NVMe on Intel S2600CP

J Hart

Active Member
Apr 23, 2015
145
100
43
42
Has anybody documented issues running BIOS 1.06.0001 with updated BMC/ME/SDR? I know for consumer boards the modding community likes updating the OROMS and ME to newer versions when possible.
I had it running that way for about a day. I didn't encounter any problems, but I was a little worried that there would be something unintended if I ran it that way.
 

TrevorX

New Member
Apr 25, 2016
27
4
3
@J Hart
HELP! Please help. Right now I have a dead server.
I followed your steps, right up to 15 - I successfully flashed firmware 1.06.0001, then pulled the USB drives and rebooted. I left the machine for a good 15 minutes, where it didn't get past the initial splash screen (showing Installed BIOS Se5C600.86B.02.06.0002 - I'm not concerned about the firmware number - I realise that will be reset when the full firmware gets flashed), with a little blinking cursor. I shut it down, inserted the target USB drive, then tried starting up again. Just does the same thing. I tried changing the BIOS recovery jumper back to the default position (because I wasn't sure if it should be returned at this point - I couldn't find that in your step-by-step) and then the server does precisely *nothing* - you turn it on, the console screen remains blank, the fans spin up, the diagnostic LEDs don't flicker, there's just an amber and a green. (I don't have the full manual here to check codes - didn't think I'd need it. Stupid me).

So WTH do I do now?? I am seriously $#!@@ing myself here - if I've bricked a perfectly good server my a$$ is on the line.
 
Last edited:

zhoulander

Active Member
Feb 1, 2016
181
46
28
Suggest you set the jumper to BIOS recovery mode and flash the ***rec.cap again. Power down, remove USB drives, set the jumper back to normal, then boot up again.
 

TrevorX

New Member
Apr 25, 2016
27
4
3
I'd love to. But with the jumper in BIOS recovery mode it just sits at the splash screen with a little cursor blinking at the bottom left of the screen. It does nothing else. It won't give me boot options, it doesn't do anything. If I switch the jumper back to default position it won't even show anything on the screen. At least with the jumper in BIOS recovery mode it shows me *something*, but I can't get it to do anything else - it's completely unresponsive.

I can't even contact Intel from here 'cause I don't have my login, so I'll have to leave this gear here and come back tomorrow or something.

FML. Seriously.

Thanks for trying to help, though. Sorry, trying not to be too down, but it's a struggle.
 

TrevorX

New Member
Apr 25, 2016
27
4
3
All out of time - it's 11:17pm here, so I have to call it a night and come tomorrow morning.
 
Last edited:

J Hart

Active Member
Apr 23, 2015
145
100
43
42
Sorry I just got in to work and had some things to take care of. I don't think it is unrecoverable. The fact that anything is coming up means that there is something in BIOS and it isn't just erased.

I think the first thing would be to have a look at the POST code. I definitely experienced this getting stuck while trying to get in to recovery mode thing. If it is sticking at a particular code, that would be where to start.

If you have any devices in the machine (especially the NVMe devices), take them out. The BIOS settings will get reset and the machine will get stuck in POST because of them. See if that works with just CPU, memory and a keyboard. No HD, nothing else.
 
  • Like
Reactions: TrevorX and legopc

T_Minus

Build. Break. Fix. Repeat
Feb 15, 2015
7,352
1,799
113
CA
Sorry I just got in to work and had some things to take care of. I don't think it is unrecoverable. The fact that anything is coming up means that there is something in BIOS and it isn't just erased.

I think the first thing would be to have a look at the POST code. I definitely experienced this getting stuck while trying to get in to recovery mode thing. If it is sticking at a particular code, that would be where to start.

If you have any devices in the machine (especially the NVMe devices), take them out. The BIOS settings will get reset and the machine will get stuck in POST because of them. See if that works with just CPU, memory and a keyboard. No HD, nothing else.
Great advice.

Start as 'bare' as you can :)
 

TrevorX

New Member
Apr 25, 2016
27
4
3
Sorry I just got in to work and had some things to take care of.
Not a problem, I appreciate your help immensely :)

I don't think it is unrecoverable. The fact that anything is coming up means that there is something in BIOS and it isn't just erased.
Yeah it isn't even possible to write to the BIOS, BMC or ME recovery controllers, and they should overwrite whatever is actually stored in the firmware. So I fail to see how you could properly brick one of these even if you were trying! That said, this has me stumped...

I think the first thing would be to have a look at the POST code. I definitely experienced this getting stuck while trying to get in to recovery mode thing. If it is sticking at a particular code, that would be where to start.
It is getting stuck at 01h - First post code after CPU reset. The System Status LED goes amber (fatal error), the System ID LED is off and LSB (LED#0) is green. All other progress code LEDs are off.

If you have any devices in the machine (especially the NVMe devices), take them out. The BIOS settings will get reset and the machine will get stuck in POST because of them. See if that works with just CPU, memory and a keyboard. No HD, nothing else.
There were no PCI-Express devices connected to the board last night, only SATA drives. I have disconnected those as well tonight so the only things plugged into the board are CPUs, RAM, fans and PSU.

