VMware ESXi and vSphere Cluster Management

What Is VMware vCenter Server?

Learn what VMware vCenter Server does, how it manages ESXi hosts and virtual machines, its architecture, deployment choices, permissions, HA, DRS, vMotion, and operational dependencies.

VMware vCenter Server is the centralized management platform for a VMware vSphere environment. It gives administrators one management plane for ESXi hosts, virtual machines, clusters, datastores, networks, permissions, alarms, policies, and performance information.

Understanding the distinction between vCenter Server and ESXi is essential. An ESXi host is a bare-metal hypervisor that provides CPU, memory, storage access, and networking for running virtual machines. vCenter Server coordinates and manages one or more ESXi hosts; it does not replace ESXi and is not normally the system where guest virtual machines run.

Why Centralized Management Matters

An administrator can connect directly to each ESXi host and manage its local virtual machines. This is suitable for a small, standalone environment, but the work becomes repetitive and inconsistent as the number of hosts grows.

AreaDirect ESXi ManagementCentralized vCenter Management
InventoryEach host is viewed separately.Hosts, clusters, VMs, datastores, networks, and policies appear in one organized inventory.
ConfigurationSettings must be repeated on individual hosts.Cluster and policy administration can be coordinated across hosts.
MobilityWorkloads are tied more closely to one host.Features such as vMotion can move running VMs between compatible hosts.
MonitoringEvents and performance data are spread across hosts.Tasks, events, alarms, and performance information are available centrally.
DelegationPermissions are mainly local to each host.Roles and permissions can be assigned throughout the inventory hierarchy.

Centralization supports resource sharing, workload mobility, consistent administration, operational visibility, and delegated access. For example, an administrator can put an ESXi host into maintenance mode after moving its running VMs to other hosts, rather than coordinating each workload manually.

If vCenter Server becomes unavailable, virtual machines that are already powered on generally continue running on their ESXi hosts. However, many centralized tasks become unavailable or limited, including inventory changes, vMotion, DRS decisions, cluster configuration, centralized monitoring, and much of the normal administrative workflow. Direct access to an ESXi host remains available for selected recovery and local administration tasks.

Core vSphere Architecture

VMware vSphere is the virtualization platform ecosystem that includes ESXi, vCenter Server, and related management, storage, networking, and availability capabilities.

A typical management path is:

  1. An administrator uses the browser-based vSphere Client.
  2. The vSphere Client connects to vCenter Server.
  3. vCenter Server communicates with registered ESXi hosts and maintains their inventory and status information.
  4. ESXi hosts run virtual machines and provide access to virtual networks, datastores, and physical resources.
  5. Clusters and vCenter services coordinate operations such as VM placement, host maintenance, and recovery.

Management traffic flows between the vCenter Server system and the management interfaces of the ESXi hosts. Separate VMkernel adapters and networks may carry services such as vMotion, storage traffic, and Fault Tolerance logging.

The vCenter Inventory Hierarchy

The inventory is vCenter Server's organized view of infrastructure objects. A common hierarchy contains:

  • Datacenter: a top-level logical container for related virtual infrastructure objects.
  • Folders: organizational containers for objects such as VMs, hosts, networks, or datastores.
  • Cluster: a group of ESXi hosts managed together for shared resources and services such as HA and DRS.
  • Host: an ESXi system registered with vCenter Server.
  • Resource pool: a logical CPU and memory allocation boundary within a host or cluster.
  • Virtual machine: a software-defined computer that runs on an ESXi host.
  • Datastore: a storage container made available to ESXi hosts for VM files and related data.
  • Network objects: virtual switches, port groups, and distributed networking objects used to connect VMs and VMkernel services.

