Academy/Holodeck Lab Setup & Operations/Holodeck Toolkit Installation and Content Staging (Holodeck 9.1)
This lab targets VCF 9.1.0.0

Holodeck Toolkit Installation and Content Staging (Holodeck 9.1)

VCF 9.1.0.0Beginneradminarchitectvcdx⏱ 90 min

Holodeck 9.1 deploys VCF 9.1.0.0, VVF 9.1.0.0 and VCF 5.2.4. Supported range: VCF 9.1.0.0 down to VCF 5.2 (9.1.0.0, 9.0.2.0, 9.0.1.0, 9.0.0.0, 5.2.4, 5.2.3, 5.2.2, 5.2.1, 5.2). Physical target host ESX 8.0 U3 or 9.0 (the 'Minimum ESX Version: 8.0 U3' line is the target-host floor, unchanged since Holodeck 9.0); nested hosts are built from the staged ESX ISO of the VCF version deployed. Release dates: VCF 9.1.0 GA 12 May 2026 (techdocs release notes; the 9.1 upgrade guide prints 11 May 2026). Holodeck 9.1.0: docs header 'May 2026'; Broadcom lifecycle portal GA date 17 Jun 2026 (portal export, 24 Sep 2026); GA announced on the VCF blog 1 Jul 2026. All three are first-party; the lifecycle portal is the release-management record. No GitHub releases are published for vmware/Holodeck — vmware.github.io/Holodeck/9.1/ is the sole source of truth. Exact toolkit build number and PowerShell module version to be captured during your own live 9.1 deployment.

What changed in 9.1:

Holodeck 9.1 deploys VCF 9.1.0.0, VVF 9.1.0.0 and VCF 5.2.4/5.2.3; full supported range VCF 9.1.0.0 down to 5.2; physical target host ESX 8.0 U3 or 9.0 (tested and supported; nested hosts run the ESX build of the chosen VCF release). SIZING (the big change): VCF 9.1 single site needs 48 CPU / 650 GB / 2.2 TB — exactly 2.0x the RAM and disk of VCF 9.0 (32 / 325 GB / 1.1 TB) and +50% CPU; dual site 96 / 1536 GB / 5 TB. VVF 9.1: 24 / 512 GB / 2 TB single, 48 / 1024 GB / 4 TB dual. HoloRouter appliance itself is now 6 vCPU / 12 GB / 75 GB (was 4/8 in 9.0.1). NEW: soft validation — pre-checks WARN rather than block below-spec deploys, so an under-spec build starts and then degrades instead of failing fast. STAGING: binaries go to /holodeck-runtime/bin/<version>/ where version is one of 9.1.0.0, 9.0.2.0, 9.0.1.0, 9.0.0.0, 5.2.4, 5.2.3, 5.2.2, 5.2.1, 5.2. Installer pairing: 9.1.0.0 installer for VCF 9.1.0.0, 9.0.2.0 installer for all 9.0.x, and VCF Installer 9.0.0.0 is NO LONGER SUPPORTED. DEPLOYMENT DRIVER: GitOps is an OVA deploy-time Extra Property (default ON) that auto-provisions GitLab and lets a full deployment run from the GitLab UI — there are zero GitOps cmdlets and it CANNOT be retrofitted later, so the driver choice is made once at OVA deploy. Other OVA properties: Enable SSH, Enable UI (Webtop). CONFIG: New-HoloDeckConfig now emits TWO config files (Site A + Site B) with Site A auto-loaded; Import-HoloDeckConfig now requires a MANDATORY -Site a|b, which breaks every 9.0.x invocation. PREREQS unchanged and still enforced: MTU 9000, trunk port group VLAN 4095 with Promiscuous/MAC Changes/Forged Transmits = Accept, NTP on all hosts, SSD/NVMe backing. AIR-GAP: ODA token via Broadcom KB 390098; VCF Download Tool at /root/vdt/bin replaces OBTU; the <version>-offline-depot-metadata.zip is MANDATORY alongside binaries or the depot silently fails.

