Deploy a VI Workload Domain on Holodeck
Objectives
- Explain what a VCF 9.0 workload domain is: its own vCenter deployed into the management domain, an NSX Manager cluster that is new or shared with another workload domain (never the management domain's), one or more vSphere clusters of commissioned hosts, and principal storage
- Provision three nested ESX hosts for the workload domain with New-HoloDeckESXiNodes and commission them into the VCF inventory with a network pool and the matching principal storage type
- Create the workload domain from VCF Operations (Inventory > VCF Instance > Add Workload Domain > Create New) with a vSphere Lifecycle Manager cluster image, vSAN storage, and a new or shared NSX Manager
- Validate the workload domain in its own vCenter (vc-wld01-a) and NSX Manager (nsx-wld01-a) and in the VCF Operations inventory
- Deploy a test workload on an NSX segment in the workload domain and verify connectivity through the HoloRouter
- Perform lifecycle operations with the documented limits: add hosts and clusters (Update-HoloDeckInstance -AdditionalCluster), respect the three-host minimum of a standard vSAN cluster, and remove the workload domain
Prerequisites
holodeck-02 complete: management-only VCF 9.0.2 instance healthy (SDDC Manager, vc-mgmt-a, nsx-mgmt-a, ops-a). Physical host capacity for three more nested hosts plus a workload vCenter and NSX Manager (see the full-stack row of the requirement tables).
Prior labs: holodeck-02
Required skills:
- VCF 9.0 constructs (instance, management domain, workload domain)
- VCF Operations and vSphere Client navigation
- vSAN and vSphere Lifecycle Manager basics
- NSX segments and gateways
Lab Environment
Management domain (esx-01a..04a) as deployed in holodeck-02. The workload domain adds three nested hosts (esx-05a, esx-06a, esx-07a in the Holodeck naming), a workload vCenter vc-wld01-a deployed into the management domain, and a workload NSX Manager nsx-wld01-a (or a shared one with another workload domain). All on the Site A VLANs of 10.1.0.0/20 behind the HoloRouter.
The workload domain's vCenter and NSX Manager are new appliances deployed into the management domain; the management vCenter and management NSX Manager are never shared with workload domains (VCF 9.0.1 documentation, Table 2 Types of VCF Domains).
Mermaid source
graph TB
subgraph MGMT[Management domain]
SM[SDDC Manager] --- VCM[vc-mgmt-a] --- NSXM[nsx-mgmt-a]
VCW[vc-wld01-a: workload vCenter appliance runs here]
NSXW[nsx-wld01-a: workload NSX Manager runs here]
OPS[VCF Operations ops-a]
end
subgraph WLD[Workload domain wld01]
E5[esx-05a] --- E6[esx-06a] --- E7[esx-07a]
end
VCW -->|manages| WLD
NSXW -->|transport nodes| WLD
OPS -->|Add Workload Domain| SMIP Addressing
| Network | Purpose | VLAN |
|---|---|---|
10.1.0.0/20 | Default Holodeck Site A master CIDR, subdivided by New-HoloDeckNetworkConfig into the Site A subnets (management, vMotion, vSAN, host overlay, edge uplinks, and others) | VLANs 0 and 10-25 (Site A default) |
10.2.0.0/20 | Default Holodeck Site B master CIDR (dual-site only) | VLANs 40-58 (Site B default) |
<external management network> | HoloRouter external management IP (DHCP or static, set at OVA deployment); the only external IP a Holodeck environment needs | External port group |
Credentials
| System | Username | Password |
|---|---|---|
| HoloRouter (SSH and pwsh) | root | Set in the OVA Additional settings at deployment |
| HoloRouter webtop | (no authentication in 9.0.2) | http://<holorouter-ip>:30000; the docs warn not to expose it externally |
| VCF Installer | admin@local | Holodeck documented default (Post Deployment table) |
| VCF Operations | admin | Holodeck documented default (Post Deployment table) |
| Management vCenter | administrator@vsphere.local | Holodeck documented default (Post Deployment table) |
| SDDC Manager | administrator@vsphere.local | Holodeck documented default (Post Deployment table) |
| Management NSX | admin | Holodeck documented default (Post Deployment table) |
| Nested ESX hosts | root | Holodeck documented default (Post Deployment table) |
| Workload vCenter | administrator@vsphere.local (or administrator@<sso-domain> with an isolated SSO domain) | Set in the Add Workload Domain wizard, or auto-generated and retrieved via the SDDC Manager API |
| Workload NSX | admin | Set in the Add Workload Domain wizard |
Tasks
Task 1 Provision and commission the workload hosts
ManageabilityA workload domain needs commissioned hosts in the inventory before the wizard can use them. Holodeck provides the nested hosts; VCF commissions them with a network pool and a principal storage type that the cluster will have to match.
