VMware ESXi and vSphere Cluster Management

Enable vSphere DRS on a VMware Cluster

Learn how to enable vSphere DRS, choose automation levels, review recommendations, configure affinity rules, and troubleshoot balancing and vMotion issues.

VMware vSphere Distributed Resource Scheduler (DRS) is a cluster capability that evaluates virtual machine resource demand against available capacity on ESXi hosts. DRS helps place workloads where they can run efficiently and can reduce CPU and memory resource contention by recommending or performing virtual machine migrations.

This lesson covers enabling DRS in the vSphere Client, selecting an automation level, reviewing recommendations, configuring placement rules, and validating DRS behavior. Menu names and available options can vary by vSphere version, edition, and client interface.

What vSphere DRS Does

An ESXi host is a physical VMware hypervisor server that runs virtual machines. A cluster is a group of ESXi hosts managed together by vCenter Server for shared resource and availability capabilities.

DRS continuously assesses demand and capacity across the hosts in a vCenter-managed cluster. Depending on its configuration, it can:

Choose an appropriate host when a powered-off VM is powered on. This is called initial placement.
Recommend moving a running VM to another host to improve resource distribution.
Automatically use vMotion to migrate eligible running VMs when balancing is justified.

vMotion is the live migration of a running VM between compatible ESXi hosts. Initial placement happens when a VM is powered on; ongoing load balancing happens later as workload demand and host conditions change. These are related DRS decisions, but they are not the same operation.

Requirements Before Enabling DRS

Before changing cluster settings, confirm the following conditions:

The target is a vCenter Server-managed cluster containing compatible ESXi hosts. DRS is not enabled by selecting an individual standalone host.
vMotion is configured and functional between the hosts that may receive migrated VMs.
Hosts have compatible CPU features, networking, storage access, and sufficient capacity for workloads that may move.
VMkernel networking for vMotion is correctly configured and reachable.
Destination hosts can access the VM's required datastores, port groups, networks, and policies.
Your vCenter account has permission to edit cluster DRS settings and manage related rules.
The installed vSphere edition and version provide DRS and the settings you expect. DRS availability, licensing, and option names can vary.

Test vMotion before enabling fully automated migrations. A successful test should include representative VMs and the host pairs that DRS is expected to use.

Enable DRS from Cluster Settings

The exact labels differ slightly between vSphere releases, but the workflow is generally:

Select the target cluster in the vSphere inventory.
Open the cluster's Manage or Configure view.
Locate vSphere DRS in the cluster settings.
Open the cluster settings editor, commonly through Edit.
Turn on or select Enable vSphere DRS.
Select an automation level and migration threshold or aggressiveness setting.
Save or apply the configuration.

Cluster inventory object
  > Manage or Configure
  > vSphere DRS
  > Edit cluster settings
  > Enable vSphere DRS
  > Select automation level and migration threshold
  > Save

Enabling DRS does not guarantee that VMs immediately migrate. DRS first evaluates demand, capacity, constraints, rules, and the selected threshold. The resulting action may be no action, a recommendation, or an automatic migration.

Configure DRS Automation

The DRS automation level controls whether DRS actions wait for administrator approval or occur automatically. The terminology can vary slightly by release, but the three common modes are manual, partially automated, and fully automated.

Automation level	Initial VM placement behavior	Load-balancing migration behavior	Best use case
Manual	DRS recommends a host when a VM is powered on.	DRS presents migration recommendations for administrator review.	Learning, validation, or environments requiring explicit change approval.
Partially automated	DRS automatically selects an initial host.	DRS recommends balancing migrations, but an administrator applies them.	Production clusters where placement can be automated but migrations need review.
Fully automated	DRS automatically selects an initial host.	DRS can automatically migrate eligible VMs using vMotion.	Tested clusters with reliable vMotion and an approved automation policy.

Use a conservative mode when first enabling DRS in production. Manual or partially automated operation lets you inspect proposed moves and confirm that monitoring, storage, networking, licensing, and application policies behave as expected. After observing the cluster, you can choose fully automated operation if the organization accepts automatic migrations.