Version note: 9.1 metadata docs-verified 2026-08-31 against vmware.github.io/Holodeck/9.1/ (release notes, cmd_reference, downloads, offline_depot, FAQ). Version targets, sizing, breaking changes and support matrix below are authoritative. STEP-LEVEL PROCEDURES are still cloned from the 9.0 twin and awaiting validation during your own live 9.1 deployment — treat individual step syntax as provisional until then.

Objectives

  • Validate the physical host meets Holodeck 9.1 requirements: target-host ESX 8.0 U3 or 9.0, the documented VCF 9.1 sizing (48 CPU / 650 GB / 2.2 TB single site), nested virtualization, trunk port group, NTP and DNS
  • Install and verify the Holodeck 9.1 toolkit and understand the two deployment drivers: PowerShell cmdlets vs GitOps (GitLab pipeline)
  • Download and stage VCF 9.1.0.0 content (and optionally VVF 9.1.0.0 / VCF 5.2.4) to the correct directory structure
  • Create and validate a Holodeck 9.1 deployment configuration, including the new dual-site (Site A + Site B) auto-generated config pair
  • Execute a pre-flight checklist and establish a baseline snapshot for subsequent 9.1 labs

Prerequisites

Clean physical ESXi host with no existing Holodeck instances. Holodeck 9.1 deploys VCF 9.1.0.0, VVF 9.1.0.0, or VCF 5.2.4 (supported range VCF 9.1.0.0 down to VCF 5.2); the physical target host must run ESX 8.0 U3 or 9.0 (nested hosts are built from the staged ESX ISO). Docs-published Holodeck 9.1 recommended sizing (single-site VCF): 48 logical cores, 650 GB RAM, 2.2 TB NVMe (dual-site: 96 cores / 1536 GB / 5 TB). Soft-validated — deployment proceeds with less but performance is impacted, nested virtualization enabled in BIOS, VLAN-trunk-capable port groups, NTP and DNS resolvable from the management network.

Required skills:

  • ESXi host installation and basic configuration
  • PowerShell fundamentals (pwsh on HoloRouter) — running cmdlets, reading output
  • JSON file editing and validation
  • Git/GitOps concepts (repositories, commits, pipelines) for the GitOps path
  • Network troubleshooting — PING, route commands, port group configuration
  • Snapshot management in vSphere
📸 Starting State: S0-91 — Fresh Host (Holodeck 9.1)

Lab Environment

Single physical ESX host (8.0 U3 or 9.0). This lab does NOT deploy the nested VCF stack: it deploys the HoloRouter 9.1 OVA (which carries the Holodeck PowerShell module and, with GitOps enabled, GitLab), stages the VCF 9.1 binaries on HoloRouter under /holodeck-runtime/bin/9.1.0.0/, and optionally rehearses with -ProvisionOnly (nested ESX hosts + VCF Installer only). A workstation is needed only for the Broadcom portal download and scp/SSH.

Mermaid source
graph TB
  WST[Workstation] -->|HTTPS| PBS[Broadcom Support Portal]
  WST -->|SSH / scp| HR91[HoloRouter 9.1 OVA on the physical ESX host]
  HR91 -->|/holodeck-runtime/bin/9.1.0.0/| BIN[ESX ISO + VCF Installer OVA]
  HR91 -->|Trunk port group VLAN 4095| PHESXI[Physical ESX host]
  HR91 -.->|GitOps enabled| GL[GitLab https://gitlab.vcf.lab]

IP Addressing

NetworkPurposeVLAN
10.1.0.0/20Default nested management supernet carried forward from the 9.0 baseline — verify 9.1 defaults against live toolkit9.0-era default VLAN range; VLAN range handling was fixed in 9.1 — verify

Credentials

SystemUsernamePassword
HoloRouter (SSH / console)rootSet in the OVA 'Password' property at deploy time
Physical ESXi HostrootSet during ESXi installation. Must be strong (length >= 8, mixed case, numbers, special chars).
Broadcom Support PortalYour Broadcom support account emailCorporate SSO or personal Broadcom account — required to download VCF content
HoloRouter 9.1 services (Vault :8200, Authentik :9443, Technitium :5380, Webtop :30000)Vault: root token; Authentik: akadmin (Getting Started) or admin (release notes — conflict, record what your build accepts); Technitium: admin; Webtop: admin; GitLab: rootHolodeck documented default (see the Holodeck 9.1 docs, Accessing Holodeck Environment)

Tasks

Task 1 Validate physical host hardware and nested virtualization

availability

Holodeck fails silently or with cryptic errors if the physical host does not meet the requirements. For 9.1 the documented floor is a target host on ESX 8.0 U3 or 9.0 and the recommended sizing is 48 CPU / 650 GB / 2.2 TB (single-site VCF 9.1); soft validation only warns below that. Validating nested virtualization up front prevents hours of wasted deployment attempts.

