Skip to content

VCP-DCV Section 2 - The VMware Solution

vCenter, clustering services, vSAN, and the features that turn a set of hosts into a managed platform.

vCenter Server

  • vCenter Server Appliance (VCSA) - the Photon OS-based appliance. The Windows version is discontinued, so VCSA is the only deployment form.
  • vCenter Single Sign-On (SSO) - the authentication service issuing tokens for vSphere components.
  • vCenter Single Sign-On domain - vsphere.local by default, distinct from your Active Directory domain. Confusing the two is a common error.
  • Enhanced Linked Mode - multiple vCenter Servers sharing roles, permissions, licenses, tags, and policies, and manageable from one client view.
  • vCenter High Availability (VCHA) - active, passive, and witness nodes protecting vCenter itself.
  • vCenter backup - file-based backup to FTP, SFTP, HTTP, HTTPS, SMB, or NFS. This is the supported protection mechanism and is examined.

Authentication and permissions

  • Identity source - Active Directory over LDAP, OpenLDAP, or an external identity provider for federated authentication.
  • Permission model - a permission is the combination of a user or group, a role, and an object, with an option to propagate to children.
  • Role - a named set of privileges. Sample roles include Administrator, Read-only, and No access, plus sample roles for common jobs.
  • Propagation - permissions applied to a parent object flow to children when propagate is enabled.
  • Effective permissions - where multiple permissions apply, user permissions on an object override group permissions on that object, and permissions set on a child object override those inherited from a parent.

Global permissions apply across all inventory hierarchies, including in Enhanced Linked Mode.

Clusters and their services

  • Cluster - the container on which HA, DRS, and vSAN are enabled.

vSphere HA

  • Purpose - restart VMs after host, guest OS, or (with configuration) datastore failure. Downtime occurs; it is a restart, not continuity.
  • Master and subordinate hosts - one host is elected master and monitors the others.
  • Heartbeat - network heartbeats, with datastore heartbeating as a secondary channel to distinguish a failed host from an isolated one.
  • Host isolation - a host that cannot reach the network but is still running. Isolation response options are leave powered on, power off, or shut down.
  • Admission control - reserves capacity so that failover can actually succeed. Policies: cluster resource percentage, slot policy, or dedicated failover hosts. Disabling admission control allows more VMs but may leave insufficient capacity to restart them.
  • VM restart priority and dependencies - ordering for restarts.
  • Proactive HA - evacuates VMs from hosts a hardware vendor reports as degrading, before failure.
  • VM Component Protection (VMCP) - responds to storage failures, specifically permanent device loss (PDL) and all-paths-down (APD).

vSphere DRS

  • Automation levels - manual, partially automated (initial placement only), fully automated (placement and ongoing migration).
  • Migration threshold - how aggressively DRS balances.
  • Affinity rules - keep VMs together (affinity) or apart (anti-affinity), and VM-to-host rules with "should" (soft, best effort) or "must" (hard, enforced) semantics.
  • Predictive DRS - uses vRealize/Aria Operations forecasting.

Anti-affinity separates VMs that must not fail together, such as two domain controllers. "Must" rules are enforced even at the cost of availability, which is a real trade-off the exam probes.

Migration technologies

  • vMotion - live compute migration. Requires shared visibility of the VM files or uses the migration network to copy them, plus a VMkernel adapter enabled for vMotion, and compatible CPUs (see EVC).
  • Storage vMotion - live migration of VM files between datastores.
  • Cross-host and cross-datastore migration - both compute and storage at once.
  • Cross vCenter vMotion - between vCenter instances.
  • Long-distance vMotion - supported up to defined round-trip latency limits.
  • Cold migration - the VM is powered off; the fewest requirements.

vSAN

  • vSAN - hyperconverged storage pooling local host devices into a single datastore, enabled at cluster level.
  • Disk group - one cache device plus one or more capacity devices, in the original architecture. Newer Express Storage Architecture (ESA) uses a single-tier pool.
  • Storage policy-based management (SPBM) - policies specify availability and performance requirements per VM or per disk, and vSAN places data to satisfy them.
  • Failures to tolerate (FTT) - how many host failures the data survives. FTT=1 with RAID-1 mirroring needs 3 hosts; RAID-5 erasure coding needs 4; RAID-6 needs 6.
  • Witness - a component providing quorum, and a dedicated witness host or appliance in stretched and two-node clusters.
  • Stretched cluster - a cluster split across two sites with a witness at a third.
  • Deduplication and compression - space efficiency, available on all-flash configurations.

FTT drives the host count. Questions giving a required FTT and asking for minimum hosts are common.

Content and lifecycle

  • Content Library - a repository of templates, ISOs, and scripts, which can be published and subscribed to across sites and vCenters.
  • VM template - a non-runnable master image for deploying VMs.
  • Clone versus template - a clone is a copy made now; a template is a master intended for repeated deployment.
  • Guest customization specification - applies hostname, network settings, and domain join during deployment.
  • vSphere Lifecycle Manager (vLCM) - manages ESXi images and firmware for a cluster using a desired-state image, replacing the older baseline approach from Update Manager.

Exam pointers

  • HA restarts, FT continues, DRS balances. Do not mix them up.
  • Admission control exists to guarantee failover capacity; disabling it risks failed restarts.
  • Datastore heartbeating distinguishes host failure from network isolation.
  • FTT and the RAID type determine the minimum vSAN host count.
  • vCenter is protected by file-based backup and optionally VCHA.
  • Permissions: user overrides group on the same object; child overrides inherited parent.
  • vLCM uses desired-state images; Update Manager baselines are the legacy approach.

Official documentation

📖 vSphere availability - HA, FT, and DRS reference 📖 vSAN documentation - architecture and policies 📖 vCenter Server installation and setup - VCSA deployment