I moved my unraid to completely new hardware (or rather old because I downgraded to a larger box but with an older CPU) and since then I have an error:

 

CPU possibly will not throttle down frequency at idle    

 

in the fix common problems module.  Additionally, when I go to the dashboard it keeps showing spikes in the CPU load but when I check htop through ssh things are totally fine.

 

Is there a way to recalibrate the dashboard so it displays cpu load correctly?  I started with a i5 6600k and downgraded to an i5 760.

You can see if the processor clock is varying with load by running

watch "cat /proc/cpuinfo | grep MHz"

at the command prompt (cancel it with CTRL-C). Intel SpeedStep is implemented on the i5 760 but, as you've changed your hardware, you might not have a speed governor selected. A simple way to do that is to install the Tips and Tweaks plugin and choose a speed governor, such as on-demand, there.

I installed Tips and Tweaks plugin and set the governor at power save and the cores are varying their frequency for sure (both before and after the governor) but I still have that error in fix common problems (after a rescan) and the load shows in the dashboard as if it is almost constantly at 100% when during ssh into unraid htop and cpuinfo both don't indicate this.

Edited August 26, 20186 yr by demonachizer

Well, the message from FCP includes the word "possibly" so maybe it's a false indication. Since you can see the clocks scaling, this is quite likely. Perhaps the Dashboard and FCP use the same method, while htop uses cpuinfo. Maybe there's something about those old Lynnfield processors that isn't quite right, or that a BIOS update might fix. There was an issue with similar vintage Xeons, if I recall.