12 Posts
0
1464
R820 with pem cant not post: cpu in socket 4 did not come out of reset
Recently, i bought a pem board for my R820 to use 4 cpu,but when i install all 4 cpus and dimms then power it up,stuck in configure memory with fan speed 100%, 10 or more seconds later the screen dispalys"Fatal error,cpu in socket 4 did not come out of reset".No matter how i swap the cpu and get the same error.After remove the cpu 4,
system boot it up,but always show 2 cpu in post screen and bios,as if there is a cpu 4 in pem board and then can't boot and show the same error.
4 cpu with same model,4 dimms in A1,B1,C1,D1.
Have anyone met this before?
Forgive my bad english,thanks!
Lzp960428
12 Posts
0
August 9th, 2021 04:00
Eventually,this problem has fixed by replacing the PEM board.
The new PEM board part number is 0DPKTV which you mentioned in your first post on this topic,its a NTPM version match to the motherboard version.
I dont konw if this is a compatibility issue between motherboard(NTPM) and PEM(TPM), or maybe it just a broken PEM installed in a good motherboard. In the first,i suspect my motherboard is broken,and now, im lucky,its good!
If somebody have the same issue with me, replacing the PEM board to the same version and try it!
Thanks for your help,Maria!
Lzp960428
12 Posts
0
August 3rd, 2021 09:00
Also i checked the cpu socket pins,did not found any damage,it was strange. bios also was up to date.
DELL-Chris H
Moderator
Moderator
•
8.4K Posts
0
August 3rd, 2021 13:00
Lzp960428,
I would start with verifying compatibility. Would you confirm the part number for the motherboard, as well as the power supplies installed? Regarding the power supplies, i ask as with four processor configurations they require two PSUs (2 x 750W or 2 x 1100W).
Let me know what you see.
Lzp960428
12 Posts
0
August 3rd, 2021 18:00
The main motherboard P/N: 0PFG1N, this one is brand new board used to replace my old board that was damaged.
PEM board P/N: 08HJ4P, this is second hand board which passed the test by the seller before he sended to me.
Maybe the psu is problem? I use two 495w psus,I know 4 processors configration need more power,but I only use 4 E5 4603V2 to test if it can boot?
DELL-Shine K
3K Posts
0
August 3rd, 2021 19:00
Can you also check whether iDRAC FW is also up to date?
Lzp960428
12 Posts
0
August 4th, 2021 01:00
IDRAC and Lifecycle FW was up to date,nothing diffrence,always get the same error.
I ordered two 1100w psus and it is already on the way.
Lzp960428
12 Posts
0
August 4th, 2021 23:00
Relplaced 2 495w to 2 1100w did not soluve this problem.
after stuck in "configure memory" without "done", display "Fatal error,cpu in socket 4 did not come out of reset" , also clear the NVRAM, but the good thing not happen as i will, is there a compatibility issue between the main board and PEM board?
Dell- Maria J
Moderator
Moderator
•
278 Posts
0
August 5th, 2021 01:00
Hi,
I've made some research and would like to provide you follow troubleshooting steps, I hope it helps to resolve your issue.
1)Could you please test server without PEM? Does server boot? In case server boots, the question could be related with compatibility between motherboard and PEM. Regarding part numbers you provided:
PFG1N - Motherboard NTPM
8HJ4P - PEM
I would like to recommend use Motherboard and PEM in follow combination:
Motherboard NTPM (like you have) with PEM NTPM or Motherboard TPM and PEM (like you have)
Part numbers of PEM NTPM are follow:
7TJ0F, DPKTV, 3H25N
2) Could you please write installed BIOS version? CPU, which were installed in server are v2, so that means this version of processors is not supported with BIOS version below 2.00.20, according follow documentation:
https://dell.to/3xqQE4f
Back-flashing to BIOS versions earlier than 02.00.20 is not allowed if Intel Xeon E5-
46xx V2 family processors are installed.
Please ask me if you have any questions.
Thank you,
Maria Januszka
#IWork4Dell
Dell | Social Outreach Services - Enterprise
Lzp960428
12 Posts
0
August 5th, 2021 02:00
Server can nomally boot up without PEM board.
if PEM only installed cpu3,it can pass the post,but only can use cpu1 and cpu2,this is nomal because there are only three configurations(1,2 or 4cpu). The strange thing is that you can't install cpu in cpu4, if you do, the server will not post.
I tried three bios versions:2.10(the original version in the motherboard),2.50 and 2.70(the newest version) and no one success.
The borad model with TPM or without TPM(NTPM),this two model can not insatll together?
Thanks for you help!
Dell- Maria J
Moderator
Moderator
•
278 Posts
0
August 5th, 2021 03:00
Hi,
Thanks for your answer,
Did you additionally check if there are no debris or damage on the QPI connector or power connections, which is located between the System board and PEM?
Motherboard and PEM should be both NTPM or both TPM.
Thank you,
Maria Januszka
#IWork4Dell
Dell | Social Outreach Services - Enterprise
Lzp960428
12 Posts
0
August 5th, 2021 03:00
Yes, i didn't found any debris or damage on the QPI connector or power connections,everything looks so good.The mother borad is a brand new board that i bought five days ago.
AFAIK,cpu keep in reset state in most suitation should be a power problem, power supply voltage below nomal operation voltage make cpu can't receive the good power signal which can cancel reset state of cpu.
Dell- Maria J
Moderator
Moderator
•
278 Posts
0
August 5th, 2021 04:00
Hi,
Could you please addiotionally provide part number of PSU you use? We could check competibility with server too.
An what power supply redundancy was set? When two identical power supplies are installed, power supply redundancy (1+1 – with redundancy or 2+0 – without redundancy) is configured in system BIOS. In redundant mode, power is supplied to the system equally from both power supplies when Hot Spare is disabled. When Hot Spare is enabled, one of the PSUs will be put into standby when system utilization is low in order to maximize efficiency.
Thank you,
Maria Januszka
#IWork4Dell
Dell | Social Outreach Services - Enterprise
Lzp960428
12 Posts
0
August 5th, 2021 20:00
PSU DP/N:0YT39Y.
Two psus part number is the same, i dont think this is a power problem,i tried all power redundancy configurations,mabye the circuit board on PEM or motherboard have broken some electronic components,like capacitor or some microchips ,i already return this PEM borad to seller,let him check this problem,i only have one R820, there is not enough conditions or environment to make me find problem,it's too hard..... if i have some progress,i will post in there.
Dell- Maria J
Moderator
Moderator
•
278 Posts
0
August 9th, 2021 05:00
Hi,
Thank you for your feedback.
I am glad that issue was resolved.
Thank you
Maria Januszka
#IWork4Dell
Dell | Social Outreach Services - Enterprise
luminousplasma
2 Posts
0
September 27th, 2022 21:00
Hey so I have a similar problem as well but I got a different motherboard P/N than the OP but we had the same P/N for the PEM. I am still trying to wrap my head around all of this but I just need to know if the reason it does not POST and goes full fan mode is because I have missed matched components. My motherboard P/N is 066N7P and my PEM is 08HJ4P. Please let me know if I need to replace the PEM or if there is something else I must do. BIOS is on the latest rev too.