Hi @JRenaud 

Thanks for posting in our business forum.

Please share a topology of your network. It seems to be a problem with your whole network setup instead of the controller on any platform.

CPU increases to 100% and it sounds like a problem where you have looped your whole network.

You can also Wireshark from the controller and see if their packet comes to the controller computer.

Hi @JRenaud 

Thanks for posting in our business forum.

JRenaud wrote

Hi @Clive_A,

 

The software controller is run in a data centre in a VM, and itself isn't part of any network it manages. It's primarily used to manage APs and some switches in remote locations.

The CPU usage is of the host OS that hits 100% usage after a few days, but a restart of the controller fixes that.

 

The managed devices don't appear to have any issues functioning, but new ones throw “General error” when trying to adopt.

When adding an AP, this is what lines up when the adopt button is pressed:

 

INFO [https-jsse-nio-8043-exec-8] [] c.t.s.o.m.d.d.m.a.a(): MSP: Assigning device: DeviceMac(xxxxxx) to customer: OmadacId(xxxxxx)
Invoke API method a failed.
java.lang.reflect.InvocationTargetException: null

Followed by 100+ lines like below:

 

along with thousands of ones like below, about 30 per second but always happening, not just when trying to adopt:

 

 

Sorry for the pictures, it doesn't seem to like the formatting/links it sees when pasting the text.

So, a data center, I believe is not as powerful as a local.

Does the VM meet this what this FAQ writes?

And if it is in a data center, have you ruled out that there is the possibility of your instance being under attack?

BTW, remove the controller ID and MAC address as they are private information.

I think this relates to the CPU problem you have or your environment. Try a local controller and see if you experience the same thing. I don't think it is a problem with the controller so far.