Repurpose HyperFlex Servers as Regular UCS Servers

Introduction

This document describes the options to re-use HyperFlex servers for different purposes if one decides to move away from the HyperFlex platform. This applies to the Unified Computing System (UCS) nodes in standard HyperFlex (HX) clusters, stretch clusters, and edge clusters.

What are the options to repurpose HyperFlex servers as regular FI-managed or standalone UCS servers?

It is possible to use HyperFlex-servers as a regular UCS-server with these caveats:

Drives

In HyperFlex, the storage controller Virtual Machine (VM) takes care of handling the disks and redundancy in pass-through mode. HX-nodes come with a Serial-attached SCSI (SAS) Host Bus Adapter (HBA) disk controller, which means that the disks are offered to the operating system as Just a Bunch of Disks (JBOD) and disk redundancy must be handled on the OS level.

A HyperFlex node has multiple types of drives installed, so, this is something to keep in mind regarding your use case and boot order.

• Housekeeping
• Capacity
• Cache
• Boot/System (SD-card on M4, M.2 SSD on M5 and M6)

Product-ID or PID

As M4 and M5 HyperFlex nodes have different PIDs, it is possible that they do not show up in a compatibility matrix for the product you are planning to use the server for. In general, the servers are equal to their Unified Computing System (UCS) variant of the same generation and type.

Some products require specific qualified Host Bus Adapter (HBA) and Drives to be fully supported.

M6 Personality

Before considering repurposing M6-based HX servers, consider talking with your account manager as there can be options for converting these nodes to the Cisco Compute Hyperconverged with Nutanix solution.

M6 servers have a server personality set when used as HyperFlex servers. They do not come with a different PID for M4 and M5 servers.

In order to reset or clear the personality, the steps mentioned can be used.

For UCSM-Managed Servers

1. Secure Shell (SSH) to the UCS Manager (UCSM) IP
2. Access the specific server:
   
   

   UCS-A# scope server <server-num>

3. Accesses personality. By default, this setting is 1 for the HX server:
   

   UCS-A/server # scope personality 1

4. Optionally, check the currently set personality:
   
   

   UCS-A/server/personality # show

5. Clear the personality:
   
   

   UCS-A/server/personality # clear personality

6. Commit and save the changes:
   
   

   UCS-A /org/service-profile* # commit-buffer

For Standalone Servers (HX Edge or Already Converted to Standalone)

1. SSH into the Cisco Integrated Management Controller (CIMC) IP.
2. Scope Chassis.
   
   

   C240-M6-01-CIMC# scope chassis

3. Clear the personality:
   
   

   C240-M6-01-CIMC /chassis # clear-personality ?
   <Personality Name> Name of personality to delete. Leave Empty if all personalities needs to be deleted

Support

The UCS hardware stays supported under warranty or contract.

Keep in mind that M4-based HX nodes went out of support on March 12, 2024.