Showing overvolt in IPMI

Lococoyote

Prominent
May 5, 2017
6
0
510
I have read and googled my butt blind, and can't find an answer.

Asrock E3C224D2I with a Xeon E3-12220 V3 CPU.

After a bios update today the IPMI is showing VCORE of 1.8 with a warning that it's critical. Going into the BIOS also shows the same voltage. Try as I might, I have not been able to tell if that is correct or not for the voltage. You would think you could look up the intended voltage in the cpu specs, but I can't find it. All CPUZ info I've found indicates it should be around 1.2 and the thresholds in the IPMI indicate that would be good.

I can not find anywhere in the BIOS to adjust the voltage, so my concern is that the CPU is getting over stressed.

Any help in case I'm at risk of toasting a $2K server would be appreciated.
Thanks
 
Stop what you're doing.

1) I don't care what the machine is doing, unplug it now.
2) Press and hold the power button for 3 seconds to drain all remaining power.
3) Open up the case, remove the CMOS battery from the motherboard.
4) Take a nap, get a snack, do literally anything while you wait a couple minutes.
5) Put the battery back in, plug the machine in, go into BIOS setup and load optimized defaults.
6) Don't let Windows boot. Boot into setup, check CPU voltage.
7) If voltage is still high, unplug it again and we'll figure out what the heck is going on.
 


well perhaps not such haste, but yes, act fast. if you can find the clear CMOS pins and short them for 5 seconds with a screw driver tip or something metal for 10 seconds, that will do the same thing.
 


Many Intel boards need to reboot while the Clear CMOS pins are shorted for the CMOS to actually clear. Then I would shut it off, boot into BIOS and check voltage.
 


yes, thank you for clarifying, do this while the PC is shut down. hitting that while it is running might not be a good idea.
 

Lococoyote

Prominent
May 5, 2017
6
0
510
Thank you both. I thought it was not right. I took a screen shot of the VCORE voltage and temp, but can't find a way to post it.
I've read my eyes red because I do have an IT background and was not dismissing the possability of a bug in the BIOS (done just before this) or a bug in the IPMI display/interpretation of the info. Part of the problem is that I knew temp would have to go up, but it shows 47C

Server has been taken down.

Thanks
 


If BIOS information doesn't change, it should stay set up. Sometimes these workstation boards apply new information upon a reset though, so no guarantees.

Like if it failed to recognize the CPU properly last time, but recognizes it properly this time, then the BIOS information might change. Since you're already having CPU issues, this is definitely a possibility.
 

Lococoyote

Prominent
May 5, 2017
6
0
510
CMOS battery was removed for better part of an hour, and voltage remains the same. Manual for motherboard does not seem to indicate that there is a CMOS clear jumper so given the responses above, not sure if the CMOS cleared or not.

Edit: I did have to reset the system date and time, so that would indicate a change.

So, was the voltage high all along and the old CMOS didn't report it correctly, or did it jump up with the CMOS upgrade today?

Too damn much work for an unpaid civilian :pt1cable:

Thank you all for the input!
 


Hmmm. If the temps are still not elevated, then it could be a bug especially if it is floating around after a clear CMOS. It sadly might be time to try to reach the motherboard company.
 

Lococoyote

Prominent
May 5, 2017
6
0
510
Robert: Thanks, after the displayed voltage did not change (still showing 1.8) after a CMOS clear I figured the same thing. I'll message the MB manufacturer and see if I get a response.

Will post back any info/solution for everyone's benefit.
 

Lococoyote

Prominent
May 5, 2017
6
0
510
Asrock to their credit answered promptly and repeated. Only suggestion provided was reset the BIOS to defaults which had no effect. We swapped about 6-8 emails and the same person was back to me on an average of 20-30 minutes. Impressive in this day and age.

He pointed out that the thing to do would be RMA the board for repair/replacement of the sensor as the server boards have a 3 year warranty (something I was not aware of). I'm not keen on having the NAS server down for some time so debating getting a Supermicro to swap out and send in the Asrock. Working with freenas people to see how much of a wrench that would throw into the OS/setup.

I agree it may be a sensor, as the CPU temp is cool.
 

zanderover

Reputable
Dec 29, 2017
1
0
4,510


Has this board/sensor issue been solved? I've run into the same issue with a similar board (226 instead of 224) and i was wondering whether you have manage to solve it