ASRock Rack X470D4U2-2T

  1. Make sure DHCP is working on your router
  2. Perform a CMOS Reset on your motherboard
  3. Attach only the IPMI network (using a working network cable :wink: ) and power cable
  4. Turn on the system

If the IPMI then doesn’t get a DHCP assigned IP, then I’m afraid you have DOA.

Bad news I’m afraid… I’ve received a response from Asrock Rack, with “official statement” from AMD on this, regarding ECC on this mobo (and AM4 in general):

Dear Mastakilla,

So many thanks for you detail experience.
We will share this information to RD 

However we got AMD official respond today

  1. AM4 support ECC function
  2. AM4 does not support ECC error reporting function

Here is the conclusion:
AM4 platform CPU (Ryzen 1000,2000,3000 series) can all support ECC correction, but not ECC report function

Best regards,
Kevin Hsiueh
Asrock Rack Incorporation

To which I responded:

Hi Kevin,

Thanks for getting back to me!

That is very unfortunate news…

Does this mean that the sensors for “DRAM ECC Error A1/A2/B1/B2” in the IPMI Event Log are unused and always will remain empty, even if memory errors do occur?
Do you know why these sensors then exist on this board? Were they simply copied over from an existing Intel /TR4 / Epyc Board, without testing them? Or were they added explicitly, but weren’t you aware of this missing feature (and also didn’t test it)?

Kind regards,

Mastakilla

And their response:

Dear Mastakilla,

According to AMD, X470 is desktop MB, and our QT won’t test ECC report function on desktop MB.
We follow AMD POR to writes specification.
In order to prevent misunderstanding, we will also remove ”DRAM ECC Error A1/A2/B1/B2” in the IPMI Event Log”.
Thanks for doing so many test and kind remind, and we will pay more attention on similar case in the future.

Best regards,
Kevin Hsiueh
Asrock Rack Incorporation

So no ECC reporting is supported…

Not entirely sure of this, but doesn’t this mean that:

  • there is no way to know for sure ECC is actually doing something or to validate that it actually works (even for Asrock Rack or AMD themselves).
  • there is no way to know if your memory is stable or not (ECC might be correcting errors all the time without you knowing about it). This is especially relevant if you want to overclock it.

I’m also not entirely sure all of this is true. Wendell told me he knew about people who reported logged error corrections on Ryzen. Perhaps AMD / Asrock Rack told me this to stop asking annoying questions about it? I certainly hope so :wink: (please prove me wrong)

3 Likes

AM4 most definitely supports ECC memory. It is a requirement for Ryzen Pro support that both the 1Gb and 10Gb versions support.

I’m smelling BS. Hopefully ASrock is not trying to get away with feature removal.

Normally with the dedicated IPMI port you can get rid of this problem.

Following up from my previous post. The problem was a short with the motherboard and case. I removed it from the case and unplugged all but the Power button from the header and was able to IPMI in. (There’s an IPMI activity light that would flash for 1/10th of a second when power was first applied then never again, so anyone reading this later you should see a blinking IPMI light between the PCIe slots.)

There was still a problem with the CPU but that probably was just the thermal paste from taking the fan on / off to try different RAM combinations. So cleaning the CPU and heatsync with Rubbing alcohol, reapplying thermal paste and I can POST!

I’m now installing proxmox from a NFS mounted ISO and the motherboard see’s all 128GB of memory. Direct support from ASRock has well Rocked.

1 Like

Using a Ryzen 5 2600 I was seeing corrected ECC errors in Memtest86 for one of my two sticks of RAM (which I then RMA’d). This was with X470D4U2, BMC 1.60, BIOS P1.50. I never tried injecting errors, though.

2 Likes

Did you also see them in the IPMI Event Log? (if you didn’t check yet, could you perhaps have a look?)

I’ve been looking at this board for an upgrade to my home server, I’m wondering if anyone can tell me how the PCI-E behaves with both slots filled with respect to bifurication?
Board spec says x4/4/4/4 and auto switch to x8 when slot 4 is filled, is it able to do x4/4+x4/4 or x4/4+x8 when the second slot is filled?