Step 1

Connect to the physical ESXi host via SSH (or vSphere Client): ssh root@<esxi-host-ip>

ESXi shell prompt (or vSphere Client is open)
Step 2

Verify the physical ESX version: esxcli system version get — it must be vSphere 8.0 U3 or 9.0 (tested and supported per the Holodeck 9.1 docs, Target Host Configuration).

Physical ESX 8.0 U3 or 9.0
The 8.0 U3 floor is the physical target host. Nested ESX hosts are not reusable templates — Holodeck builds them from the ESX ISO staged under /holodeck-runtime/bin for the VCF version you deploy (ESX 9.1 for VCF 9.1.0.0).
Step 3

Verify nested virtualization support (vHV): check vSphere Client > ESXi host > Configure > System > Processors for hardware-assisted nested paging (Intel EPT / AMD NPT), or the equivalent esxcli kernel setting.

vHV enabled / hardware-assisted nested paging supported
If vHV is disabled in BIOS, reboot the physical host and enable it. Common on new servers where it is disabled by default.
Step 4

Check total physical RAM: esxcli hardware memory get. Holodeck 9.1 docs (Pre-requisites, Physical Host requirements, read 1 Oct 2026): VCF 9.1 single site 48 CPU / 650 GB RAM / 2.2 TB SSD-NVMe; dual site 96 CPU / 1536 GB / 5 TB; VVF 9.1 single site 24 CPU / 512 GB / 2 TB. These are recommended values — Holodeck 9.1 soft validation warns below them but deploys. (VCF 9.0 was 32 CPU / 325 GB / 1.1 TB single site.)

Physical memory at or above 650 GB for single-site VCF 9.1 (1536 GB dual-site); if below, expect the 9.1 pre-check WARN and reduced performance
Step 5

Check CPU core count: esxcli hardware cpu global get. Docs-published 9.1 recommendation: 48 logical cores single-site (96 dual-site).

Logical cores >= 48 single-site (>= 96 dual-site)
Step 6

Check available datastore space: esxcli storage filesystem list. Docs-published 9.1 recommendation: 2.2 TB SSD/NVMe single-site (5 TB dual-site), including staged content.

At least one SSD/NVMe datastore with >= 2.2 TB free (>= 5 TB dual-site)
Step 7

Verify network configuration: confirm at least one port group supports VLAN trunking and the planned VLAN range does not conflict with production. VLAN range handling was fixed in Holodeck 9.1 — verify default VLAN range and any changed -VLANRangeStart semantics against live 9.1 toolkit.

VLAN-trunk-capable port group present; no VLAN conflicts with existing networks
Step 8

Verify NTP and DNS: esxcli system ntp get and confirm DNS resolution (e.g. nslookup broadcom.com). Note: on the 9.1 HoloRouter, DNS is served by Technitium DNS (:5380) rather than the 9.0-era dnsmasq pods — upstream-forwarder behavior must be re-validated on first live 9.1 deploy.

NTP synchronized; DNS resolving
Holodeck bring-up is extremely sensitive to clock skew and DNS failures — unresolved NTP/DNS issues surface later as obscure SSL certificate errors.

Validation Gate

Check: Checklist: physical ESX 8.0 U3 or 9.0: YES/NO, vHV enabled: YES/NO, RAM >= 650 GB (single-site VCF 9.1): YES/NO, cores >= 48: YES/NO, datastore >= 2.2 TB free SSD/NVMe: YES/NO, VLAN trunking configured: YES/NO, NTP synchronized: YES/NO, DNS resolving: YES/NO

Expected: All items YES. If any is NO, document the gap and remediate before proceeding.

