An administrator needs to perform maintenance on a datastore that Is running the vSphere Cluster Services (vCLS) virtual machines (VMs). Which…

vSphere Cluster Services (vCLS) Retreat Mode is a specific feature designed for scenarios where maintenance is required on datastores hosting vCLS VMs. When an administrator needs to evacuate a datastore, they can enable Retreat Mode. This action instructs the vCLS service to automatically move its virtual machines to other available and compliant datastores within the cluster. This process avoids the need for manual intervention like Storage vMotion on the system-managed vCLS VMs, ensuring that cluster services like DRS remain functional while the datastore maintenance is performed.

A. vSphere Distributed Resource Scheduler (DRS): DRS is a consumer of vCLS services for its health and availability; it is not the tool used to manage the placement of vCLS VMs for datastore maintenance.

B. vSphere vMotion: vSphere vMotion migrates the active memory and execution state of a virtual machine between hosts, not its storage files between datastores.

C. vSphere Fault Tolerance: vSphere Fault Tolerance provides continuous availability for individual virtual machines by creating a secondary copy, but it is not a feature for managing datastore maintenance.

1. VMware vSphere Documentation: vSphere Resource Management Guide (vSphere 8)

"vSphere Cluster Services (vCLS)"

section "vCLS Retreat Mode".

The documentation states

"If you want to perform a maintenance on a datastore and you need to move the vCLS VMs from that datastore

you can use Retreat Mode. vCLS VMs are migrated to other datastores in the cluster."

2. VMware Knowledge Base Article 80472: vSphere Cluster Services (vCLS) in vSphere 7.0 Update 1 and newer versions.

Under the section "How to move vCLS VMs off a particular datastore (Datastore Maintenance Mode)

" the article details the procedure for using advanced settings to specify allowed datastores

which is the underlying mechanism for Retreat Mode. It confirms this is the sanctioned method to evacuate vCLS VMs from a datastore undergoing maintenance.