Hub Randomly Restarted Overnight - Can we determine why?

Platform Version
2.4.3.177
Hardware Version
C-8

I got a notification overnight that my hub had gone "Away" and gone back to "Night" at 2:19am. After investigation, this was due to it restarting (I have a rule configured to fix mode manager properly when the hub starts up) and I'd like to determine what triggered it. The power did not go out (everything is on a UPS anyway) so I don't believe this to be a power issue.

All I can see in the logs are "System startup with build: 2.4.3.177".

Are there any logs anywhere else that can determine what happened / what triggered this? I'm slightly worried that maybe my hub was accessed remotely, but I don't know why triggering a reboot would be all that someone would've done.

In your logs, click on "hub events" in the top bar. It might show the reason why

Unfortunately it doesn't. All I see is the "systemStart" event at 2:19am. The last Hub Event before that was 2/13.

Maybe your hubs memory dropped to a critical point causing your hub to initiate a self reboot. You can try the below link (Assuming it still works and the history wasn't erased due to the reboot) to check your hubs memory level history.

http://yourhubsip/hub/advanced/freeOSMemoryHistory

Just a thought

Appreciate that. Unfortunately looks like the reboot wiped the history. First line is dated 02-17 02:18:48.

I doubt it was hacked, but I would maybe install the Hub Information Driver v3 and keep an eye on things like free memory, in case it happens again. It could have been a one off, a power blip, who knows? I would only be concerned after the next occurrence of it happening.

Loading that driver up now. Thanks!

Gopher (Victor) recently patched a Java Free Mem (IIRC?) leak that could (again, IIRC) cause an unsolicited reboot... Since you're still on 2.4.3, perhaps that's what you're running into.

1 Like

The hub always wipes the memory history on reboot.

If it reboots to low memory, it makes an entry in the event log. That was actually shown as the reason mine rebooted about a week ago in this comment

Definitely wasn't low memory then. I have no entries there.

I don't think that the jvm free memory reboot creates an event in the log

1 Like

@gopher.ny Victor, is this correct? … if so, can an entry be added?

I’m pretty sure it was @danabw who had this happen in the beta. Perhaps he can chime in.

I had an "unexplained" reboot (a couple of them, actually) and Victor confirmed they were due to low Java memory report on the hub. Events looked like below:

"hub memory is critically low, rebooting now"

Victor's comment on this was:

He also said they were fixes on .126, and I haven't seen any of these events since that release:

This below looks like a normal restart message/event. I'd expect this maybe if there was a scheduled reboot setup on the hub...

2 Likes

I got two of these: Hub memory is critically low, rebooting now

On 125 & 126, none recently.

1 Like

There isn't. I have used an app to reboot hubs on a schedule in the past, but never on this hub and not on any in a year or more. This is the first time this hub has ever done that, that I know of. That's what made me nervous.

This is very unlikely to be the cause unless you have an unusual remote access strategy configured for your hub.

Is it possible the hub’s power blipped even momentarily last night?

Either the circuit it’s plugged into, or the hub’s USB power supply?

2 Likes

I don't. I do use Hubitat Remote access and the password is completely random (not re-used anywhere else on the internet).

I'd be surprised. Nothing else went offline overnight (I have 2 Hubitat hubs, many Wi-Fi cameras, a Firewall that tells me every little thing that goes on with my devices and ISP, etc.) and the C8 is plugged into a UPS battery backup (along with some of those other devices I mentioned).

1 Like

I had one of those recently on a C7, same build. Victor had a look at the engineering logs and they were clean so suspected a power glitch, but like you I have it hooked up to a UPS along with other things (rpi, router) that didn’t reboot. If it’s a one off probably not worth sweating it. Can always check / replace the power cable and/or power brick.

Same story here (.177). But my logs showed the critical memory error. I never heard back from the review of my engineering logs though. Firmware has been updated twice since then. So, if it was the java memory issue, it was fixed I hope.