ESXi host fails to boot after Cisco nenic/nfnic driver upgrade on ESXi

Products

• VMware vSphere ESXi

Issue/Introduction

After upgrading Cisco VIC drivers (nenic and nfnic) on an ESXi 8.x host, the server fails to boot the operating system and drops into the UEFI EFI Shell.

Symptoms:

• The EFI Shell displays NVMe or local storage as block devices (e.g., BLK12, BLK13) but fails to map them as File System devices (FS0:).
• Error: "Mapping not found" or "No bootable device found."
• Impacted Components: Cisco UCS B-Series or C-Series servers, Cisco VIC (Virtual Interface Card).

Environment

VMware ESXi 7.x

VMware ESXi 8.x

Cause

The driver upgrade triggers a re-enumeration of the PCIe bus. If the Cisco UCS vNIC/vHBA Placement Policy is not explicitly defined, the virtual interfaces may shift their PCIe addresses. Because UEFI boot entries are tied to specific hardware paths, this shift renders the existing boot entry invalid, forcing the system to the EFI Shell as it can no longer find the EFI System Partition (ESP).

Resolution

To resolve this issue, the PCIe placement must be stabilized in the Cisco UCS Service Profile.

1. Log in to Cisco UCS Manager (UCSM) or Cisco Intersight.
2. Navigate to the Service Profile or Service Profile Template associated with the affected host modify vNIC/vHBA Placement. 
   See Cisco documentation on steps to do this: 
   Setting the vNIC/vHBA Placement
    

Additional Information

If the issue persists after correcting placement, ensure the Cisco VIC Firmware (Adapter Bundle) is compatible with the driver versions.

Drivers on the 5.0.x (nfnic) and 2.0.x (nenic) trains typically require UCS Infrastructure versions 4.2(3) or 5.2(x) depending on the VIC generation.

For more details on how ESXi Determine the ordering see KB 24534