How to find out why hub went offline

My hub went offline around 11pm. I didn’t notice because I generally find it to be kind of flaky, and just assumed some sensors weren’t being responsive. When I tried to use a pico remote at 1:30 nothing happened I went to the hub events logs and they stop around 11 and restart at 1:40 when I unplugged and replugged in the hub

How do I diagnose why it went offline? And why doesn’t the iOS app have some functionality to tell you when it does?

Was it just plugged into an outlet, or do you have it plugged into a battery backup system (UPS).

Do you remember if the power light was green or red, before you unplugged it and plugged it back in?

It’s plugged into a regular outlet. I don’t remember what color the led was when I unplugged it, green or blue perhaps, in think I would have noticed something like a red/orange/yellow.

Did you check the past logs?

1 Like

This should not be happening. I suggest disabling ALL apps and devices with custom drivers.

Then enable them one at a time until you identify ones that reduce stability. The only times either of my HEs is rebooted is when there’s a platform update.

Edit - I should add that in the past I’ve brought one of my hubs to its knees by:

  1. a badly written driver (I wrote it)
  2. leaving stranded zwave devices (my fault for force removing them without exclusion).
  3. someone else’s port of a ST app

Edit: In the event you weren't sure, here is how you disable devices and apps while identifying issues that lead to instability (you do not have to uninstall them).

2 Likes

Also on the hardware side make note of any new devices you may have added/removed or modified recently. Other have reported times when a bad device has been able to lock up the hub.

And before its asked, no there are no tools to diagnose runaway processes on the hub.

2 Likes

Red, green, and blue are the only possible choices for C-5 hubs. The next time it happens, check if you can reach the hub at port 80. If that doesn't work, try port 8081.

3 Likes

I was about to be surprised that no one told you not to pull the plug on your hub. Thankfully @aaiyar mentioned it somewhat, but I'll don my HE bad cop hat for a second.

You should not pull the power cord until you have tried ALL other methods of rebooting your hub. Doing so could lead to corruption and much bigger head aches.

When having trouble, please try accessing the hub at

{you hub's IP address}:8081

If anyone reading this has never seen the screen below, I encourage you to discover it (there is a reason we do not wait till the building is burning to have a fire drill!). :fire_extinguisher: :fire: :fire_engine:

8 Likes

Some of the flakiness is with core system functions like motion lighting, modes and scenes, so its hard to blame 3rd party code, since these issues have been happening since day one before any 3P apps or drivers were installed.

Only things that has been added in the last 48 hours is an IKEA tradfri light (added via Hue) and a new rule for that light from an existing motion sensor.

yes, as I said, there was nothing suspicious looking in the logs before it went offline.

You should describe exactly what you mean by flakiness, which is a little too generic to debug.. And please include logs.

Don't know what happened there without accompanying logs. My suspicion is that the browser back button was accidentally pressed somewhere along the way so the scene device wasn't created.

@bertabcd1234 addressed this. Don't use "0" for a dimmer level if you aren't using it for a particular mode.

This seems to have been two issues. One which was a conflict between a built-in app and a community app. The last post in that thread (log screenshot) do not appear to reflect a built-in app.

Fixed in the latest iOS beta as described yesterday by @bobbyD.

1 Like

Sorry for the troubles. Please send us an email at support@hubitat.com with your hub UID and we will further research. If these problems happened since day one, you may have a fauly hub. We would like to further research to make sure that you are not dealing with a hardware malfunction. Regarding your open tickets, sorry for delays on those, we do not have a public bug tracker.

4 Likes