Check capacity against the full-stack row of the requirement tables (VCF 9.0 with Automation: 24-32 CPU, 325 GB RAM, 1.1 TB single site). The workload domain adds three nested hosts plus the vc-wld01-a and nsx-wld01-a appliances in the management domain.
On the HoloRouter in pwsh (config imported) create three nested hosts: New-HoloDeckESXiNodes -Nodes 3 -CPU <vCPU> -MemoryInGb <GB> -Site a -vSANMode OSA (match the vSAN mode of the management domain unless you deliberately want an ESA cluster). Holodeck builds them from the staged ESX ISO on the Site A networks with DNS records (esx-05a.., check with Get-HoloDeckAppNetwork -Site a).
New-HoloDeckESXiNodes -Nodes 3 -CPU 8 -MemoryInGb 64 -Site a -vSANMode OSAReview the documented commissioning criteria that matter in the nested lab: supported ESX version, management IP on a VMkernel adapter on a standard switch, forward and reverse DNS, host name equal to the FQDN, self-signed certificate regenerated on the FQDN, hardware health OK, no pre-existing partitions on the vSAN disks, a network pool that supports the storage type, and the gateway inside the management subnet. Holodeck prepares the hosts this way; verify one host's DNS and certificate.
Create or confirm a network pool for the workload domain (vMotion and vSAN ranges from Get-HoloDeckSubnet -Site a) and then commission the hosts: in the vSphere Client of the management vCenter, Global Inventory Lists > Hosts > Unassigned Hosts > Commission Hosts; review the checklist (Select All, Proceed); add the three FQDNs with storage type vSAN, vSAN type vSAN HCI, ESA unticked for OSA; select the network pool; validate and commission. (The SDDC Manager UI is the documented alternative.)
Record the commissioning outcome: host FQDNs, storage type, network pool, and the fact that a standard vSAN cluster needs a minimum of three hosts (VCF 9.0.1 documentation, vSAN cluster requirements), so three is the smallest workload cluster you can build.
Validation Gate
Check: Three nested hosts created with New-HoloDeckESXiNodes, DNS present, commissioned with vSAN HCI (OSA) and a network pool; three-host minimum recorded.
Expected: Hosts available in the Unassigned Hosts inventory for the workload domain.
Common Errors
Task 2 Create the workload domain in VCF Operations
ManageabilityIn VCF 9.0 the Add Workload Domain wizard lives in VCF Operations (Inventory > Detailed View > VCF Instance > Add Workload Domain > Create New). It deploys a new vCenter into the management domain and a new or shared NSX Manager, then builds the cluster from a vSphere Lifecycle Manager image.
Confirm the general requirements: a vSphere Lifecycle Manager cluster image is available for the default cluster (Managing vSphere Lifecycle Manager Images), the hosts are commissioned with the target principal storage type, a network pool with free IPs exists, a static IP pool or DHCP exists for the host overlay (TEP) VLAN, and the vCenter and NSX Manager FQDNs resolve (add them with Set-HoloDeckDNSConfig if the Holodeck records do not already cover vc-wld01-a and nsx-wld01-a).