Common Errors

Toolkit pre-check reports a version or resource warning
Cause: Physical ESX below 8.0 U3, or host resources below the recommended 9.1 sizing (soft validation warns and continues)
Fix: Upgrade the physical host to ESX 8.0 U3 or 9.0; add RAM/disk or accept the documented performance impact and record it.
Nested VMs are extremely slow or freeze frequently
Cause: vHV not enabled in BIOS — nested VMs fall back to software emulation
Fix: Reboot the physical host, enable Intel VT-x (EPT) / AMD-V (NPT) in BIOS, and re-verify from ESXi.

Task 2 Deploy the HoloRouter 9.1 OVA and choose a deployment driver

manageability

In Holodeck 9.1 the PowerShell module ships inside the HoloRouter OVA — there is no separate toolkit to install on a workstation. The deployment driver (PowerShell cmdlets or the GitOps GitLab pipeline) is fixed by the OVA's Enable GitOps property at deploy time and cannot be retrofitted, so the driver decision is made here, once.

Step 1

Download the HoloRouter 9.1 OVA from the Broadcom Support Portal (Holodeck 9.1.0 free download, linked from the Holodeck 9.1 docs Downloads page). No GitHub release packages exist for vmware/Holodeck.

HoloRouter 9.1 OVA downloaded
Step 2

Decide and document the deployment driver: (a) PowerShell cmdlets on HoloRouter (imperative, verbose output, fine-grained control), or (b) GitOps — GitLab pipeline pre-loaded on HoloRouter, deployments run from the GitLab UI. Record the rationale (repeatability, auditability, shared-host coordination). GitOps is an OVA property: it cannot be added later.

Driver decision documented with rationale
Step 3

Deploy the OVA to the physical ESX host (or its vCenter). On the Additional settings screen set the Router Network fields (Hostname, Management IP / CIDR / Gateway / DNS / NTP, DNS Domain, and Password, which sets the HoloRouter admin password), then under Extra Properties set Enable SSH, Enable UI (Webtop) and Enable GitOps (on by default) to match your driver decision (9.1 Getting Started). Size: 6 vCPU / 12 GB / 75 GB.

HoloRouter powered on with the chosen properties
Step 4

ssh root@<holorouter-ip>, run pwsh, and record the built-in Holodeck cmdlet inventory: Get-Command *HoloDeck* (plus Get-Module for the version; the 9.1 Getting Started page says everything is packaged into a PowerShell module called Holodeck, built into the HoloRouter OVA). Note the 9.1 change: New-HoloDeckConfig writes Site A and Site B config files, with Site A loaded.

Cmdlet inventory and module version recorded in the lab notebook
Step 5

GitOps path only: browse to https://gitlab.vcf.lab (user root) and confirm the pre-loaded Holodeck project and pipeline are present and idle.

GitLab project and pipeline visible; nothing run yet

Validation Gate

Check: Verify: (1) HoloRouter deployed with the intended Extra Properties, (2) SSH + pwsh session works and the Holodeck cmdlet inventory is recorded, (3) driver decision documented, (4) GitLab pipeline visible if GitOps was enabled

Expected: All checks pass. You are ready to stage binaries.

Common Errors

No Holodeck cmdlets in pwsh
Cause: The session is not on HoloRouter (for example a workstation PowerShell), or pwsh was not started
Fix: ssh root@<holorouter-ip>, start pwsh, and run Get-Command *HoloDeck* there — the module ships inside the OVA.
GitLab is missing
Cause: The OVA was deployed with Enable GitOps off
Fix: GitOps cannot be retrofitted: redeploy the OVA with Enable GitOps on, or continue with the PowerShell driver.

Task 3 Download and stage the VCF 9.1 binaries on HoloRouter

manageability

Holodeck reads the ESX ISO and VCF Installer OVA (Cloud Builder for 5.2.x) from a fixed per-version folder on HoloRouter: /holodeck-runtime/bin/<version>/. There is no content-path setting and no separate manifest to build; the discipline that remains is downloading the right pair and checking the hashes.