An administrator registers an ESXi host with vCenter Server by supplying its management address and authorized credentials. The host then reports inventory, configuration, health, and status information to vCenter.

vCenter Server Components

ComponentPrimary FunctionAdministrative Importance
vCenter Server ApplianceRuns the vCenter management services as an appliance.Provides the central management platform and its supporting services.
vSphere ClientBrowser-based administrative interface.Provides access to inventory, configuration, monitoring, and workflows.
Embedded databaseStores inventory, configuration, permissions, tasks, events, alarms, and performance data.Supports operational history and management state; its growth requires planning.
vCenter Single Sign-OnAuthenticates users and connects vCenter to identity sources.Provides the foundation for centralized authentication and authorization.
Managed ESXi hosts and VMsProvide the compute resources and workloads represented in the inventory.Are the infrastructure controlled through vCenter Server.
Associated servicesMay include certificates, licensing, content libraries, lifecycle management, and monitoring services.Extend administration, software maintenance, content distribution, and operational control.

vCenter Server Appliance

The vCenter Server Appliance (VCSA) is the standard deployment model for current vSphere environments. It is a Linux-based appliance supplied as a supported virtual appliance rather than as software installed on a general-purpose Windows server.

Older vSphere versions also supported Windows-installed vCenter Server deployments. Those deployments and the historical desktop and web client terminology may appear in older documentation, but current environments generally use the VCSA and the browser-based vSphere Client.

Database and Management State

The embedded database stores information needed to operate the management plane. Examples include inventory objects, configuration state, roles and permissions, tasks, events, alarms, tags, resource pools, and performance metrics. Retention settings and inventory growth affect database size, appliance capacity, backup duration, and recovery planning.

Single Sign-On

vCenter Single Sign-On (SSO) is the authentication service used by vCenter Server. It supports vCenter identities and can connect to identity sources such as Active Directory and LDAP-compatible directories. SSO authentication is followed by authorization decisions based on users, groups, roles, privileges, and permissions.

Deployment Models and Sizing

Current deployments normally use the appliance-based model. Earlier architectures separated vCenter services from a Platform Services Controller, which could be embedded with vCenter or deployed externally. This embedded-versus-external distinction is historical architecture; current designs use integrated services rather than planning new external Platform Services Controller deployments.

Deployment Preparation

Deploying the VCSA requires more than selecting an appliance size. Plan for:

  • The number of ESXi hosts.
  • The number of powered-on and total registered virtual machines.
  • Inventory growth, including folders, tags, networks, datastores, and clusters.
  • Performance data retention and event history.
  • DNS records, including forward and reverse resolution where required.
  • Static IP addressing, routing, gateway configuration, and firewall rules.
  • Reliable time synchronization through NTP or an equivalent design.
  • Network redundancy and management traffic placement.
  • Backup, restore, and disaster recovery objectives.
  • Management-platform availability and recovery requirements.
Planning InputWhy It MattersExamples
Hosts and VMsDetermines appliance resource requirements and supported scale.Number of hosts, powered-on VMs, and total registered VMs.
Inventory growthAffects database use and operational complexity.New clusters, folders, tags, datastores, and networks.
Performance retentionLonger retention requires additional storage and can affect database growth.Capacity and performance history used for trend analysis.
Network designReliable management communication is required for host registration and services.DNS, IP addresses, routing, firewalls, and redundant management paths.
Backup and recoveryProtects the central management state and reduces outage duration.Scheduled appliance backups, recovery procedures, and tested restore plans.
Version and configurationSupported limits differ by vSphere version and design.Use the sizing and scale guidance for the exact release and configuration.

Supported scale limits are version- and configuration-dependent. Do not treat a host or VM limit from an older release as a universal current limit. Use the documentation and sizing guidance for the exact vSphere version, edition, and deployment design.

Features Coordinated by vCenter Server

FeatureWhat It DoesTypical Dependencies
vSphere HARestarts protected VMs on surviving hosts after an ESXi host failure.Healthy cluster communication, accessible storage, sufficient failover capacity, compatible hosts, and correct admission control.
DRSEvaluates workload demand and host capacity to recommend or automate VM placement and balancing.Cluster configuration, compatible hosts, licensing, accurate resource information, and suitable automation settings.
vMotionMoves a running VM from one compatible ESXi host to another.VMkernel vMotion networking, network reachability, CPU compatibility, licensing, destination resources, and compatible VM configuration.
Storage vMotionMoves VM files between datastores while the VM remains running.Compatible datastores, storage connectivity, permissions, licensing, and sufficient destination capacity.
Fault ToleranceMaintains a secondary execution instance for selected workloads to provide continuous availability if a host fails.Additional CPU, networking, storage, compatibility, licensing, and workload requirements.