Migration Threshold or Aggressiveness

The migration threshold, sometimes called the DRS aggressiveness or sensitivity setting, controls how much benefit DRS expects before proposing or performing a migration. A conservative setting results in fewer migrations and requires a stronger imbalance. A more aggressive setting allows more balancing actions when DRS sees a worthwhile improvement.

Increasing aggressiveness can improve distribution, but it can also increase vMotion traffic, storage activity, and VM movement. Change the setting gradually and compare migration frequency, host contention, application performance, and operational policy after each change.

Verify That DRS Is Enabled and Operating

Open the cluster Summary view.
Confirm that the vSphere DRS status panel is present and reports that DRS is enabled.
Review the cluster balance or resource-distribution indicator.
Open the cluster Monitor area.
Locate vSphere DRS recommendations and review pending actions.

A recommendation to move a VM from one host to another means DRS believes the move could improve cluster balance or placement. Review the VM, source host, destination host, expected resource impact, and any rules before applying it.

A cluster can appear imbalanced without an immediate migration. The imbalance may not exceed the selected threshold, or no suitable destination may have enough capacity. Affinity rules, reservations, host differences, vMotion incompatibility, attached devices, and other policies can also prevent a useful move.

Check	Where to inspect	Expected result	If the result is unexpected
DRS status	Cluster Summary	DRS status is enabled.	Confirm the correct cluster is selected, permissions are sufficient, and the feature is available for the edition and version.
Balance information	Cluster Summary	Resource distribution is reported without unexplained warnings.	Review capacity, reservations, rules, and migration eligibility.
Recommendations	Cluster Monitor and vSphere DRS recommendations	Pending recommendations show the proposed VM and host move.	Check automation mode, thresholds, and DRS health.
Migration readiness	vMotion tests, host networking, storage, and VM configuration	Eligible VMs can move between intended hosts.	Resolve vMotion, storage, networking, CPU, device, or policy issues.

Review and Apply DRS Recommendations

In manual mode, both initial placement and balancing actions normally appear as recommendations. In partially automated mode, DRS can perform initial placement automatically while balancing migrations remain pending for administrator approval. In fully automated mode, DRS can perform eligible placement and migration actions without manual approval.

When reviewing a recommendation, validate:

The VM being moved and its current source host.
The proposed destination host and its available CPU and memory capacity.
Datastore, network, CPU compatibility, and vMotion readiness.
VM reservations, shares, limits, attached devices, security settings, and application policies.
VM-host and VM-VM placement rules.
The expected impact on network and storage resources.

Apply a recommendation only when it complies with operational policy. Recommendations can disappear or change before application because workload demand, host utilization, rules, reservations, or available capacity may change.

DRS Affinity and Anti-Affinity Rules

DRS rules add placement constraints to normal resource balancing. A VM-host affinity rule controls whether selected VMs should run on, must run on, should avoid, or must avoid designated hosts. A VM-VM affinity rule keeps selected VMs together. A VM-VM anti-affinity rule separates selected VMs across hosts.

Rule type	Placement effect	Typical use case	Potential trade-off
VM-host affinity	Keeps or prefers selected VMs on designated hosts, or keeps them away from those hosts.	Licensing, hardware, or application requirements.	Reduces the number of destination hosts and can create capacity pressure.
VM-VM affinity	Keeps selected VMs together when possible or required.	Related service tiers that benefit from host locality.	Concentrates workloads and may reduce fault tolerance.
VM-VM anti-affinity	Places selected VMs on different hosts when possible or required.	Redundant application nodes or separate fault domains.	May be impossible during shortages or maintenance.

Where supported by the vSphere version, distinguish mandatory rules from preferential rules. A mandatory rule must be honored, while a preferential rule is a desired placement that DRS may relax to satisfy resource or availability needs. Strict rules can produce unmet recommendations or reduce balancing efficiency.