Step 1

Download the ESX 9.1 ISO and the VCF Installer 9.1.0.0 OVA from the Broadcom Support Portal (or from Firefox inside Webtop). Pairing: the 9.1.0.0 installer for VCF 9.1.0.0, the 9.0.2.0 installer for any 9.0.x target; VCF Installer 9.0.0.0 is no longer supported. For an air-gapped build the <version>-offline-depot-metadata.zip is also required.

ESX ISO and VCF Installer OVA downloaded
Do not interrupt downloads mid-transfer — a truncated image fails later in the build.
Step 2

Copy both files to HoloRouter: scp <ESX ISO> root@<holorouter-ip>:/holodeck-runtime/bin/9.1.0.0/ and scp <VCF Installer OVA> root@<holorouter-ip>:/holodeck-runtime/bin/9.1.0.0/

Both files copied
Step 3

Verify the SHA-256 of each file on HoloRouter (sha256sum) against the values on the Broadcom portal.

All hashes match; any mismatch = delete and re-download that file
Step 4

Confirm the folder content: ls -l /holodeck-runtime/bin/9.1.0.0/

ESX ISO and VCF Installer OVA listed with the expected sizes

Validation Gate

Check: Verify: (1) the ESX ISO and VCF Installer OVA for 9.1.0.0 are in /holodeck-runtime/bin/9.1.0.0/ on HoloRouter, (2) both SHA-256 values match the portal

Expected: Binaries staged and verified for VCF 9.1.0.0

Common Errors

The deployment cannot find the binaries
Cause: Files copied to the wrong folder (for example the physical ESX datastore or a workstation folder) or into the wrong version folder
Fix: Copy them to /holodeck-runtime/bin/<exact version>/ on HoloRouter (9.1.0.0 for VCF 9.1).
Cannot download content from the Broadcom portal
Cause: Entitlement/download-token issue, expired link, or a firewall blocking the CDN
Fix: Verify the account entitlement, regenerate the link, or use Firefox inside Webtop.

Task 4 Create and validate the Holodeck 9.1 deployment configuration

manageability

The deployment configuration is the source of truth for the instance. 9.1 changes the workflow: New-HoloDeckConfig now auto-generates Site A AND Site B configurations in one pass (dual-site by default). On shared hosts this is a double-edged sword — the auto-generated Site B must be reviewed for IP/VLAN collisions with any existing instances before it is ever used. This mirrors the VCF Planning & Preparation Workbook discipline.

Step 1

Generate the configuration with New-HoloDeckConfig. Confirmed 9.1 behavior: this now produces BOTH Site A and Site B configurations in a single pass. Syntax (Holodeck 9.1 docs): New-HoloDeckConfig -TargetHost <ESX or vCenter> -UserName <user> -Password <password> [-Description <text>]; Site A is auto-loaded into $config. Output layout: verify against the live 9.1 toolkit.

Site A and Site B configuration artifacts generated; ConfigID (or equivalent) recorded
Step 2

Record the deployment parameters you will pass to New-HoloDeckInstance in holodeck91-02 (target host and credentials are already set by New-HoloDeckConfig): -Version '9.1.0.0' (or VVF 9.1.0.0 / VCF 5.2.4 for those deployment types), -InstanceID, -CIDR (a /20; default 10.1.0.0/20), -VLANRangeStart, -DNSDomain, -vSANMode, -DepotType. The VCF version is a New-HoloDeckInstance parameter, not a config field. Do not hand-edit the config JSON — the Holodeck docs state that customers are not expected to interact with the config file directly or edit it. Valid -Version value strings for the 9.1 toolkit (DOCUMENTED: 9.1 release notes + FAQ, read 24 Sep 2026): '9.1.0.0', '9.0.2.0', '9.0.1.0', '9.0.0.0', '5.2.4', '5.2.3', '5.2.2', '5.2.1', '5.2' — confirm with Get-Help New-HoloDeckInstance on the live 9.1 toolkit (the generic cmd_reference page still shows the older six-value list; upstream lag). 9.1.1.0 is not a valid -Version value as of 24 Sep 2026; no toolkit release for 9.1.1 has surfaced (VCF 9.1.1.0 GA 3 Sep 2026 per techdocs; maintainer advice on issue #151, 6 Jul 2026: greenfield with 9.1.0.0 binaries rather than patch versions).

