Server Critical Fault (Service Information: Power On Fault, System Board, AUX/Main EFUSE

[Gflex78]

Hi,

We have one DL 380 Gen9 (M3G77A) with dual PSU 754381-001 800 W firmware 1.00.

We bought the GPU kit with and the GPU riser. We put in a HP nVidia Tesla P100 16GB PCIe 3.0 x16 Q0E21A and when we power on the server we receive the error

Server Critical Fault (Service Information: Power On Fault, System Board,  AUX/Main EFUSE Regulator 1 (10h))

Any help would be appreciated.

[Suman_1978 · Accepted solution]

Hi,

This should be related to Power cabling or Power supplies.  Usually we see this if power cables are not connected properly or insuffucient power.

Please check the QuickSpecs page# 31.  Also this advisory.

Thank You!
I work with HPE but opinions expressed here are mine.
Recent Support Video Releases

[Gflex78]

Hi Summan!

Thanks for pointing me in the right direction.. Found this note

For doublewide GPU support you are required to have at least 1x1400W Power Supply (720620-B21),

We will replace them and post and update.

[Gflex78]

Hi Suman,

We replaced the PSU with 2 x 720620-B21 HP 1400W Flex Slot Platinum Plus Hot Plug Power Supply according to the table below. But we still get the same error. The  advisory that you post is regarding Gen 10 servers and the firmware cannot be updated with firmware from that link and I cannot find any firmware to update for the 720620-B21 HP 1400W Flex Slot Platinum Plus Hot Plug Power Supply. I don’t know how they can have different firmware. One has 1.01 and has 1.10

Do you know if there is the same issue with 720620-B21 and if where we can find firmware updates for 720620-B21 ?

If there is no issue, do you have any ideas on what else it could be?

Power Supplies Bay Present Status PDS Hotplug Model Spare Serial Number Capacity Firmware 

720620-B21  754383-001  5DRNA0ANF390FQ  1400 Watts  1.10  

720620-B21  754383-001  5DRMV0A4D171LA  1400 Watts  1.01

[Suman_1978]

Hi,

I could not find any firmware for that Power supply.
But for your issue, I would suggest to update Firmwares via SPP.
If still not fixed, please get a support case logged with HPE.

https://community.hpe.com/t5/proliant-servers-ml-dl-sl/dl360-g9-bad-fuse-system-board-p12v-main-aux-regulator-1-10h/td-p/7057584
https://community.hpe.com/t5/bladesystem-server-blades/server-critical-fault-bl460c-gen-9/td-p/7158833

Thank You!
I work with HPE but opinions expressed here are mine.
Recent Support Video Releases

[Gflex78]

Hi Suman,

The idea to just put in the PSU with the 1.10 firmware never crossed our mind.

We tried it today and the server boots.

Really strange that the firmware update for the PSU cannot be applied on Gen9 servers.

And there is no update for it in PSP.

Now we are trying to find another PSU with firmware 1.10

Thanks for your help, you receive a gold star in my book!

[mkbean]

I don't have a contract/warranty with HP and thus can't download the SPP. Do you know of another way to potentially upgrade the PSU to 1.10 so I can use the Tesla card I got? 

[Beginner1]

Were you able to resolve this and please how did you resolve it?

[Gflex78]

No, we were never able to resolve this. We tried all cabling and changed PCI risers, but the server wouldn’t boot with the card. I talked to one of our suppliers, and they said they only recommend buying CTO (Configure to Order) servers with high-power-consuming GPUs. Otherwise, you’re in for a long and bumpy ride with firmware mismatches, cabling issues, and other problems. So, they only recommend buying GPUs that don’t require power cables if the server is not CTO. Additionally, there isn’t much documentation on cabling and related topics.

[Beginner1]

My dl380 gen9 server tries to boot but goes off. When I accessed the ILo to see what's happening I discovered that the list of Firmwares in the system information > firmware information, have all disappeared (removed without any action taken to remove them) apart from the ILo firmware which is left in the firmware list. Also the device inventory lists are not showing and I checked the IML, I saw messages like; 1 server critical fault (Service information: Input Power Loss, Power Supply, Power Supply 1 (03)). 1 server critical fault (Service information: Input Power Loss, System Board, AUX/Main EFUSE Regulator 1 (10h)). 56 Critical temperature threshold exceeded. Do you have any ideal of what I can do to resolve this issue?

[mkbean]

For me it was buying the right cable for the GPU. Seems odd but it worked.

I got this on eBay
10pin to 8pin Cable for HP ML350 G9 and Nvidia K80/M40/M60/P40/P100 PCIE GPU

[Beginner1]

My server is blinking green and red lights and if it tries to reboot it goes off and start blinking yellow and red lights. I can only access ILo of the dl380 gen9 server from another computer. It doesn't boot and no display in the monitor. No GPU in the riser card and the server is not coming up.