Can Enabling CPU C-States cause Disk Errors?

[phoenixdiigital]

With a goal to reduce my UnRAID servers power usage I started to investigate enabling CStates in the BIOS (see attached).

 

In the following few weeks I started getting  Disk Errors on different disks(Disks 6 & 7). I did some extended SMART tests and the drives reported back fine. I should also note these disks have been in the array for over a year and the errors only started cropping up after enabling CStates.

 

In one instance the disk went into emulated mode and I recovered from parity to the same disk. It's been fine for about a month.

 

I'm fine to leave Cstates enabled even with these occasional non critical disk errors so I'm not really asking for help. This post is more just a question of why would enabled CStates be causing these errors?

 

Some reading on C-States for anyone interested

 

CStates Explanation - https://metebalci.com/blog/a-minimum-complete-tutorial-of-cpu-power-management-c-states-and-p-states/
More CState Information - https://www.thomas-krenn.com/en/wiki/Processor_P-states_and_C-states
 

Just before hitting submit I dawned on me. I wonder if the errors are caused by the "SAS2308 PCI-Express Fusion-MPT SAS-2" and the server hitting a state where it tries to send the PCI card to sleep? Pretty sure I avoided enabling PCI power saving features of the BIOS out of concern of this.

 

I've attached a diag too.

 

tower-diagnostics-20240621-0813.zip

[JorgeB]

9 hours ago, phoenixdiigital said:

I started to investigate enabling CStates in the BIOS

This is usually not a problem, except for some early Ryzen CPUs, most boards come with this enabled by default.

 

Your issue appears to be this:

 

 

[phoenixdiigital]

29 minutes ago, JorgeB said:

Your issue appears to be this:

 

 

Very interesting. Thanks for the heads up.

Apologies to C-States it seems for accusing them of messing things up. 

 

Though these disks have been in there for quite some time. Maybe I was just lucky.... Edit: Actually I probably did upgrade UnRAID during that time so possibly the issue aligns with that.

Edited June 21 by phoenixdiigital