Deployment parameters recorded for holodeck91-02; config file left unedited
Step 3

Review management cluster resource sizing. The toolkit template (/holodeck-runtime/templates/config.json) sets nested host memory to 80 GB by default — the Holodeck release-notes workaround for NSX Edge deployment failures raises management and workload host memory from 80 GB to 128 GB. The vCPU default and any 9.1 or VNA-related sizing deltas: verify in that template on the live 9.1 toolkit. Keep total consumption <= 80% of physical capacity.

Resource sizing documented and within the 80% envelope
Step 4

Review networking: management CIDR, VLAN range, and external router IP. VLAN range handling was fixed in 9.1 and Site B port-group issues were resolved — verify the corrected defaults and confirm the auto-generated Site B ranges do not collide with anything existing on the physical network (critical on shared hosts).

Network settings reviewed for BOTH auto-generated sites; no collisions
Step 5

Set strong passwords for all credential fields and record them securely. Note: the 9.1 HoloRouter bundles HashiCorp Vault (:8200) for secrets management — whether deployment credentials flow through Vault: verify against live 9.1 toolkit.

All password fields set and documented securely
Step 6

Inspect the loaded configuration without editing it: run $config in the HoloRouter pwsh session and confirm the target host and the Site A / Site B networking values. There is no version or content-path field to check — the version is the -Version parameter of New-HoloDeckInstance, and binaries are found by version under /holodeck-runtime/bin/<version>/.

Configuration values display without error
Step 7

Note how validation works before you commit resources: Holodeck runs its built-in pre-checks (binaries present, target reachable, and the 9.1 resource checks with soft validation) automatically at the start of New-HoloDeckInstance; a separate validate-only or dry-run cmdlet did not surface in the 9.1 command reference. The nearest rehearsal is the optional -ProvisionOnly run in the next task. Expect stricter VCF Installer validation in 9.1 than in 9.0.

Pre-check behaviour understood; any soft-validation warnings will be recorded at deployment time

Validation Gate

Check: Checklist: (1) Site A + Site B configs generated and reviewed, (2) -Version value chosen for New-HoloDeckInstance (9.1.0.0, or VVF/5.2.4 target), (3) binaries present in /holodeck-runtime/bin/9.1.0.0/, (4) resources <= 80% of capacity, (5) no IP/VLAN collisions for either site, (6) passwords set, (7) built-in pre-checks noted (they run at the start of New-HoloDeckInstance)

Expected: Configuration ready for the deployment phase in holodeck91-02

Common Errors

Auto-generated Site B config collides with an existing instance on a shared host
Cause: 9.1 generates Site B automatically; on multi-instance shared hosts the defaults may overlap existing IP/VLAN allocations
Fix: Review and adjust the Site B CIDR/VLAN allocations before any deploy. Coordinate with other instance owners on the shared host.
Configuration edits made after a partial deployment do not take effect
Cause: 9.0-era behavior cached host specs on the first deployment run; whether 9.1 retains this behavior: verify against live 9.1 toolkit
Fix: Assume the safe pattern: destroy the instance, create a fresh config with New-HoloDeckConfig (one config per deployment), and redeploy from scratch.

Task 5 Execute pre-flight checklist and create baseline snapshot

recoverability

Production operations always run a pre-flight checklist before deploying infrastructure. A baseline now gives a known-good rollback point. New-HoloDeckInstance has no separate 'prepare' phase: without -ProvisionOnly it deploys the full stack, so this lab stops at a provision-only rehearsal or an idle GitLab pipeline.

Step 1

Document the environment baseline in deployment-baseline.txt: physical ESXi version, RAM, cores, datastore free space, NTP status, Holodeck module version (from HoloRouter pwsh), binary SHA-256 values, and the (redacted) configuration.

deployment-baseline.txt created with all items documented
Step 2

Final readiness check before deploying: confirm the config loads ($config, or Get-HoloDeckConfig then Import-HoloDeckConfig -ConfigId <id> -Site a), the target host is reachable from HoloRouter, and the binaries are in /holodeck-runtime/bin/9.1.0.0/. Holodeck's own pre-checks run automatically when New-HoloDeckInstance starts (including the optional -ProvisionOnly rehearsal); remediate anything they flag before the full run.

