Machine check error scary warning

[donatobox]

Hi all, thank you for your help!

 

My unraid build hasn't been the most stable, since building it last week, it has gone unresponsive twice. I'm experimenting with turning off "Aggressive Link Power Management" which might help.

 

But, I've been receiving these  "Machine Check Error" warnings, which may be the true source of the flakes. I first noticed the warning yesterday, and then a new one popped up over night during a parity build.

 

(in case the mcelog data isn't included, I copied output here - https://pastebin.com/V4UzpeL3)

donatobox-diagnostics-20220730-1012.zip

Edited July 30, 2022 by donatobox

[Squid]

See if disabling Secure Boot in the BIOS helps you out

[Squid]

But, mcelog is reporting memory errors

SOCKET 0 CHANNEL 0 DIMM any
corrected memory errors:
	1 total
	1 in 24h

and you should replace the dimm

[donatobox]

I replaced the dimm, but keep getting the same warning for the same dimm slot.

 

I'm going to try using a different slot next (even though it's not recommended to skip the first slot in the supermicro manual.

 

 

[ChatNoir]

Maybe try to re-socket the CPU and check the pins. A bad contact can sometimes RAM issues as the memory controller is in CPU.

Or the MB slot is somehow faulty ?

[donatobox]

It's been about 24 hrs without any MCE errors, so I'm hopeful that using a different slot fixed the problem. But I'll keep checking for a few more days before I get complacent