Can ESXi hosts with different memory types (RDIMM/LRDIMM), sizes, and speeds coexist within the same vSphere cluster?
search cancel

Can ESXi hosts with different memory types (RDIMM/LRDIMM), sizes, and speeds coexist within the same vSphere cluster?

book

Article ID: 447816

calendar_today

Updated On:

Products

VMware vCenter Server

Issue/Introduction

Administrators often need to know if ESXi hosts with varying physical memory configurations. Such as different RAM types (e.g., RDIMM vs. LRDIMM), module sizes (e.g., 16 GB vs. 64 GB), and clock speeds (e.g., 3200 MHz)—can safely coexist within the same VMware vSphere cluster. This article clarifies cluster compatibility and outlines the specific impacts on vMotion, Distributed Resource Scheduler (DRS), and High Availability (HA).

Environment

VMware vSphere ESXi 7.x

VMware vSphere ESXi 8.x

Cause

Informational inquiry regarding hardware compatibility and vSphere cluster design.

Resolution

Yes, VMware vSphere supports clusters containing hosts with heterogeneous memory configurations. vSphere abstracts physical hardware into a shared resource pool, allowing for variations in RAM capacity, speed, and type across different hosts in the same cluster.

Technical Guidelines:

  • Host-Level Constraints: While the cluster supports mixed memory across different hosts, server manufacturers prohibit mixing memory technologies (such as RDIMMs and LRDIMMs) within a single physical server motherboard. Population must follow the hardware vendor's specific guidelines per server chassis.
  • vMotion Compatibility: Memory type and speed do not affect vMotion compatibility. Virtual machines can seamlessly vMotion between hosts with different RAM setups without interruption.
  • DRS Behavior: The vSphere DRS algorithm prioritizes virtual machine resource satisfaction. Provided that VMs are receiving the resources required to operate optimally on their current host, DRS will not initiate migrations merely to equalize the number of VMs across the cluster. 
  • HA Configuration: In clusters with unbalanced host capacities, configure the vSphere HA admission control policy to use Cluster Resource Percentage. This setting calculates failover capacity based on total aggregated resources rather than assuming all hosts are identically sized.

Additional Information

Performance Best Practices for VMware vSphere 8.0 Update 3