I must apologise for my manic-sounding posts last night. I was not handling the stress at all well. Fortunately everyone has been very understanding thus far - sometimes stuff just happens. I need to get it sorted ASAP though!

My thoughts at this stage:

1) The flash MUST have worked initially or it would have thrown up an error when I tried to write 1.06.0001, wouldn't it? I was previously on 02.06.0002, so that should have had the security bit enabled which would have stopped 1.06.0001 from writing, right? So the fact that it successfully wrote 1.06.0001 should mean that the recovery to 02.05.0004 must have been successful...

2) Would a BMC or ME conflict be causing this? Should I try the recovery process for those, too?

3) My USB flash drives are reliable and work fine - every USB 2.0 one I've tried has been detected and been able to flash firmware successfully to the board. However, I can't see what it's doing when it is using BIOS recovery - surely if the drive is detected correctly it should simply follow the process when the jumper is set in recovery mode position?
 

TrevorX

New Member
Apr 25, 2016
27
4
3
Oh I should point out that the only way to get the LED error codes is to run it in BIOS recovery mode. If the BIOS recovery jumper is set to normal position, the only LED that is operational is the System Status LED, which starts solid green then starts flashing. None of the other LEDs does anything.

Interestingly, I just tried connecting to the BMC over the network and managed to get it to respond, kindof. Check out this screenshot of what it's showing. It doesn't actually allow me to access it - it's a dead page. But the fact that it is showing that is interesting.
 

Attachments

TrevorX

New Member
Apr 25, 2016
27
4
3
I've tried different BIOS recovery files from different firmware versions, I've tried four different flash drives (one an old 256mb drive from around 2003), I've tried using the BMC and ME recovery procedures, I've tried USB flash drives one at a time in every one of the four rear USB ports - nothing I do has any effect whatsoever. I've also spoken with Intel and they think the board is unrecoverable - according to them, there are no specific compatibility requirements for USB flash drives, pretty much anything should work, so they ruled out the possibility that I might need some odd or unique drive to get BIOS recovery to work out of hand - as long as it is FAT32 formatted and has the files in the root of the drive, then it should just work.

So now I'm back in the unenviable position of having a dead board that will need physical replacement. Not sure how this is even possible if the BIOS recovery routine should be protected from pretty much everything, but Intel couldn't shine any light on that either (I guess at the end of the day they're support staff, not hardware engineers).

If anyone has any additional suggestions at this point I'd love to hear them - right now I'm off to try to price a replacement :-/
 

TrevorX

New Member
Apr 25, 2016
27
4
3
I definitely experienced this getting stuck while trying to get in to recovery mode thing. If it is sticking at a particular code, that would be where to start.
Can you please confirm if you have seen precisely this same error and behaviour? Or if yours was just something similar, but not quite exactly the same? I'm at the point where absolutely nothing changes, and Intel have no suggestions for me either, so I'm going to have to replace the board. I'd hate to do so if there was still some chance of recovery :-/ The only thing I can think of at this point is that the board is being extremely finicky about the flash drives - can you please let me know what utility / process you used to format the drive that worked, and maybe even what brands you have confirmed work? I have a 4GB USB 2.0 Corsair Flash Voyager (original version) that should be perfect - I've never seen one of these have issues. But you never know... I've tried formatting with RMPrepUSB and Windows Explorer, both in FAT32. I've tried reducing the partition size to 2GB too, and confirmed with Intel that there are no known compatibility issues with drives and firmware flashing (but you have noticed it yourself, so there definitely must be, even if they're not aware of it).

Thanks again,

Trevor
 

J Hart

Active Member
Apr 23, 2015
145
100
43
42
I really am not sure where it went wrong. The machine crashing right at 01h is suspicious. The only thing I can think of is that the BIOS somehow is corrupted. I definitely had mismatched BMC, ME, SDR for a bit and didn't have it crash anything on boot(I went from 2.06 -> 1.06).

I think probably the best bet would be to contact Intel.

Not a problem, I appreciate your help immensely :)

Yeah it isn't even possible to write to the BIOS, BMC or ME recovery controllers, and they should overwrite whatever is actually stored in the firmware. So I fail to see how you could properly brick one of these even if you were trying! That said, this has me stumped...

It is getting stuck at 01h - First post code after CPU reset. The System Status LED goes amber (fatal error), the System ID LED is off and LSB (LED#0) is green. All other progress code LEDs are off.

There were no PCI-Express devices connected to the board last night, only SATA drives. I have disconnected those as well tonight so the only things plugged into the board are CPUs, RAM, fans and PSU.

I must apologise for my manic-sounding posts last night. I was not handling the stress at all well. Fortunately everyone has been very understanding thus far - sometimes stuff just happens. I need to get it sorted ASAP though!

