3V0-12.26 Advanced VMware Cloud Foundation 9.0 Architect exam badge

VMware

Advanced VMware Cloud Foundation 9.0 Architect (3V0-12.26) practice questions

★★★★★★ 4.4 43 reviews — aggregated from community votes by visitors who practiced this bank

Questions
64
Exam length
135 min
Passing score
300
  • Instant feedback per question
  • Explanation for every option
  • Free, no signup

Exam overview 3V0-12.26

A customer asks for 99.99% availability across two regions and expects the design to be defended, not merely drawn — that is the level 3V0-12.26, the Advanced VMware Cloud Foundation 9.0 Architect exam, operates at. Passing leads to the VMware Certified Advanced Professional – VMware Cloud Foundation Architect credential (VCAP-VCF Architect), delivered by Broadcom through Pearson VUE. The paper has 60 items in 135 minutes with a 300 pass mark on the 100-500 scaled score.

The blueprint is the mirror image of the Administrator paper. Sections 1, 2 and 3 hold all 20 testable objectives — 1.1 to 1.5 on IT architectures and standards, 2.1 to 2.3 on VMware products and solutions, and 3.1 to 3.12 on planning and designing the solution — while Sections 4 and 5 have none. Nothing asks you where to click or which CLI command to run; the whole paper stays at Day 0 and Day 1 architecture. Candidates who sat it report almost entirely multiple-choice items with only one or two drag-and-drop questions, and enough time to review flagged answers.

Architects, cloud architects, consultants and senior engineers are the intended audience. The method matters as much as the products: requirements, assumptions, constraints and risks must be separated and traced, then weighed against the AMPRS qualities of availability, manageability, performance, recoverability and security. Design work runs from conceptual through logical to physical outcomes, across management and workload domains, region boundaries and Cloud Zone assignments, VMApps versus AllApps organisations, projects and tenancy, monitoring strategy, and recovery design. NSX VPCs with stateful services on Tier-0 and Tier-1 gateways are a recurring design constraint rather than a configuration exercise.

The question bank is built around that judgement. An item may give you a business goal and three candidate designs, then ask which requirement makes one of them invalid; the explanation walks the trade-off and shows why each alternative fails. A timed set helps you get through 60 scenarios without second-guessing, and every explanation gives you a portable rule instead of a memorised letter.