In VCF Operations select Inventory > Detailed View, expand VCF Instances, browse to your instance and click Add Workload Domain > Create New. Review the prerequisites, Select All, Proceed.
General Information: Workload Domain Name wld01 (3-20 characters; include location in production); Enable vSphere Supervisor: deactivate for this lab (if left on, the wizard creates a Supervisor with NSX VPC networking, one control plane VM and one vSphere zone, and you must add an NSX Edge cluster afterwards); SSO Domain Name: leave empty to join the instance's SSO, or enter one (for example wld.sso) to create an isolated domain; Password Creation: set the SSO administrator password (8-20 characters, mixed classes, no dictionary words) or choose auto-generation (retrievable via the SDDC Manager API).
vCenter: enter the FQDN vc-wld01-a.site-a.vcf.lab (note the wizard's statement that the workload domain vCenter is deployed in the management domain) and the vCenter root password (15-20 characters, mixed classes). Cluster: name wld01-cl01. Select the cluster image (hosts are remediated to it if their add-ons differ).
NSX Manager: choose Create new (Deployment Size: Standard single node to save lab resources, or High-Availability three nodes as recommended; Appliance Size Medium) with FQDN nsx-wld01-a.site-a.vcf.lab (and two more FQDNs for HA), or Join existing to share an NSX Manager with another workload domain if one exists. The management domain's NSX Manager cannot be selected. Configure Network Connectivity: Centralized (configure edge clusters afterwards) or Distributed (VLAN, gateway CIDR, external and transit IP blocks for VPCs).
Storage: vSAN (leave Enable vSAN ESA unticked for OSA hosts). vSAN Storage details: Failures to Tolerate 1 (minimum hosts follow), space efficiency optional. Host selection: pick the three commissioned hosts. Review and Finish.
Monitor the task in VCF Operations (Fleet Management > Tasks) and in SDDC Manager (Inventory > Workload Domains > wld01 shows Activating). Expect vCenter deployment, NSX Manager deployment (if new), host preparation and cluster creation.
Record the design facts as you observed them: the workload vCenter and NSX Manager appliances appear as VMs in the management vCenter inventory; the workload cluster is managed by vc-wld01-a; SDDC Manager and VCF Operations manage both domains.
Validation Gate
Check: Wizard completed with vSphere Supervisor off, new vCenter, new or shared NSX Manager, vSAN OSA cluster of three hosts; wld01 Active.
Expected: A VI workload domain with its own vCenter and NSX Manager.
Common Errors
Task 3 Validate the workload domain
AvailabilityValidate in the places where a design review would look: the workload vCenter, the workload NSX Manager, SDDC Manager and VCF Operations, and the HoloRouter's view of the new appliances.
Open https://vc-wld01-a.site-a.vcf.lab (administrator@vsphere.local, or administrator@<sso-domain> if isolated). Confirm cluster wld01-cl01 with esx-05a, esx-06a, esx-07a, vSAN healthy, the vSphere Distributed Switch created by the wizard, and the cluster image compliance.
Open https://nsx-wld01-a.site-a.vcf.lab (admin). System > Fabric > Hosts shows the three workload hosts prepared and Up; System > Appliances shows the manager (cluster) Stable.
In the management vCenter, find the vc-wld01-a and nsx-wld01-a appliance VMs running on the management cluster (esx-01a..04a).
In SDDC Manager (Inventory > Workload Domains) and VCF Operations (Inventory) confirm wld01 Active with its vCenter, NSX Manager and cluster listed.
On the HoloRouter: Get-HoloDeckAppNetwork -Site a shows the new FQDNs; Get-HoloDeckDNSConfig -FQDN vc-wld01-a.site-a.vcf.lab resolves. Check Get-HoloDeckSubnet -Site a for the workload VLANs in use.
