Unrecoverable Machine Check Exception

[finlike27]

Hi all,

I've just built my first FreeNAS system, and am stuck at install stage. I suspect that something is wrong with hardware, but don't know how to diagnose this further. Can anyone assist?

Build

Fractal Design Node 804 Case
Corsair 550RMx

Supermicro X11SSM-F Motherboard
Intel Celeron G4920 3.20Ghz
Kingston KVR24N17S8/8 8GB [x2] Kingston KVR24E17S8/8I 8GB [x2]

WD Red Nas 4TB [x4]
Sandisk 120GB SSD Plus


I created a bootable USB stick with FreeNAS-11.1-U6 and tried to install onto the 120GB Sandisk, but got an error as per the attached image.

Can anyone assist with the error message? Is there something I should try?

Thanks!

 

[DrKK]

Well, the X11SSM-F is one of the best boards you can possibly buy for a FreeNAS, so good job choosing that. My main system is an X11SSM. The error code you're getting is hardware related.

The X11SSM, as far as I know, requires unbuffered, ECC RAM for servers. A quick search on your Kingston SKU makes no mention that this is ECC. Since the hardware error you are getting says something about "uncorrectable", there is a high-degree of probability this is related, since ECC is about memory error correction. If your memory is not ECC, there is a 95% chance this is the problem.

Other ideas:

The CPU you have (I've never seen someone use a Celeron on the X11 for FreeNAS, but it should work) is recently launched. Therefore, you need to have the absolute most up-to-date BIOS for your system. There is a 50-50 chance a BIOS update will solve this problem,if the problem is not the aforementioned memory.

 

[DrKK]

Yeah I can seem to confirm that the memory you have is not good enough for that board.

You need ECC Unbuffered/Unregistered RAM (UDIMM). It's more expensive. Check the qualified RAM list on the SuperMicro site.

 

[DrKK]

This is the Kingston part I use in my X11SSM:

KVR21E15D8

 

[finlike27]

Hey DrKK,

Thanks for jumping in to help.

That's a really embarrasing mistake - I double checked ... and I put the wrong model number in, the RAM is actually: Kingston KVR24E17S8/8I 8GB [x2] - which I'm pretty sure is ECC.

I'll edit the first post to make that clear.

Regarding the BIOS - the board is currently on 2.1a [03/07/2018]. I'll update later and report.

Is there anything else I should attempt?

Cheers.

 

[DrKK]

My board is 2.1a. Current BIOS appears to be version 2.2. I guess you already know this.

That's what I'd be doing first.

Updating SM BIOS is a huge pain in the ass. :)

 

[finlike27]

There you go, v2.2 fixed the problem.

Thank you!!

 

[DrKK]

Fascinating. Well I'm sure this thread will help someone else out there. :)

Also, for the record, this conversation inspired me to update my BIOS on my X11SSM to 2.2 last night. Of course, it made no difference for me, since I was already fine with 2.1a.

 

[finlike27]

@DrKK - this is really strange - can we reopen this?

The BIOS update allowed me to install, but I was getting all sorts of instability with the GUI. I would be able to log in, and then 5 minutes later - the GUI would simply hang - and the same 'unrecoverable machine' error would throw up via IPMI.

I've been searching through the forums, and thought that the Sandisk 120GB SSD Plus might have been at fault, as per this bug report.

Disconnected the Sandisk 120GB SSD Plus, and tried a fresh install of FreeNAS-11.2-RC2 on to a clean USB stick - and I'm getting the same errors thrown my way.

Any suggestions?

 

[DrKK]

Almost certainly hardware related.

Bad power supplies cause this kind of thing, as does bad RAM.

I don't know that I have any specific advice. You might try our IRC channels (#freenas and ##freenas-social), or the FreeBSD channels, since this is really a FreeBSD failure, not a FreeNAS failure.

 

[finlike27]

Thanks - perhaps my question needs to be clearer.

Are there any scripts that can help narrow what the issue is? Is it bad RAM? Motherboard? CPU?

 

[Chris Moore]

Supermicro X11SSM-F Motherboard [BIOS v2.2]
Intel Celeron G4920 3.20Ghz CPU
Kingston KVR24E17S8/8I 8GB [x2]

Two questions:

- Is the CPU compatible with the MB? The CPU is generation 8, and the MB chipset is Intel C236 (supports 6/7)
- Is the RAM compatible with the MB? I've found different spec sheets, some of which claim the RAM is UDIMM and some DIMM

When you have already started one thread on a problem, it is usually not a good idea to post additional information in a different thread. Anyone that would help you will help you in the original thread if you post the correct information about the system.

 

[Ericloewe]

Things are much easier when there is no crossposting going on: https://forums.freenas.org/index.php?threads/x11ssm-f-build-combo.71869/

As for the mystery error, bad power seems less likely, now that it's revealed that the CPU is unsupported in that motherboard. It's a Coffee Lake CPU, which is not supported with the C23x PCH and requires a C24x PCH board, like the Supermicro X11SCH-F.