Sample questions 7 of 64 shown

  1. Q1
    A regulated enterprise must ensure complete recoverability of VMware Cloud Foundation (VCF) management components based on the following scenarios:
    • Post upgrade failures
    • VCF instance total loss
    • Workload domain component loss
    The following conditions must be met:
    • Snapshots are restricted
    • Backups must survive regional failures
    Which three design decisions directly address the design trait of recoverability? (Choose three.)
    1. Replicate VCF Operations and VCF Operations for Networks image-based backups offsite.
    2. Use VCF-native prechecks and postchecks to validate rollback-supporting upgrade paths.
    3. Maintain cold-standby clones of NSX Manager appliances for rapid redeployment.
    4. Store the SDDC Manager, vCenter, NSX Manager and VCF Automation backups in an external replicated object store accessible from multiple failure domains.
    5. Use isolated backup repositories for each management component to guarantee version pinning.
    6. Depend on vSphere HA for management Virtual Machine (VM) recovery across clusters.
    Recoverability across the three stated scenarios maps to three distinct mechanisms. Post-upgrade failure is covered by VCF-native prechecks and postchecks, which validate that the target state is reachable and that a rollback path exists before remediation begins (B). Total loss of a VCF instance, combined with a "no snapshots" restriction and the requirement that backups survive a regional failure, requires image-based backups of the management components stored in an external, replicated object store reachable from multiple failure domains (D). Protecting the observability layer - VCF Operations and VCF Operations for Networks - by replicating their image-based backups offsite (A) preserves monitoring and alerting history and speeds rebuild. Cold-standby NSX Manager clones (C) are not a supported recovery mechanism and add management burden; per-component isolated repositories (E) defeat centralized, versioned recovery and cannot survive a regional loss; and vSphere HA (F) only restarts VMs inside surviving clusters and provides no protection against instance loss or upgrade rollback.
  2. Q2
    An architect is designing a VMware Cloud Foundation (VCF) Single Sign-On (SSO) architecture for a customer environment. The following requirements have been identified:
    • REQ01: Workloads must be supported across multiple physical datacenter locations (DC01 and DC02).
    • REQ02: Administrative access to the platform must use two-factor authentication.
    • REQ03: The design must minimize operational complexity when managing multiple VCF instances.
    Which two design decisions should be documented to meet these requirements? (Choose two.)
    1. Deploy VCF Identity Broker (VIDB) in the management domain of every VCF instance across all sites.
    2. Deploy VCF Identity Broker (VIDB) in the management domain of each VCF instance at DC02.
    3. Deploy VCF Identity Broker (VIDB) in the first VCF instance management domain at DC01.
    4. Configure additional VCF instances within the same region to use the VIDB from the first VCF instance at DC02.
    5. Configure additional VCF instances within the same private cloud to use the VIDB from the first VCF instance at DC01.
    VCF Identity Broker provides the fleet-level identity plane that brokers authentication between VCF instances and enterprise identity providers, including 2FA via SAML/OIDC federation to an external IdP. REQ03 (minimize operational complexity) means avoiding one identity provider instance per VCF instance, so a single VIDB is deployed in the management domain of the first VCF instance at the primary site, DC01 (C). The remaining VCF instances in the same private cloud are then configured as consumers of that VIDB (E), which keeps a single identity policy, a single 2FA configuration, and a single integration point for the IdP. Deploying VIDB in every instance (A) or in each DC02 instance (B) multiplies identity configuration and certificate/trust management. Option D places the authoritative VIDB at the secondary site while the first instance lives at DC01, which inverts the natural ordering and increases cross-site dependency for the primary instance.
  3. Q3
    An architect is designing for a VMware Cloud Foundation (VCF) workload domain to use vSAN, with the following requirements:
    • Must support large-scale analytics workloads with ingestion patterns that show burst activity.
    • Must ensure stable latency and highly sustained throughput even during cluster expansion and internal component balancing.
    Which two design decisions directly address the design trait of performance? (Choose two.)
    1. Ensure consistent hardware generations across nodes.
    2. Disable automatic rebalancing.
    3. Use vSAN ESA to attain Adaptive Write Path.
    4. Use compression-only mode for vSAN OSA.
    5. Use vSAN OSA with RAID-6 storage policies.
    The Express Storage Architecture (ESA) introduces the Adaptive Write Path, which separates the write path into a fast log-structured ingest tier and a capacity-optimized, full-stripe write tier. That design absorbs burst ingestion without inflating write amplification and keeps latency stable while the cluster rebalances or expands (C). Uniform hardware generations across nodes (A) matter because vSAN performance is gated by the slowest contributor: mixed CPU, memory, and NVMe generations produce uneven queue depths and skew component placement, so consistent generations keep the Adaptive Write Path balanced. Disabling automatic rebalancing (B) leaves the cluster in a skewed state after expansion and is not a performance control. Compression-only on OSA (D) and RAID-6 on OSA (E) are OSA-era capacity levers - RAID-6 in particular increases write amplification and rebuild cost - and neither addresses burst ingest or sustained throughput.
  4. Q4
    A customer plans to deploy VMware Cloud Foundation (VCF) across two availability zones (AZs) within the same region. The customer requires:
    • Failover of NSX North/South routing with minimal loss of forwarding
    • No loss of functionality due to host failures and Edge Node maintenance
    • No cross-availability-zone stretched clusters are allowed for compliance
    Which two design decisions correctly address the design trait of availability? (Choose two.)
    1. Use Edge Node anti-affinity rules to ensure routing services survive a single host failure.
    2. Deploy multiple Tier-0 gateways in Active/Active Equal-Cost Multi-Path (ECMP) mode with Edge Nodes distributed across fault domains.
    3. Configure all Tier-0 uplinks to share a single Bidirectional Forwarding Detection (BFD) profile to simplify failover signaling.
    4. Pin all Tier-0 gateway services to a single Edge Cluster to reduce control plane transitions.
    5. Run NSX Global Manager in Primary/Secondary mode across AZs to maintain policy continuity.
    Because stretched clusters are prohibited, availability must come from the network and compute layers rather than from a single stretched vSphere cluster. Active/Active Tier-0 gateways with ECMP, with Edge Nodes spread across the two AZs and across fault domains, keeps forwarding symmetric and removes the Edge node as a single point of failure; a surviving AZ continues to advertise and forward with minimal packet loss (B). Edge Node anti-affinity rules keep the Edge VM placement spread across hosts within an AZ so that a single host failure cannot take out the majority of an Edge cluster's routing capacity, which also covers the "no loss of functionality during Edge maintenance" case (A). Sharing one BFD profile across all uplinks (C) does not improve resilience and can cause correlated, flapping failover. Pinning services to a single Edge Cluster (D) creates exactly the single point of failure the design must avoid. NSX Global Manager Primary/Secondary (E) is a multi-site policy construct and, with no stretched clusters, does not deliver data-plane availability.