Record versions: workload vCenter and NSX builds against the VCF 9.0.2 BOM shown in VCF Operations.
Take the snapshot set holodeck-05-wld-created from the outer layer after Stop-HoloDeckInstance -InstanceID holo (then Start it again).
Validation Gate
Check: Workload vCenter and NSX healthy; appliances located in the management domain; SDDC Manager and VCF Operations consistent; HoloRouter DNS consistent; versions recorded; snapshot taken.
Expected: Validated workload domain with evidence.
Common Errors
Task 4 Deploy a test workload into the workload domain
PerformanceA domain is validated by a workload that uses its compute, storage and NSX networking end to end.
In nsx-wld01-a create an overlay segment web-seg (for example 172.16.10.0/24, gateway 172.16.10.1) connected to a Tier-1 gateway. If the domain has no edge cluster, keep the test east-west only, or deploy an NSX Edge cluster first (Configure Centralized Network Connectivity with Edge Clusters) and peer the Tier-0 with the HoloRouter (BGP values from Get-HoloDeckBGPConfig -Site a).
In vc-wld01-a deploy a small VM (a Linux ISO or an OVA from a content library) on wld01-cl01, vSAN datastore, connected to web-seg, with the vSAN default storage policy.
Test east-west: a second VM on the same or another segment can ping the first. Test north-south if an edge cluster and BGP exist: the VM reaches the HoloRouter and the external network; from the workstation (holodeck-04 routes) reach the VM's overlay IP if the Tier-0 advertises it to the HoloRouter.
Check vSAN placement and compliance for the VM (Monitor > vSAN > Physical disk placement) and note the FTT=1 object layout across three hosts (two data components and a witness component).
Apply a simple DFW rule in nsx-wld01-a (Security > Distributed Firewall) allowing only SSH/HTTP to the VM from the management subnet, with logging; verify the default rule is still Allow and discuss when you would change it to Drop.
Record resource consumption of the workload domain on the physical host (CPU, memory, datastore) to see the cost of the domain itself versus the workload.
Write the validation entry: what the test proved (compute, vSAN, NSX overlay, DFW) and what it did not (performance, HCL, physical NIC redundancy).
Validation Gate
Check: Segment, VM, connectivity, vSAN placement, DFW rule and resource figures recorded.
Expected: Workload domain proven end to end in the lab.
Common Errors
Task 5 Lifecycle operations: add a cluster, respect vSAN minimums, remove the domain
ManageabilityScale operations have hard limits in VCF: a standard vSAN cluster needs three hosts; a two-host vSAN cluster is only valid with an external witness host. Holodeck adds a documented Day 2 cmdlet for clusters, and removal is a planned, ordered operation.
Add a cluster the Holodeck way: Update-HoloDeckInstance -Site a -AdditionalCluster -VIDomain Workload deploys a 3-node vSphere cluster in the workload domain (separate parameter set; cannot be combined with other Day 2 switches). Alternatively add commissioned hosts to the existing cluster from the vSphere Client (cluster > Actions > Add Hosts > Add Unassigned Hosts).
Update-HoloDeckInstance -Site a -AdditionalCluster -VIDomain WorkloadScale-in rules: write down that a standard vSAN cluster must contain a minimum of three ESX hosts (VCF 9.0.1 documentation), that FTT=1 RAID-1 needs three fault domains (two replicas plus a witness component), and that a two-host vSAN cluster exists only with an external witness host. Therefore do not scale wld01-cl01 below three hosts; if you added a fourth host, remove it to return to three (host in maintenance mode with full data migration, then remove from cluster and decommission to Unassigned Hosts).
Capacity reasoning: losing one host of three removes one third of the cluster's capacity (not half) and leaves the cluster unable to rebuild FTT=1 until the host returns; record how you would size N+1 in production.
