HyperFlex DPI Installer Hardware Validation Error

[BARishel]

We have 3 HXAF240C-M5SX that we are trying to deploy, but are running into errors during the DPI's validation phase.  The same error displays on all 3 servers. 

Each server has the following config:

2 Intel(R) Xeon(R) Gold 5120 28 cores

192 GB DDR4-2666-MHz RDIMM/dual rank/x4/1.2v

1 2.5in U.2 HGST SN200 NVMe High Perf. High Endurance

1 2.5 inch Enterprise Value 6G SATA SSD 240 GB

6 2.5 inch Enterprise Value 6G SATA SSD 3.8 TB

I also see this in CLI when running a show mini-storage, but I don't see it under inventory. 

UCSM CLI:  /server # show mini-storage

Mini Storage Module:
ID: 2
Type: M2
Model: UCS-MSTOR-M2

Data Platform Installer Error - During Validation Phase.  Each of the 3 servers displays this same error.  

Hardware

Validating server 'sys/rack-unit-2' has M2 SATA disks

Hardware

Validating server 'sys/rack-unit-2' has '1' M2 SATA disks. Failed: Server has '0' M2 SATA disk(s)

We have attempted to do the decommission / re-ack process from a bug we found that shows similar error, but when CLI displays duplicate raid ID's in a show raid-controller error (we are not experiencing this).  

We have also seen this error with using the 3.5a of DPI and the 4.5 of DPI. 

Thank you for your help in advance. 

Bryan Rishel

[Steven Tardy]

Most often the M.2 or really the M.2 controller isn't well seated and "push the contoller in harder to the motherboard" will resolve this.

You can check the `sensors` output for the MINI_STORAGE value from UCSM CLI:

UCSM# connect cimc 1
[ help ]# sensors

And check for the MINI_STORAGE line to have "280", if MINI_STORAGE has "180" then definitely physically reseat:

MINI_STORAGE_PRS | disc -> | discrete | 0x0280 | na | na | na | na | na | na |

The other thing I check is boot the server and [F6] to load the boot menu. The M.2 disk *should* show in the list of boot options if the BIOS can see the disk. If the boot menu does NOT show the M.2, the try a physical reseat.

[BARishel]

Ran the commands on all 3 servers, and received the same results. 

MINI_STORAGE_PRS | disc -> | discrete | 0x0280 | na | na | na | na | na | na |

From the boot menu, I only see these options:

UEFI Built-in EFI Shell

Enter Setup

[Steven Tardy]

Hmm, Still can't tell if this is a physical issue or a configuration issue.

Does the M.2 disk show under:

  UCSM / Equipment / Servers / Server x / Inventory / Storage / Disks / Storage Controller PCH 8 / Expand / Disk 1

This disk should show up and if you select show "PID : UCS-M2-240GB" on the bottom right.

If the disks shows in this spot, then there does NOT seem like a physical issue but possibly a UCS configuration issue.

What does UCSM show under:

  UCSM / Servers / Service Profiles / sub-org / HX / rack-unit-1 / Boot Order / Boot Policy Instance

The Hyperflex out of the box settings should work, but would need to see what this is configured.

[BARishel]

---

@Steven Tardy wrote:

Hmm, Still can't tell if this is a physical issue or a configuration issue.

Does the M.2 disk show under:

  UCSM / Equipment / Servers / Server x / Inventory / Storage / Disks / Storage Controller PCH 8 / Expand / Disk 1

This disk should show up and if you select show "PID : UCS-M2-240GB" on the bottom right.

If the disks shows in this spot, then there does NOT seem like a physical issue but possibly a UCS configuration issue.

Answer: No it does not show up on any of the 3 servers.  

What does UCSM show under:

  UCSM / Servers / Service Profiles / sub-org / HX / rack-unit-1 / Boot Order / Boot Policy Instance

The Hyperflex out of the box settings should work, but would need to see what this is configured.

Answer: We do not have the Service Profile options available as of yet because we have not reached the UCSM Configuration point in the DPI. 

---

[Steven Tardy]

Since none of the servers show the disk in inventory, then I would try to physically reseat.

Both the M.2 into the M.2 controller and the M.2 controller onto the motherboard.

[BARishel]

Thank you Steven.  That will be my next step later today when I go into the DC.  I will keep you posted.  

Bryan 

[BARishel · Accepted solution]

Steve, 

We were able to resolve the issue by performing a decommission, re-ack, and then clean firmware update on the servers to 4.1(2a)C.  Validation passed successfully this time and now the M2 disks are showing under the PCH 8 controller.  

Fingers crossed that is completes the configuration step successfully this time. 

Bryan Rishel