These features are not guaranteed merely because vCenter Server is installed. Availability depends on the vSphere edition, licensing, host and VM compatibility, shared infrastructure, storage, networking, resource capacity, and configuration.

Example: HA and DRS Cluster

Suppose four compatible ESXi hosts are placed in one cluster. Shared storage is available to the hosts, and redundant networks provide management and vMotion connectivity. After validating capacity and compatibility, the administrator enables HA to restart VMs after a host outage and configures DRS to make placement recommendations or balance workloads automatically.

Example: Maintenance With vMotion

Before patching an ESXi host, an administrator uses vMotion to move running VMs to compatible hosts. The host is then placed in maintenance mode, patched, and returned to service. This workflow reduces application interruption but depends on adequate destination resources and correctly configured vMotion networking.

Identity, Access Control, and Administration

vCenter authorization uses privileges, roles, and permissions:

  • A privilege is a specific allowed action, such as powering on a VM or changing a network.
  • A role is a collection of privileges.
  • A permission assigns a user or group a role on an inventory object.
  • Permission inheritance allows access assigned at a higher inventory level to flow to eligible child objects.

A practical access-control design uses directory groups, assigns permissions at the narrowest suitable inventory level, and separates duties. For example, a help desk group might be allowed to power VMs on and off within one folder, while virtualization administrators manage hosts, clusters, networking, and permissions.

Follow least privilege: grant only the capabilities required for a job, review inherited permissions, avoid broad administrator access, and limit routine use of built-in administrator accounts.

Inventory, Monitoring, and Operational Data

vCenter records the objects and state needed to manage the environment. This includes:

  • Datacenters, folders, clusters, hosts, VMs, networks, and datastores.
  • Resource pools, tags, policies, and configuration state.
  • Users, groups, roles, permissions, and inheritance.
  • Tasks and events showing completed, failed, or in-progress operations.
  • Alarms for host, VM, storage, networking, capacity, and connectivity conditions.
  • Performance metrics used to investigate resource demand and trends.

Administrators use event history and alarms to distinguish configuration problems from failures. For example, an alarm may identify datastore capacity pressure, a disconnected host, a failed VM operation, or a networking condition. Retention settings should be included in capacity planning because performance history and event data increase database and storage requirements over time.

Administrative Workflow

  1. Deploy vCenter Server Appliance: Use the supported VCSA deployment workflow. Provide valid DNS records, static network settings, gateway information, NTP settings, and administrator credentials. Select a deployment size using the guidance for the target vSphere version and inventory.
  2. Open the vSphere Client: Connect through the vCenter HTTPS address and verify that the appliance services are healthy.
  3. Configure identity and licensing: Add directory identity sources as appropriate, establish administrative groups, and apply licenses.
  4. Create inventory containers: Create a datacenter, folders, and clusters that reflect the operational design.
  5. Add ESXi hosts: Supply each host's management address and authorized credentials. Validate the host identity before accepting its certificate, then assign the host to the correct datacenter or cluster.
  6. Configure networking and storage: Validate virtual switches, port groups, VMkernel adapters, VLANs, datastores, redundancy, and host consistency.
  7. Create or register VMs: Creating a new VM provisions a new virtual machine. Registering a VM adds an existing VM configuration to the inventory. Adding an ESXi host is a separate operation and does not create a VM.
  8. Enable cluster services: After hosts are validated, configure HA and DRS. Review admission control, failover capacity, automation mode, compatibility warnings, and restart priorities before enabling production automation.
  9. Apply permissions and policies: Assign directory groups and narrowly scoped roles at appropriate inventory levels, then test the resulting access.

Cluster Preparation Checklist

  • Confirm compatible ESXi versions, CPUs, licenses, and firmware where applicable.
  • Configure consistent virtual networking on participating hosts.
  • Create and test VMkernel connectivity for management and vMotion.
  • Configure additional VMkernel service networks, such as storage or Fault Tolerance logging, when required.
  • Present compatible shared storage when the selected feature requires it.
  • Check redundant management paths, datastore accessibility, and available failover capacity.
  • Enable HA and DRS only after reviewing warnings and the intended automation behavior.

