next one and repeat the process until it’s suddenly no longer stable.

However if you are also unable to reach the diagnostic tool, start looking at the connection between your hub and your router, as something is not right there. You should be able to reach the diagnostic tool regardless of what’s happening on the hub platform.

1 Like

I do have all of my IOT (including Hubitat hub) on a separate VLAN (unifi). I know that is mixed response here. I can look at my rules a bit more. Again, no real issues before 2.2.5.

Right, but there’s no reason for not being able to access the diagnostic tool on 8081 in any of the versions. It’s independent of the platform stability for good reason.

That’s a flag that you cannot reach it and temporarily removing the hub from the VLAN during troubleshooting would be my recommendation. No way to find problems like that easily without taking things back to a very basic connection scenario. If you connect the hub directly to a router for example and the UI unresponsiveness stops, you then know where to look next.

If the unresponsiveness continues, but the diagnostic tool is now accessible, then you have some evidence that there might be a platform version and VLAN setup conflict.

1 Like

How many of the Inovelli switches do you have and what firmware version do you have on them.

I know there was a bug in the original firmware where disabling power reporting actually caused it to flood the network. Think it was fixed in 1.16 (on the switches) and a different version on the dimmer. I had the issue originally and it was killing my network as I have about 30 of them.

Could be worth looking at as well.

1 Like

Here's an outside possibility... Do you have jumbo frames enabled on the switch?

Support helped me troubleshoot some lockup issues a while back, and the deep logs showed that the hub was getting frames of "invalid" sizes/sizes it did not expect or couldn't handle that would apparently cause the nic to die/lockup.

I turned off jumbo frames on the unifi switch they were connected to, and the hub that locked up periodically has never locked up again.

Was weird, as it only happened on 1 of the 2 hubs on that switch. But literally the only thing I changed when the lockups stopped was turning off jumbo frames.

They are off.

Set the hubitat IP to my core vlan tonight, we will see what happens.

1 Like

I have (1) first gen nzw30, and (2) lzw31 reds. I'll take a look at the fw on them, thx.

Locked up again tonight, no :8081 access. Moving it to core network vlan didn't help.

Since I'm only dealing with switches sensors, I'm okay without having security. What is the recommended security for my mix? I thought S0 created more logging? Are there any adjustments I can make to device security through the hub, and not pulling out a zwave stick/software?

You will hear different opinions in the forum. A few believe that you should have your home network locked down like Fort Knox so that someone hiding in the bushes can't see that you turned on the bathroom light. Most, though, believe, as I do, that security should only be on the perimeter (border entry) devices like locks, garage door, etc.

No, it causes 3X the packets of S2 security. Again, unless it's a border device, "None" is probably the best choice. S0 is an older protocol, not the preferred choice now if your device has other choices.

Security is chosen at pairing (inclusion) time. If you want to change it, you have to exclude the device (which can break automations and dashboards unless you swap in a placeholder virtual device for every "in use by" instance on the device's page), reset the device to factory defaults, then include again, unchecking all security box choices during the inclusion process. Then, after inclusion, swap back for the placeholder virtual device.

1 Like

This is fine. Anything to get out of this daily morning reset my router routine.

I had a ZEN switch and Inovelli Red Dimmer become unresponsive, and flooding and taking some Ring First Alert smoke/co link to the ring alarm hub offline (still worked as standalone). I reset those with a circuit breaker, just excluded/included again. I thought my options in re-inclusion were S2 Class 2, S2 Class 1, S2 Class 0, S0. I didn't see any with 'none'.

Looking at the z-wave details in settings, my current security on all my devices is 'none'. Since i did this yesterday, and had to power cycle my hub this morning, this wasn't the cause of my issues.

None is obtained by unchecking all choices.

2 Likes

Rolled back to 2.2.4. Also, all zwave security is now set to none.

I think i'm down to excluding all my devices, updating FW, then re-adding/including devices back in slowly. I have a couple of zwave repeaters on order, maybe that can also help.

Did you end up excluding and re-adding your devices? I am having the same issues except that I can reset it by logging into the diagnostic tool.

Honestly, they figured out i had a bad hub. I received a replacement, none of the same issues. I would open up a ticket for them to research a bit.

1 Like