Decommission: after removing a host from a cluster, decommission it (vSphere Client > Global Inventory Lists > Hosts > Unassigned Hosts > Decommission Selected Hosts). A host that was in a VCF domain must be re-imaged before it is commissioned again; a host that was never in a domain does not.
Remove the test workload (VM, segment, DFW rule), delete any NSX edge cluster on wld01 first, then delete the workload domain from VCF Operations (Inventory > Detailed View > wld01 > Actions > Delete Domain) or SDDC Manager. Deletion is irreversible: the domain's clusters, datastores, workload vCenter and NSX Manager (if not shared) are removed, and the hosts return to the free pool in a 'need cleanup' state (VCF 9.0.1 docs p.1529-1530). Network pools are kept.
Decide whether to keep the three nested hosts for later labs or remove them; record the physical capacity returned.
Take the snapshot set holodeck-05-complete (Stop-HoloDeckInstance first, Start after).
Write the lifecycle entry: operations performed, documented limits applied, time taken for domain creation and deletion.
Validation Gate
Check: Cluster added via Update-HoloDeckInstance or hosts added; cluster kept at the three-host minimum; decommission rules applied; domain removed cleanly; snapshot taken; entry written.
Expected: Lifecycle operations executed within the documented limits.
Common Errors
Final Validation
A VI workload domain is created on Holodeck 9.0.2 the documented way: nested hosts provisioned and commissioned, the Add Workload Domain wizard in VCF Operations deploying a dedicated vCenter and a new or shared NSX Manager, validation in every console, a test workload on an NSX segment, and lifecycle operations within the vSAN minimums, ending with a clean removal.
✓ Hosts → Three nested hosts from New-HoloDeckESXiNodes, commissioned vSAN HCI with a network pool
✓ Domain → wld01 Active with vc-wld01-a and nsx-wld01-a (new or shared), vSAN OSA cluster of three hosts, Supervisor off
✓ Validation → Workload vCenter and NSX healthy; appliances in the management domain; consoles consistent; versions per BOM
✓ Workload → VM on an NSX segment with east-west (and north-south if edge) connectivity; vSAN FTT=1 placement; DFW rule
✓ Lifecycle → Cluster or hosts added; three-host minimum kept; decommission rules; domain removed; snapshot
Cleanup / Restore
Snapshot: holodeck-05-complete
• Remove the test VM, segment and DFW rule before deleting the domain
• Decommission the three 'need cleanup' hosts; to reuse them for holodeck-06/07, redeploy (re-image) the nested hosts and commission them again
• Record physical capacity before and after
Design Reflection (VCDX)
The workload domain is where VCF's separation of concerns becomes visible: management components in the management domain, each workload domain with its own vCenter (hosted in the management domain) and an NSX Manager that is new or shared with another workload domain, never the management one. Panels will ask why that rule exists (blast radius, lifecycle independence, scale), how you decide between a new and a shared NSX Manager, why Supervisor needs an edge cluster with centralized connectivity, and what the vSAN three-host minimum and FTT maths do to your sizing. Use the lab to practise the wizard's decisions, not to size production.
Requirements
- Workload domain with its own vCenter deployed into the management domain
- NSX Manager new or shared with another workload domain, never the management domain's
- vSAN OSA cluster of at least three hosts with FTT=1
- Hosts commissioned with the storage type of the cluster and a network pool
Constraints
- Standard vSAN cluster minimum three hosts; two-host clusters need an external witness
- vSphere Lifecycle Manager image required for the default cluster
- Supervisor with centralized connectivity needs an NSX Edge cluster afterwards
- Password policies (SSO 8-20, vCenter root 15-20 characters)
Assumptions
- Physical capacity for three more hosts and two appliances
- holodeck-02 management domain healthy
- DNS for the new FQDNs on the HoloRouter
- TEP IP pool or DHCP for the host overlay VLAN
Risks
- Sharing the management NSX — Impact: not possible by design; a design that assumes it fails — Mitigation: new or shared-with-workload NSX only
- Scaling below three hosts — Impact: FTT=1 cannot be met — Mitigation: three-host minimum rule
- Supervisor left on — Impact: activation stuck without edges — Mitigation: decide Supervisor and connectivity before the wizard
- Capacity exhaustion — Impact: appliance crashes — Mitigation: full-stack sizing row applied
Self-Assessment Discussion Prompts
- Why does each workload domain get its own vCenter, and why is it deployed into the management domain rather than into its own cluster?