Dependencies and Limitations

Common dependencies for vCenter-managed operations include reliable DNS, NTP, IP reachability, firewall allowances, compatible host CPUs, sufficient resources, shared storage where applicable, and correctly configured VMkernel networking.

Each ESXi host also has a direct management interface. It can be used for limited local administration or recovery if vCenter access is lost, but direct host management does not provide the complete centralized inventory, cluster coordination, permissions model, or automation capabilities of vCenter Server.

Availability of the management platform matters even when workloads continue running during a vCenter outage. vCenter Server should be monitored, backed up, and included in disaster recovery procedures. A backup protects management configuration and inventory state; it does not replace a tested recovery plan for the ESXi hosts, storage, networks, and workloads.

Troubleshooting Common Problems

An ESXi Host Cannot Be Added

  • Validate forward and reverse DNS resolution where required.
  • Verify IP reachability and firewall rules between vCenter Server and the ESXi management interface.
  • Confirm the host credentials and check that the host is not already managed by another vCenter instance.
  • Review the certificate prompt only after validating the host identity.
  • Check time synchronization on vCenter and the host.

vMotion Is Unavailable or Fails

  • Confirm that each host has a VMkernel adapter with vMotion enabled.
  • Test VMkernel connectivity and verify VLAN consistency.
  • Check CPU compatibility, licensing, host compatibility, and VM configuration constraints.
  • Verify destination host resources and any storage requirements.

HA Cannot Be Enabled or VMs Do Not Restart

  • Check cluster health, host connectivity, datastore accessibility, and management network redundancy.
  • Review HA admission control and confirm that sufficient failover capacity exists.
  • Inspect host isolation settings and HA-related events.
  • Confirm that the VM is eligible for protection and that its restart priority is appropriate.

A User Can Log In but Cannot Perform an Action

  • Identify the inventory object where the permission was assigned.
  • Check the role's privileges, directory group membership, inheritance, and conflicting assignments.
  • Verify that the user authenticated through the intended identity source.

The vSphere Client Cannot Connect

  • Confirm that the VCSA is powered on and its management services are healthy.
  • Check DNS, routing, HTTPS connectivity, certificate validity, and time synchronization.
  • Review appliance resource pressure and available disk capacity.
  • Use direct ESXi access only for necessary recovery tasks while investigating the central management outage.

Key Terms

  • VMware vSphere: The VMware virtualization platform ecosystem containing ESXi, vCenter Server, and related capabilities.
  • ESXi host: A bare-metal hypervisor that runs virtual machines.
  • VCSA: The appliance-based vCenter Server deployment.
  • Datastore: A storage container made available to ESXi hosts for VM files and related data.
  • VMkernel adapter: An ESXi networking interface used for management and services such as vMotion, storage, and Fault Tolerance logging.
  • vMotion: Live migration of a running VM between compatible hosts.
  • Storage vMotion: Migration of VM storage files between datastores while the VM remains running.
  • HA: A cluster service that restarts affected VMs on surviving hosts after an ESXi host failure.
  • DRS: A cluster service that recommends or automates VM placement and balancing.
  • Fault Tolerance: A service that uses a secondary execution instance to provide continuous availability for selected workloads.

Exam-Relevant Summary

  • ESXi is the hypervisor that runs VMs; vCenter Server is the centralized management platform.
  • vCenter Server organizes infrastructure in an inventory and coordinates hosts, clusters, permissions, alarms, tasks, and performance data.
  • The current standard deployment is the Linux-based vCenter Server Appliance.
  • The vSphere Client is the current browser-based management interface.
  • HA, DRS, vMotion, Storage vMotion, and Fault Tolerance require appropriate licensing, configuration, compatibility, networking, storage, and capacity.
  • Loss of vCenter access generally does not immediately stop already-running VMs, but centralized management and automation are impaired.
  • DNS, time synchronization, network reachability, backups, monitoring, and tested recovery procedures are foundational dependencies.
  • Use directory groups, least privilege, narrowly scoped roles, and carefully reviewed permission inheritance.

For a related reference, see What Is vCenter Server?