My C5 hub keeps locking up randomly and it's getting very annoying.
Sometimes it works months or weeks or even days between lockup. It shows a green light when in this non-responsive state and requires me to pull the power to reset it.
I resorted to trying a rule to soft reset the hub every 3 days, but that did not solve the problem either.
I only have 32 devices on the hub. I cannot see anything in logs, but maybe I don't know where to look.
Here you can see it locked up during the night after 2am. I noticed after 4:30am and cycled the power.
Do you mean a reboot or a "Soft Reset"? They're two different things. A reboot can be done via a Rule. A "Soft Reset" cannot.
I would suggest you try a real "Soft Reset" via the Diagnostic Tool, and then restore the hub's database. That may resolve the issue. Instructions are here:
Also you might want to monitor your memory / resources. Consider installing the HubInfo driver..
Other things to watch out for are overheating (make sure your hub is not on a smooth plastic surface for example) and electrical issues. A lot of use have our systems on power backups.
Also look at the stats for your custom/community apps and drivers - anything that pings the cloud/network etc can have a big potential impact on resources.
Good point. Power supplies have been known to go bad. As have USB cables. If the issues persist after your Soft Reset, I suggest replacing both the power supply and the microUSB cable.
That's where the HubInfo driver can help monitor things like a spike in activity, temp or memory reduction. You can set up watchdog rules to reboot if you get too much of a mem drop for example.
dBSize of 3M is very small. Happy to discuss numbers if you have questions. Free memory normally starts out after a reboot in the 450+ range and will drop slowly - I used to have an alert set at 250K, but have recently lowered it to 200K based on recent improvements. Temperature I have a warning at 140°F (60°C) for my C7 (C5 and C4s run a little warmer than the C7) - normally only hits it at reboot. CpuPCT or CpuLoad has dramatically dropped with recent releases but varies depending on your mix of apps and drivers, so look for consistency in your numbers.
Continuing discussion from yesterday . I think your Denon driver is still causing your hub problems. I suggest temporarily removing the integration to see if that resolves the frequent lockups.
That number will grow as events happen on your hub. The size it stabilizes as I think mostly depends on the event and state history size limits you set:
Mine start out around 7 and stabilize around 23, but I have over 200 devices on each hub.
Free memory starts around 450,000 like @thebearmay said, and usually stabilizes around 300,000.
Does event and state history size setting have that much affect on stability?
Mine are still at default settings. Should I adjust? I see you set your's much lower.
For very busy things, or things where I don't care if there is a long history, I set both events and states to 1 or 2. Weather would be a great example of this, there are dozens upon dozens of items like current temperature, min temp, max temp, rainfall, cloudy/sunny, and so on. That can fill the database in a hurry if you track of too many events. And besides, who needs to go back a month or even a week for weather? You need today and tomorrow and possibly the next day.
Same for busy things like multi-sensors, do you really need to look back into history for weeks if not months? Probably not. A few of each event would be fine.
For most everything else, maybe a simple light switch, the defaults are good in my opinion. And I do want a history of locks and garage doors, but they aren't busy devices.