Readiness items confirmed; no failing item
Step 3

Create a baseline snapshot where possible, or document that the physical host cannot be snapshotted.

Baseline snapshot created OR documented as not applicable
Step 4

Review and sign off the pre-flight checklist (mirrors the VCF Planning & Preparation Workbook). Every item must be a confident YES.

Checklist completed with all YES responses
Step 5

Optional provision-only rehearsal (PowerShell path): New-HoloDeckInstance -Version 9.1.0.0 -InstanceID <id> -ManagementOnly -ProvisionOnly — deploys the nested ESX hosts and the VCF Installer only and writes the manual walkthrough scripts to /holodeck-runtime/specs/. Otherwise stop here: the full deployment is holodeck91-02/03. GitOps path: confirm GitLab (https://gitlab.vcf.lab) shows the pre-loaded Holodeck project and pipeline; do not run the pipeline yet.

Either the -ProvisionOnly run completes (nested ESX + VCF Installer up, specs written) or the GitLab project and pipeline are visible and untriggered; instance ID recorded
Step 6

Verify the state: Get-HoloDeckInstance -InstanceID <id> (ProvisionOnly path) or the GitLab project page (GitOps path). New-HoloDeckInstance has no separate 'prepare' phase — run without -ProvisionOnly it deploys the full stack.

Instance details show the provisioned nested hosts / VCF Installer for 9.1.0.0, or the pipeline is listed and idle
Step 7

Take a snapshot of HoloRouter (and of any nested hosts from the optional rehearsal): name it 'holodeck91-01-complete'.

Snapshot 'holodeck91-01-complete' created on all Holodeck VMs deployed so far

Validation Gate

Check: Checklist: (1) deployment-baseline.txt populated, (2) readiness items confirmed (config loads, target reachable, binaries staged), (3) optional -ProvisionOnly rehearsal completed or GitLab pipeline confirmed idle, (4) state verified, (5) snapshot 'holodeck91-01-complete' exists

Expected: All items complete. Lab holodeck91-01 is finished; ready for holodeck91-02 (management domain deployment).

Common Errors

The -ProvisionOnly rehearsal stops at HoloRouter or network configuration
Cause: The 9.1 router runs a different services stack (Technitium DNS, reverse proxy, Vault, Authentik) from 9.0
Fix: Capture the toolkit output and the HoloRouter service status and diagnose on the 9.1 router; do not apply 9.0-era dnsmasq or FRR fixes blindly.
The rehearsal fails with a target-host connectivity error
Cause: Physical ESX host unreachable from HoloRouter, wrong credentials, or the trunk port group not set (VLAN 4095, Promiscuous/MAC changes/Forged transmits = Accept)
Fix: Check reachability from HoloRouter, the credentials in the config, and the port group security settings.

Final Validation

HoloRouter 9.1 is deployed with a deliberately chosen driver (PowerShell vs GitOps), the VCF 9.1.0.0 binaries are staged in /holodeck-runtime/bin/9.1.0.0/ with verified hashes, the Site A + Site B configuration pair has been reviewed for collisions, and either a -ProvisionOnly rehearsal ran or the GitLab pipeline is confirmed idle, with a baseline snapshot in place.

✓ Holodeck toolkit version → Holodeck module version recorded from pwsh on HoloRouter

✓ Support matrix understood → Candidate can state: deploys VCF 9.1.0.0 / VVF 9.1.0.0 / VCF 5.2.4; supported range VCF 9.1.0.0 down to VCF 5.2; physical target host ESX 8.0 U3 or 9.0

✓ Physical host resources → Validated against the documented 9.1 sizing: 48 CPU / 650 GB / 2.2 TB single-site (96 / 1536 GB / 5 TB dual-site); shortfalls recorded as accepted performance impact

✓ Nested virtualization enabled → vHV enabled in BIOS and visible to ESXi

✓ VCF 9.1 content staged → ESX ISO and VCF Installer OVA in /holodeck-runtime/bin/9.1.0.0/ with matching SHA-256

✓ Configuration pair reviewed → Auto-generated Site A + Site B configs reviewed; no IP/VLAN collisions; resources <= 80% of capacity