The corrected ECC errors are not in the IPMI Event Log. The only things in my log are “Timestamp Clock Synch - Asserted”. A few corrected ECC errors were listed in my ESXi logs, though. They looked like:

2019-09-27T15:06:53.331Z cpu0:2097725)MCA: 136: CE Poll G0 Bf Sd42040000000011b A40000011966b480 Mc000000000000000 P242cd6880/10 Memory error, read

That’s interesting…

@ https://www.ixsystems.com/community/threads/freenas-build-with-10gbe-and-ryzen.77752/page-3 someone with broken DIMMs recently replied that he also had MCA errors (in FreeBSD). But his MCA errors were only regarding the CPU cache (although replacing his faulty memory did fix them). Your MCA errors clearly state “Memory error” however, so it seems like somehow it was reported to the OS…

Could it be that these came from the infinity fabric then? Because infinity fabric memory errors are reported even when using non-ECC memory…
If they are really coming from the memory itself though, then this seems to indicate that your memory errors somehow did get reported :astonished:

Did you perhaps make a screenshot / picture of when Memtest86 reported these “corrected memory errors” on your Ryzen 2600? Did it clearly say that it was a “corrected memory error”? (I never saw this myself, so I have no clue what it would look like)

[SOLVED]
Turns out there is a BIOS setting that needed to be changed:

Advanced -> AMD PBS -> PCIe x16/2x8/4x4 Switch

It was set to 1x16, but if changed to 2x8, both cards and all 8 HDDs showed up in Proxmox.


Hello everyone!

I’ve recently bought this motherboard for a Proxmox server.
The trouble I have is that I have installed a Dell perch H330 flashed to a IT HBA with a guide from servethehome: Google >flash H330< and you will find it (sorry but I’m not allowed to post links)

The HBA can:

  1. Show boot message and find all 8 HDDs on startup.
  2. If I set CSM OpROM to UEFI only, it can display itself in the MB Advanced options.

But for the life of me, cannot display itself or the drives anywhere else in Proxmox and therefore is useless.

How have you setup your bios to let the HBA card and its drives show to the OS?

Server specs:
X470D4U2-2T (BMC fw 1.70.00, BIOS fw P3.30)
Ryzen 7 3700x
2x 16GB kingston ECC ( on the supported RAM list of the MB)
1x 1TB WD red SSD for VMs
2x 120GB Kingston SSD ZFS mirror for Proxmox
2x 120GB Kingston SSD ZFS mirror for Freenas
Dell perch H330 flashed to Dell HBA330 IT mode fw 16.17.00.05
8x 4TB WD red

EDIT:

  • I have installed the HBA on the bottom PCI slot named “PCIE4” in the manual, but it only shows in the MB bios when CSM PCIE6 is set to UEFI.
  • I have also installed a 4 port Intel NIC that worked fine in another pc in PCIE6, that doesn’t show up anywhere either.

Hope Im in the right place.

Turns out there is a BIOS setting that needed to be changed:

Advanced -> AMD PBS -> PCIe x16/2x8/4x4 Switch

It was set to 1x16, but if changed to 2x8, both cards and all 8 HDDs showed up in Proxmox.

1 Like

I had the same issue when I installed my 3700x. Had to manually set 2x8 so all three pcie slots would work.

1 Like

I’m really interested in using this motherboard and a 3700X but for live performance, and using my collection of 1U server cases.

I’m confident there’s plenty of IPC and acceptable latency, but hope to get opinions on heat, build quality and stability.
Won’t be using GFX Card, but the ASPEED chip for meager 2D monitoring.

Love reading posts from folks here. Way above my pay grade, plus I get to learn.

Thanks for any advice.

3700X and a 1U, you might want to look at the Dynatron L3. I bought a couple directly from Dynatron via there website, and got the AM4 brackets to go with it.

I use nothing but Dynatron since P4’s.
Great solutions for tight space.
Plus those Supermicro Barrel fans x 3 blow right through the DIMM’s and take air out the back.

Are these ASRock Rack AMD boards stable?
I’m using a C236, Z97 and H97.
100% but never tried AMD.

I am using mine for an unraid server in a 2U SM chassis. I haven’t had really any issues with it. I am using a LSI HBA in the top slot and a riser card for U.2 in the bottom slot, and have that hooked up to an Intel P4500.

My ECC RAM is OC’d from 2400 to 2666, no issues.

From the platform side, I haven’t really had a single issue with my setup. Linux Kernel is coming along well for AMD too, expecting good things in the 5.x space. HW pass through seems to give users a space to vent at, but I take some of that with a grain of salt as in the unraid forums HWPT can be a mixed bag from all vendors and platforms and cards.

1 Like

I can only speak about the non-10 GbE X470D4U variant but if it is actually booted up and running then it runs solid, using 4 x 32 GB Samsung ECC UDIMMs, DDR4-2666@3200.

The onboard 2D ASpeed GPU dilvers something like a 15 fps signal, if that’s enough for you then you might give it a try.

1 Like

Thanks gents.
Trying to read up on the X570 version in itx.
I’ll definitely stick with ASRock Rack though.
Exciting times.

hey

i am using all 6 SATA slots and pcie4 and pcie6.
can i use both m.2 or do i lose something if i use the m.2 slots?
i am running out of space and would like to add 1 or 2 m.2 ssds