High CPU usage

[colinhowlin]

Correct, CPU usage dropped once SSH was disabled.
Would correlate with the multiple dropbear processes running also.
Like I said, I'll need to wait until later to reboot but will report change in CPU once I do.
Will try to do some further testing at that stage.

[colinhowlin]

So only getting back to this now.
Two of the three switches that were displaying high CPU usage are once again showing high usage a couple of days after rebooting to re-enable the SSH server.
I've attached output of /var/log/messages, as well as ps and top output.

I've disabled SSH server on one of the switches which didn't affect CPU usage at all - still in the 90's.
I've since disabled SNMP and all discovery protocols and still getting 90+ CPU usage.

I'm also seeing random spikes in traffic - see attached screenshot as well.

[colinhowlin]

Spikes in traffic might not be accurate.
See attached screenshot with one of our Cambium PMP450i ports highligted showing traffic in excess of 180Mbps.
Actual AP traffic is closer to 70Mbps at the moment.

Backhaul from this site is a Ceragon IP20C and Netonix is saying that is pushing over 600Mbps at the minute.
However, the radio and the other end of the link are showing a more believable figure of about 200Mbps.

[Eric Stern]

Try disabling the manager and see what happens.

[colinhowlin]

Hi Eric

I've uninstalled Netonix Manager - plan to install v 1.0.1 later today.
No change in CPU usage on the affected switch.

With SNMP, SSH and Discovery all disabled, I'm still seeing 90%+ CPU on the affected switch.
The other two are around 50%.
One of the other two has SSH server enabled so I'm starting to doubt that Dropbear is the problem.

Just rebooted the affected switch and CPPU reading normal again.
Perhaps one of the services was not fully disabled without reboot.
Having to reboot to apply changes is slowing my troubleshooting.
Will re-enable SSH server on affected switch and reboot to rule out Dropbear.