Hub frequently rebooting

General Discussion
1

Since 5-6 days the hub frequently reboots. When it’s not rebooting it works fine. I obviously think it’s related with the changes that are being implemented, it’s a weird behavior though.

Any explanation?

1 Like
2

I’m experiencing the same thing. What firmware version is your hub on?

Tagging @Cory_Heuschkel

1 Like
3

the fw is 45.00011

4

Thanks for the reply. I’m on the firmware beta, and on 46.5

I was not having hub reboots like this with the previous beta releases. Hopefully the person I tagged above will be able to help.

1 Like
5

Thank you for reporting @sulisenator and including me @johnconstantelo. Matteo - would you please describe your interpreation of hub reboots? Do you notice this in Graph, or just from your hub falling offline? Can you provide any information of LED behavior?

Thanks,
Cory Heuschkel

6

Sure, it goes offline (app notification), blue led, no automation execution (even the local ones) then after 5-8 minutes turns back online (app notification), green light, and stuff is back to normal. Rebooting it’s my guess, obviously I’m not sure it’s actually rebooting. But if it just went offline it should execute local automation, no?

It happened around 12-15 times up to yesterday, now it’s 24h that it doesn’t happen.

Thanks for the support.

7

It sounds like you’ve overloaded your hub. What you’ve described is what happens when the hub is out of memory and starts freezing up. Is it a v3? How many drivers do you have running?

8

Mine is a v2. I didn’t have this issue until the latest beta firmware update.

9

I have a v3, and the 46.x firmware seems to have lowered the amount of memory available to drivers. I had to drop 3 more drivers to get back to stable. Now only able to run 11…

I haven’t seen anyone with memory issues on a v2. Try running logcat wide open for a while to see if the hub goes into periods where it restarts drivers once per minute. That and the memory warning when adding a device by brand seem like the earliest signs of full memory.

10

There are a couple of API entries that might be of interest:

    "driverCountLimitStatus": "Ok",
    "driverMemoryLimitStatus": "Ok"

I’ve not seen them say anything other than Ok on my lightly loaded hubs so I’ve no idea what else they can say. There must be some reason they exist though.They are on the /installed-hubs/{{hubId}} endpoint.

11

Good idea. Ill try that when i get home tonight. Im only using 4 custom drivers, and all the rest are ST’s.

12

You might be right, it’s a V3 and I was just over 50 drivers, I uninstalled some and now I’m below 50, what’s an acceptable number of drivers to avoid issues? Is the number of devices relevant as well?

13

I think it’s the number of running drivers (at least one device assigned to the driver) that matters most, though I’m sure number of devices has an impact too. I would aim for under 20 drivers, but even that may be too many. I have 125 devices and am maxing out at 11 drivers on the beta firmware. Any more and my hub begins restarting a different driver every minute.

@orangebucket Here’s what they say when at the point where the nearly-full memory warning pops up.

"driverCountLimitStatus": "Ok",
"driverMemoryLimitStatus": "SoftLimit",
1 Like