My thoughts at this stage:

1) The flash MUST have worked initially or it would have thrown up an error when I tried to write 1.06.0001, wouldn't it? I was previously on 02.06.0002, so that should have had the security bit enabled which would have stopped 1.06.0001 from writing, right? So the fact that it successfully wrote 1.06.0001 should mean that the recovery to 02.05.0004 must have been successful...

2) Would a BMC or ME conflict be causing this? Should I try the recovery process for those, too?

3) My USB flash drives are reliable and work fine - every USB 2.0 one I've tried has been detected and been able to flash firmware successfully to the board. However, I can't see what it's doing when it is using BIOS recovery - surely if the drive is detected correctly it should simply follow the process when the jumper is set in recovery mode position?
 
  • Like
Reactions: TrevorX

J Hart

Active Member
Apr 23, 2015
145
100
43
42
Can you please confirm if you have seen precisely this same error and behaviour? Or if yours was just something similar, but not quite exactly the same? I'm at the point where absolutely nothing changes, and Intel have no suggestions for me either, so I'm going to have to replace the board. I'd hate to do so if there was still some chance of recovery :-/ The only thing I can think of at this point is that the board is being extremely finicky about the flash drives - can you please let me know what utility / process you used to format the drive that worked, and maybe even what brands you have confirmed work? I have a 4GB USB 2.0 Corsair Flash Voyager (original version) that should be perfect - I've never seen one of these have issues. But you never know... I've tried formatting with RMPrepUSB and Windows Explorer, both in FAT32. I've tried reducing the partition size to 2GB too, and confirmed with Intel that there are no known compatibility issues with drives and firmware flashing (but you have noticed it yourself, so there definitely must be, even if they're not aware of it).

Thanks again,

Trevor

Mine had an issue where it would enter the recovery part of the BIOS(I think it was 11100100 and then go to amber with a CATERR. This was definitely something to do with 01.06 recovery on the USB drive. If I used a USB stick with 02.06 recovery on it, it would boot into the EFI shell just fine. I wasn't able to solve that short of just adding the 01.06 after booting to EFI.
 
  • Like
Reactions: TrevorX

TrevorX

New Member
Apr 25, 2016
27
4
3
Thanks J. Yes, that is not consistent to what I'm seeing - it really isn't doing anything. Intel are as stumped as I am :-/

Thanks for your help, anyway :)
 

Awesomesauce

Member
Apr 20, 2016
104
9
18
41
What bios version it comes from the seller?

Any advangtages of the newest bios and the old one?

I also want pci 3.0
 

daynomate

New Member
May 9, 2016
1
0
1
45
Great news re: PCI-e 3.0! I'm contemplating this board vs ASRock vs Lenovo... Natex have a great deal with this board, 2 cpu's and 64GB RAM but I was worried about the bandwidth issues and the lack of 16x electrical slot. Other than that NVMe issue to work-around is there any other gotchas that might put this board behind say the ASRock Rack EP2C602-4L/D16?
 

ecosse

Active Member
Jul 2, 2013
440
103
43
Thanks J. Yes, that is not consistent to what I'm seeing - it really isn't doing anything. Intel are as stumped as I am :-/

Thanks for your help, anyway :)
TrevorX - how did you get on with this in the end. I flashed one of my m/b to the latest version and I'm now regretting it. I'm wondering whether to attempt an downgrade or just buy another one.

Thanks
 

zackiv31

Member
May 16, 2016
88
16
8
38
Sorry to revive an old thread, but @jhart did you do anything special to be able to be able to boot from PCIe? I've installed Ubuntu in EFI mode to an NVMe PCIe on an S2600ip board, but it never shows up in the boot options menu after the install.
 

bigdata

New Member
Aug 11, 2016
1
0
1
30
some incredible information here thank you to everyone.

I was just after some advice on the pcie issue with the s2600cpj2. I will only be using my pcie slots for graphics cards and was wondering if I were to update my bios and be stuck with pcie 2.0 functionality, would this affect my graphics card's performance? It is my understanding that graphics cards don't use the full bandwidth of pcie lanes but I am getting a new pascal architecture card and am a bit worried that it may be getting close to the max bandwidth.(I mainly want to make sure i wont be restricted by updating my bios as I have messaged Intel about a problem installing network card drivers and I have a feeling they will ask me to update my bios).

any help would be much appreciated :)
 

J Hart

Active Member
Apr 23, 2015
145
100
43
42
I can't find a more recent comparison, but I know I've seen one. Newer graphics cards in PCI Express 2 slots? GeForce GTX Titan Two-Way SLI Scaling: PCIe 2 vs. PCIe 3 Doing normal graphics card stuff the PCIe bandwidth is rarely a limitation. However if you are doing OpenCL or something using a graphics card it really depends on the program. Ultimately I don't think there are many(or maybe any) GPUs where the DMA/RDMA memory bandwidth on the card exceeds PCIe 3.0 x16(128 GT/s or about 16 GB/s). I can't find exact numbers for it but I doubt that it can be reached with a GeForce card but Titan X might be able to max it out with their dual DMA controllers.
 
  • Like
Reactions: bigdata