First time posting here, and having a real "awesome" time trying to build my new home server. First mobo arrived DoA (maybe not as the second mobo is exhibiting the same issue).
Here is my build:
chassis: SC826BE1C4-R1K23LPB
mobo: H11DSi-NT
Backplane: BPN-SAS3-826EL1-N4
CPU: 2x EPYC 7551
RAM:
(Current) 4x 32GB DDR4 2666MHz 2Rx4 ECC RDIMM Memory Compatible with MEM-DR432L-HL01-ER26 from ebay (incompatible despite what it says)
(New for testing) 4x ORIGINAL REPLACEMENT MEMORY MEM-DR432L-HL01-ER26 - Supermicro 1x 32GB DDR4-2666 RDIMM PC4-21300V-R Dual Rank x4 Replacement from memory-dot-net (works)
PSU: Dual redundant 1280W (1000W because American 120VAC)
The Issue: At this point it looks like my server issue is either the CPU or the RAM. The second motherboard is giving the same results, getting stuck at post code 16. I'm leaning towards the CPU, but I will have to wait for two more to arrive, as well as the new ram, to be able to test this theory, or maybe it really doesn't like trying to run on only 4 x 32GB of ram. Also, post code 16 has to do with North Bridge initialization, and occurs prior to mem-test from what I can tell. This, along with the fact that this has occurred on two separate motherboards of the same model with every conceivable combination of one or both CPU paired with various ram configurations given what I have available, is what leads me to believe that the CPUs are the issue. With no memory installed the machine is stuck at post code 10, as it should be, instead of 16. Post code 10 is - PEI execution up to and including memory detection. Post code 16 is - Pre-Memory North Bridge initialization (North Bridge module specific). My friends in one of the discord servers I subscribe to and I have come to the conclusion that my issue might be over tightening the EPYC CPUs, as I did not own a torque driver until this point in this build, and I have also bought an official Threadripper torque driver to see if my digital one was inaccurate. I also have two more CPUs, and another set of 4x DIMMS to test with arriving later this week if that doesn't work. I set the torque wrench to 14.07. It was either that or 13.98, and used it for both the cpu and the cooler. I am currently testing with 4 DIMMS and only CPU Socket 1. It is still stuck at post code 16, which is the northbridge initialization. I will now test my other CPU. I have hit a serious wall. I have tried two different CPU in socket 1 on two separate H11DSi-NT motherboards with different bios versions installed, and got the same result, post code 16.
Yay for self built ebay home servers.... (At least I have learned a lot, but at a high cost....)
Update: new RAM arrived, issue was (yes was!😁) twofold: RAM was NOT compatible, AND one of the CPU is vendor locked. System now boots just fine, though I will have to test the two CPU that arrive tomorrow for vendor lock. I hope they are not, as 7551 not P are apparently hard to find.