Workflow for Creating a DRS Rule

Identify the VMs and, for VM-host rules, the approved or excluded hosts or host groups.
Open the cluster's DRS rules settings.
Create a VM-host, VM-VM affinity, or VM-VM anti-affinity rule.
Select the affected VMs and hosts where applicable.
Choose the rule behavior, such as affinity, anti-affinity, must-run, or should-run behavior available in the deployed version.
Enable the rule and save the configuration.
Validate current placement, future destination choices, recommendations, and maintenance-mode behavior.

For example, two redundant application VMs can use a VM-VM anti-affinity rule so DRS attempts to place them on different ESXi hosts. A licensing-bound VM can use a VM-host affinity rule to restrict it to approved hosts. In both cases, verify that the remaining hosts have enough capacity.

Operational Considerations

Network and storage: vMotion consumes network bandwidth and can create storage-related activity. Monitor these resources, especially during periods of high workload demand or frequent migrations.
Capacity: DRS cannot balance a cluster when no destination host has sufficient CPU, memory, storage access, or policy compatibility.
Maintenance mode: DRS can help evacuate eligible VMs when a host enters maintenance mode. It cannot override incompatible constraints, mandatory rules, unavailable storage or networks, or insufficient destination capacity.
VM eligibility: VM settings, attached devices, host feature differences, CPU compatibility, reservations, security policies, and other configuration details can make a VM ineligible for vMotion or DRS relocation.
Rules: Document intentional DRS rules, including their business reason and owner. Review them periodically as hosts, applications, and capacity change.
Monitoring: Monitor DRS recommendations, migrations, resource contention, vMotion health, and application performance after initial enablement and after changing thresholds.

Troubleshooting DRS

DRS Is Enabled but No Migrations Occur

Possible causes include a cluster that does not exceed the selected migration threshold, a cluster that is already balanced enough, manual automation mode, unavailable capacity, restrictive rules, or VM migration incompatibility.

Review the automation level and pending recommendations.
Check cluster resource utilization and balance information.
Validate vMotion compatibility between relevant hosts.
Inspect DRS rules, reservations, and destination capacity.

A Recommended VM Migration Cannot Be Applied

Check whether vMotion networking is available and correctly configured. Confirm that the destination host can access required datastores, networks, and policies. Also review CPU compatibility, attached devices, security settings, VM configuration constraints, and placement rules.

The Cluster Remains Imbalanced

Review host capacity, reservations, and limits.
Consider a cautious increase to the migration threshold only after confirming that migrations are safe.
Assess whether strict affinity or anti-affinity rules are necessary and correctly scoped.
Identify VMs that cannot use vMotion and determine the specific incompatibility.

DRS Settings or Rule Options Are Unavailable

Confirm that a vCenter-managed cluster object, rather than an individual host, is selected.
Validate that your account has the required vCenter privileges.
Confirm DRS availability for the installed vSphere edition and version.

Practical Enablement Patterns

Administrator-Controlled Migrations

An administrator enables DRS on an existing cluster and selects manual or partially automated operation. DRS then recommends moving a resource-intensive VM from a busy host to a less-loaded host. The administrator reviews the source, destination, rules, capacity, and application impact before applying the recommendation.

Fully Automated Balancing

A cluster with tested vMotion, shared storage or the applicable storage architecture, compatible hosts, and sufficient capacity is configured for fully automated DRS with a moderate migration threshold. DRS automatically redistributes eligible VMs as demand changes. The administrator monitors migration activity and adjusts policy if movement is excessive or insufficient.

Exam-Relevant Notes

DRS is a vCenter cluster service, not a feature configured independently on one ESXi host.
Initial placement occurs when a VM is powered on; load balancing can later use vMotion.
Manual mode requires administrator action for recommendations.
Partially automated mode commonly automates initial placement but leaves balancing migrations for approval.
Fully automated mode can perform eligible placement and balancing migrations automatically.
Rules constrain DRS and can prevent an otherwise beneficial migration.
DRS cannot bypass vMotion incompatibility, missing capacity, unavailable storage or networking, or mandatory placement constraints.

For related configuration, see the vSphere DRS enablement workflow.