Introduction

The Distributed Resource Scheduler (DRS) feature of vSphere offers administrators a hands-off method of balancing workloads across compute clusters and datastores. VMware DRS can be run in a fully-automated mode or in a partially-automated mode for compute, where DRS makes recommendations based on standard deviations and other mathematical computations for balancing workloads in a vSphere cluster. Storage DRS is a storage-specific application of the functionality that uses factors such as latency and capacity to balance storage in a cluster of datastores.

As with anything that is fully-automated, there will always be exceptions that administrators have to account for. There can be virtual machines that should never coreside on the same host, such as domain controllers in the same forest. There can be virtual machines that should always be on the same host to improve the performance due to high amounts of network traffic between them. These sorts of exceptions are managed with...