Hub goes into nutso mode every 4 days or so

I was having tghe hub reboot automatically on zigbee shutdown.
This becuase I would get zigbee shutdown messages and stuff would stop working. It has been happening about a month or more now.. The house is empty and no significant changes other than f/w upgrades when this started happening.

It recently started happening every 4 days or so.

Recently i got on the hub to look at something and i could see it was about to happen.

hub was at 136 degrees, cpu at 45% based on the hub stats app.

However, looking in logs, zwave logs, and zigbee logs, as well as device and app resource status pages I could not see anything abnormal or what the heck was chewing up so much cpu time.

Any ideas how to isolate this and track the issue down.

On reboot temp and cpu usage drops immediately (or withing 10 minutes after restart when everything is loaded). temp back to 98 degrees cpu back to around 12-15 %

Have you looked at the detailed run-time stats? The details tell you which apps and which device drivers are consuming your system resources. A usage of 45% seems quite high. My system is only 2% busy.

In my system Echo Speaks and Actiontiles integration consume the largest portion of the resources. Both are Cloud based integrations, so the usage is not surprising. Most of the other apps and devices put a minimal load on the resources.

Once you find out which device or app is draining your resources, then you can take steps to correct it. Perhaps you have a device that is checking in a lot more frequently than needed. You also might have an app that is putting the hub into a processor loop. The stats should tell you if that is happening.

i already said i looked at that if you read above.. in addition, no echo speaks.. also the highest in mine are cloud based and again no suprises there.. .. also it works find for days and the only thing that has changed since i left home has been basically f/w updates on the hub. it is not an indiv. app or device as the stats are no higher than normal.. and i don't see any individual one causing the issue.. the percentages are basically the same as when the hub is running well .

Which one of the hubs?
Also, try 2.2.8, it has non-trivial performance improvements over 2.2.7... if nothing else, the hub shouldn't run as hot with 2.2.8 on it.

1 Like

the nh hub .. ok i will upgrade and try it.. was waiting to make sure things stabalize. as i get nervous upgrading a hub where i am 1400 miles away if something goes wrong..

i suppose a device could have failed.. and cause the issue. as i have had a zwave device that everythign was routing through fail and cause issues.. but not high cpu usage .... just many devices not working that were still trying to route through it.. Also dont think a reboot would fix it for days in this case .

Ok, having no physical access to the hub is a reasonable concern.
I'd say schedule daily reboots on it in the meantime, and update when convenient, then. I'll check the hub in the meantime, too, maybe something will jump out.

thanks.. i will update the local hub first as i have many similar devices and if no issues will update the remote hub.

Since you didn’t mention it, it was probably normal, but just in case -> how was the DB size?

normal for me.. it starts at high and levels out aroun d 60-70 meg. when running here is the graph. sorry thats not db size i should add that line to the graph.. but it was aound 70 at the time i rebooted it for high cpu usage.. that is not abnormal for me.. the hub was running fine with a db closer to 100 before i pruned/reduced many of the event and attribute sizes.

1 Like