  5. Q5
    An architect is creating a design for a VMware Cloud Foundation (VCF) private cloud. The solution must provide:
    • Protection against a single rack failure scenario
    • Maximum useable disk space possible
    Constraints include:
    • vSAN ESA as principal storage
    • A single cluster of 16 ESX hosts
    • There are a total of four racks available
    Which three design choices should the architect recommend to meet the requirements? (Choose three.)
    1. Rack 16 ESX hosts in equal numbers evenly across four racks. All racks of hosts will be actively used.
    2. Implement a vSAN ESA Storage Policy that uses RAID-6.
    3. Implement a vSAN ESA Storage Policy that uses RAID-5.
    4. Configure each host as a vSAN Fault Domain.
    5. Group hosts within each rack into a vSAN Fault Domain.
    6. Rack 16 ESX hosts in equal numbers evenly across four racks. The fourth rack of hosts will be reserved as hot spares.
    Rack-level failure tolerance requires the fault domain boundary to match the physical rack boundary, so hosts inside each rack are grouped into one vSAN fault domain (E). Placing the 16 hosts evenly across all four racks with every rack active (A) gives four equal fault domains of four hosts each, so the loss of any single rack still leaves the cluster able to rebuild and serve. With fault domains already providing rack-level tolerance, RAID-5 erasure coding delivers the maximum usable capacity for the requirement: it consumes one parity element instead of RAID-6's two, and the additional failure it tolerates is already covered at the rack layer (C). RAID-6 (B) would satisfy the resilience requirement but consumes materially more capacity, contradicting "maximum usable disk space". Treating each host as its own fault domain (D) ignores rack topology, so a rack failure would still take out multiple fault domains. Reserving the fourth rack as hot spares (F) leaves only three racks serving production, so a single rack loss removes a third of the hosts and the spare rack does not participate in the fault domain layout.
  6. Q6
    An architect is designing a VMware Cloud Foundation (VCF) deployment for a customer transitioning to a hybrid cloud environment. The customer has provided the following requirements:
    • Networking Capacity: The existing data center infrastructure includes 10 GbE switches and NICs, but the customer plans to upgrade to 25 GbE within the next 12 months to support higher throughput for management traffic, overlays, and future workload domains.
    • Storage: The Management Domain must provide at least 15 TB of usable storage initially, with 20-30% headroom for organic growth in management components over the next 2 years. The customer prefers an all-flash configuration for performance.
    • Growth: The deployment will start with the Management Domain and expand to include up to 2 additional Workload Domains within 18 months, requiring scalability in compute and storage without major redesign.
    Based on the customer's requirements and VMware Cloud Foundation design guidelines, identify the key design decisions for the physical design of the VCF Management Domain. Which three actions support the requirements for hosts, networking, storage, and growth? (Choose three.)
    1. Allow heterogeneous host configurations to leverage existing hardware inventory, as long as overall cluster capacity meets minimum thresholds.
    2. Equip each ESX host with dual 10 GbE NICs initially, ensuring support for Jumbo Frames (MTU 9000) and planning for a non-disruptive upgrade to 25 GbE NICs to align with future bandwidth needs.
    3. Implement vSAN ESA with an all-flash configuration, providing a minimum of 20 TB raw capacity across the cluster accounting for FTT=1, 30% vSAN overhead, and 20% growth reserve.
    4. Configure each ESX host with at least 22 physical CPU cores and 512 GB of RAM to accommodate the base management components plus resources for managing additional Workload Domains.
    5. Deploy a minimum of 3 ESX hosts in the Management Domain to optimize initial hardware costs while maintaining basic redundancy.
    6. Implement vSAN OSA with a hybrid configuration, providing a minimum of 20 TB raw capacity across the cluster accounting for FTT=1, 30% vSAN overhead, and 20% growth reserve, using at least 2 disk groups per host for performance and fault tolerance.
    B honors the stated upgrade timeline: dual 10 GbE per host with jumbo frames end-to-end satisfies today's requirement while planning a non-disruptive move to 25 GbE, so no re-cabling or host rebuild is needed inside 12 months. C matches both the capacity and the medium: 15 TB usable with 20-30% headroom, plus FTT=1 and vSAN overhead and a growth reserve, lands near 20 TB raw, and ESA all-flash meets the performance preference with a single-tier architecture that scales by adding hosts. D ensures the management domain can host the full VCF management stack (SDDC Manager, vCenter, NSX Managers, VCF Operations, VIDB) and still absorb the additional management objects created by two more workload domains. Heterogeneous hosts (A) are unsupported for a VCF management domain cluster - a cluster must be uniform in CPU family and storage configuration, and vLCM single-image management requires like hardware. Three hosts (E) is the minimum for vSAN but leaves no maintenance headroom once the additional workload domains are onboarded, and it is below the recommended management domain sizing. A hybrid OSA design (F) contradicts the all-flash requirement and OSA's disk-group model adds lifecycle overhead compared with ESA.
  7. Q7
    As part of a design for a VMware Cloud Foundation (VCF) solution, an architect has documented the following dependencies and constraints:
    • CON001 - Internet access will not be permitted from anywhere within the VCF solution.
    • CON002 - The password must not be stored in plain text anywhere within the VCF solution.
    • DEP001 - The customer must make the required VCF binaries accessible to the VCF Installer appliance during the deployment phase.
    Which design decision should the architect include in the design for the download of the VCF binaries for initial installation?
    1. The Offline Bundle Transfer Utility will be used on the VCF Installer appliance.
    2. The VCF Installer appliance will be configured to connect to an online depot.
    3. The VCF Installer appliance will be configured to connect to an offline depot.
    4. The VCF Download Tool will be used on the VCF Installer appliance.
    CON001 removes any path to Broadcom's online software depot from inside the solution, so the VCF Installer must consume a local repository. The design decision is therefore to configure the installer against an offline depot that holds the required binaries and manifests, which also satisfies DEP001 - the customer is accountable for making those binaries reachable during deployment. The Offline Bundle Transfer Utility and the legacy VCF Download Tool (A, D) are run on a separate, Internet-connected staging system to populate that depot; running them on the installer itself would require the Internet access that CON001 forbids, and neither is the configuration the installer is designed around at deployment time. Pointing the installer at an online depot (B) violates CON001 outright. The offline-depot model also simplifies CON002, because the VCF environment never stores credentials for direct depot access.