✓ Provision-only rehearsal or GitLab pipeline readiness → -ProvisionOnly run completed for 9.1.0.0, or the pre-loaded GitLab pipeline is confirmed and idle

✓ Baseline snapshot → holodeck91-01-complete snapshot exists

Cleanup / Restore

Snapshot: holodeck91-01-complete

• Snapshot 'holodeck91-01-complete' has been created on the HoloRouter and any nested hosts provisioned in the optional rehearsal

• Document the instance ID (or GitOps pipeline reference) — needed for holodeck91-02

• Document the deployment-driver decision and configuration paths — referenced in holodeck91-02

• Store deployment-baseline.txt safely as a reference artifact

Design Reflection (VCDX)

A VCDX panelist would probe: 'How did you validate lab infrastructure readiness before deploying VCF 9.1?' — answer with the pre-flight checklist (versions including the physical-host ESX 8.0 U3 floor, vHV, resources, NTP/DNS, VLANs). New for 9.1: 'Why did you choose the GitOps driver over PowerShell (or vice versa)?' — be ready to argue repeatability (declarative config in Git, pipeline-driven) vs control/visibility (imperative cmdlets, verbose output), and to explain what the auto-generated dual-site config changes about your collision-avoidance process on shared infrastructure.

Requirements

  • Physical infrastructure capable of nested virtualization at the documented Holodeck 9.1 sizing (VCF 9.1 single site: 48 CPU, 650 GB RAM, 2.2 TB SSD/NVMe)
  • Holodeck 9.1 toolkit aligned with VCF 9.1.0.0 content (or VVF 9.1.0.0 / VCF 5.2.4 targets)
  • Physical target host on ESX 8.0 U3 or 9.0; nested ESX version follows the staged ESX ISO of the VCF release
  • Content staged and checksum-verified before any deployment attempt
  • Deliberate, documented choice of deployment driver (PowerShell vs GitOps)

Constraints

  • Supported deployment range is bounded: VCF 9.1.0.0 down to VCF 5.2 — older targets are out of scope on this toolkit
  • Single physical host = single fault domain; the entire lab is lost if the host fails
  • GitOps path resource overhead (GitLab appliance) is unquantified until live verification
  • 9.0-era community KB has zero verified 9.1 content — troubleshooting references must be re-baselined

Assumptions

  • The host meets or deliberately falls short of the documented 9.1 sizing; any shortfall is recorded as an accepted performance impact (soft validation)
  • Broadcom entitlement is active for 9.1 content downloads
  • The dual-site auto-generated Site B config can be safely adjusted before use on shared hosts
  • PowerShell cmdlet path remains fully supported alongside GitOps

Risks

  • Unverified 9.1 sizing causes resource exhaustion mid-deployment — MITIGATION: conservative <=80% allocation, monitor during deploy, capture actuals for the lab notebook
  • Auto-generated Site B config collides with existing shared-host instances — MITIGATION: review both site configs before any deploy; coordinate teardown windows
  • Assuming 9.0-era fixes (dnsmasq DNS edits, FRR restarts) apply to the 9.1 services stack — MITIGATION: re-diagnose from live logs; treat all 9.0 workarounds as unverified
  • Content/toolkit/manifest version skew — MITIGATION: align all three to 9.1 and enforce checksum validation

Self-Assessment Discussion Prompts

  1. The 9.1 support matrix spans VCF 9.1.0.0 down to VCF 5.2. What does that range enable for upgrade-path rehearsal, and what does the ESX 8.0 U3 / 9.0 target-host requirement rule out on older lab hardware?
  2. Argue both sides of the GitOps vs PowerShell driver decision for a solo lab vs a shared team lab. Which quality dimensions (manageability, recoverability) dominate each case?
  3. New-HoloDeckConfig now emits Site A and Site B in one pass. What new failure mode does this introduce on a shared physical host, and what process change mitigates it?
  4. The 9.1 HoloRouter replaces the dnsmasq-era DNS path with Technitium. Why should every 9.0-era DNS workaround be treated as unverified, and how would you re-baseline them?
  5. What would you capture during the first live 9.1 deployment to convert this stub lab into a fully verified manual?

References

Was this page useful?
Type to search. ↑ ↓ to move, Enter to open, Esc to close.