- When would you share an NSX Manager between workload domains, and why can the management NSX Manager never be shared?
- What does the vSAN three-host minimum do to your smallest viable workload domain, and when does a two-node-plus-witness design make sense?
- Which wizard decisions (SSO domain, Supervisor, connectivity model, NSX size) are design decisions you must defend?
- How does the Holodeck full-stack deployment differ from the manual wizard route, and what did the manual route teach you?
- How would you present the capacity cost of the domain itself (vCenter, NSX) versus the workloads it hosts?
Extensions
Full-stack deployment comparison
Deploy a second instance with New-HoloDeckInstance without -ManagementOnly (-WorkloadDomainType IsolatedSSO where supported, -NsxEdgeClusterWkldDomain) and compare the automatically created workload domain with your manual one.
harderShared NSX Manager
Create a second workload domain that joins the existing nsx-wld01-a and document the shared-NSX implications for lifecycle and failure domains.
harderSupervisor on the workload domain
Deploy an NSX Edge cluster with an active-standby Tier-0 peered to the HoloRouter, then enable vSphere Supervisor with centralized connectivity (or use Update-HoloDeckInstance -DeploySupervisor where available in your Holodeck version).
hardervSAN ESA workload cluster
Commission three hosts for vSAN HCI with ESA and build an ESA workload domain; compare with the FAQ guidance on nested ESA.
sameIsolated SSO domain
Create the domain with its own SSO domain and explore the identity implications with VCF SSO (vcp-admin-04).
same⚠ Known Pitfalls
References
- Holodeck 9.0.2 documentation (index): Environment Overview, Holorouter Overview, Concepts, Pre-requisites, Deployment, During Deployment, Post Deployment, Day 2 Operations, How ToTier 1 — Official
Primary source for every HoloRouter and toolkit step in this lab. Fetched 1 Oct 2026. - Holodeck 9.0.2 command referenceTier 1 — Official
New-HoloDeckInstance, Update-HoloDeckInstance, Get-HoloDeckInstance, New/Get/Import-HoloDeckConfig, Reset-HoloDeckState, New-HoloDeckNetworkConfig, Set-/Reset-HoloRouter, Get-HoloDeckSubnet/AppNetwork/BGPConfig/DNSConfig/OverlaySubnet/AppIpPools, Get-HolodeckServiceIPPools. - Holodeck 9.0.2 FAQ and troubleshootingTier 1 — Official
Pre-check errors (binary not found, no trunk port group, NTP not running), ESX build SSH wait, memory tiering, nested vSAN ESA. - Holodeck 9.0.2 downloads and release notes (known issues)Tier 1 — Official
HoloRouter OVA and Offline Depot OVA; release notes at /release_notes/ carry the Known Issues per release. - Holodeck GitHub repository: issues and discussionsTier 2 — Supporting source
Where to report bugs; community troubleshooting. The releases page carries no binaries (OVA downloads are on the Broadcom Support Portal). - VMware Cloud Foundation 9.0 documentation: Managing VCF Domains; Create a New Workload Domain Using VCF Operations; Commission ESX Hosts; vSAN cluster requirements; vSAN stretched clustersTier 1 — Official
VCF 9.0.1 PDF pp.467-468 (constructs), 1497-1505 (commission hosts, create workload domain wizard), 1663 (vSAN minimum three hosts), 1692-1695 and 1833 (stretched clusters).