Answer all 64 questions with instant feedback →

Reader reviews 43 ratings

4.4 ★★★★★★ Based on 43 ratings
  • These VMware exam dumps for 3V0-12.26 saved me weeks of study time. The questions cover every domain thoroughly.

    ★★★★★ Jessica W.

  • I bought the 3V0-12.26 question bank a week before my exam and passed with 90%+. The questions are that good.

    ★★★★★ Maria V.

  • Highly recommended for VMware certification. The 3V0-12.26 practice questions are worth every penny.

    ★★★★★★ Andrew M.

  • The most realistic 3V0-12.26 practice test I have ever used. The question style and difficulty match the real exam perfectly.

    ★★★★★★ Brian J.

  • Recommended this to my entire study group for 3V0-12.26. Four out of four passed. That says something.

    ★★★★★★ Audrey K.

  • Bought lifetime access for the 3V0-12.26 bank and it’s been great. Still use it to brush up even after passing the cert.

    ★★★★★★ Sophie L.

Log in to rate this exam and leave a review.

Frequently asked questions

What does the Architect blueprint leave out?

Sections 4 and 5. Install, Configure and Administrate and Troubleshoot and Optimize carry no testable objectives, so all 20 sit in Sections 1 to 3 — architectures and standards, VMware products, and plan and design. The result is a paper about design judgement, not click-paths or CLI work. Administrators often find it a different style of exam.

How are business and technical requirements separated?

A business requirement states the outcome in the organisation's terms, such as an uptime target; a technical requirement turns it into an infrastructure capability the design can implement and test. Mixing them undercuts traceability. The exam uses workshop scenarios where you classify each statement as a requirement, assumption, constraint or risk, then pick the design that satisfies it.

Which design quality wins when requirements collide?

Whichever the scenario's stated business objective makes dominant — AMPRS is not a fixed ranking. Availability, manageability, performance, recoverability and security all matter, but one decision usually improves one at the expense of another, and the item expects you to name that trade-off. Study the qualities as competing forces, not a checklist.

Which organisation model suits a multi-tenant design?

It depends on whether the consumer needs vSphere clusters directly or a Kubernetes-backed tenancy layer. VMApps organisations integrate straight into vSphere clusters; AllApps organisations add the Supervisor, namespaces and project-level governance for multi-tenant consumption. The exam frames this as a sizing and tenancy decision, so the reasoning behind the choice matters more than reciting services.

What does a design question look like in the bank?

Typically a requirement list followed by several plausible designs where only one survives every constraint — say a two-region recovery target plus an existing hardware contract. Our practice questions trace the requirement to each rejected option and name the AMPRS quality at stake. Working through that shape trains you to rule designs out instead of picking the first reasonable one.

Why read the explanation after choosing an architecture?

Because on a design paper the same reasoning recurs in disguise. An explanation showing why a stateful service rules out ECMP, or why a hardware contract is a constraint not a requirement, gives you a rule to carry forward. Scores only say how many you missed; the question bank's reasoning shows which part of your method needs work.