, ,

How to Create a VI Workload Domain in VCF 9.1 (VCF 9.1 Deployment Series, Part 8)

Step by step guide to creating a VI workload domain in VCF 9.1 using VCF Operations, from commissioning ESX hosts to validating a new vCenter, NSX Manager and vSAN datastore.

VCF 9.1 Deployment · Part 8 of 17
Status in a VCF 9.1 deployment: OPTIONAL. A VI workload domain adds an isolated pool of compute, storage and network for tenant or application workloads, kept separate from the management domain. It depends on a running management domain, VCF Operations, commissioned ESX hosts, an available network pool, and working DNS forward and reverse records plus NTP.
TL;DR
  • You run the whole workflow from VCF Operations, Inventory, Detailed View, Add Workload Domain, Create New.
  • Prerequisites: commissioned ESX hosts on the correct storage type, a network pool, a host overlay TEP VLAN with a static IP pool or DHCP, and resolvable vCenter and NSX Manager FQDNs.
  • A new domain deploys its own vCenter and can create or reuse an NSX Manager instance. NSX Manager appliance sizes are Medium, Large and Extra Large, in Standard single node or three node High Availability.
  • Storage choices are vSAN, NFS, VMFS on FC or vVols. vSAN ESA needs at least the failure tolerance minimum of hosts.
  • Key steps run down the left rail: General Information, vCenter, Cluster, NSX, Storage, Hosts, Distributed Switch, vSphere Supervisor, then Validation.
  • Validation runs a system check before you click Finish, then you watch progress under Fleet Management, Tasks.

This guide walks through creating a VI workload domain in VMware Cloud Foundation 9.1. You will commission ESX hosts, run the Create New workload domain wizard in VCF Operations, and end with a validated domain that carries its own vCenter, an NSX Manager instance and a vSAN, NFS or Fibre Channel datastore. A workload domain is optional because a VCF instance runs on the management domain alone, but most production designs add one or more workload domains so tenant workloads never share the same vCenter or NSX Manager as the management components.

Before you start, confirm that the management domain is healthy, that VCF Operations is reachable, and that you finished the VCF Operations setup in Part 7. You also need the network and naming groundwork from Part 3 on network, DNS and NTP, because the wizard fails fast if forward or reverse DNS records are missing for the new vCenter and NSX Manager.

Prerequisites

Meet every item below before you open the wizard. Missing DNS records and uncommissioned hosts are the two most common reasons a first attempt stops at validation.

RequirementDetailExample
Commissioned hostsESX hosts added to the free pool on the same storage type as the cluster4 hosts, vSAN HCI with ESA
Network poolA pool with free IPs that supports the chosen storage typewld01-pool
Host overlay VLANHOST TEP VLAN with a static IP pool or DHCP advertising addressesVLAN 1644
DNS recordsForward and reverse entries for the new vCenter and NSX Manager FQDNswld01-vc.corp.local
NIC speedAt least one physical NIC at 10 Gbps or faster per host2 x 25 GbE
Cluster imageA vSphere Lifecycle Manager image available for the default clusterESX 9.1 base image
Commission hostsGeneral infoand vCenterCluster andimageNSX ManagerStorage andhostsvDS andSupervisorValidation, thenclick Finish
Order of the workload domain wizard, from host commissioning to the final validation and Finish.

Step 1 commission the ESX hosts and check readiness

Adding ESX hosts to the free pool is called commissioning, and a host can only join a cluster of the same storage type it was commissioned with. Commission a vSAN ESA host for a vSAN ESA cluster, an NFS host for an NFS cluster, and so on. Prepare each host first with a resolvable FQDN, a regenerated self signed certificate, deleted disk partitions and a management gateway inside the management subnet.

  1. Log in to the vSphere Client for the management domain.
  2. Click Global Inventory Lists, then Hosts, then Unassigned Hosts.
  3. Click Commission Hosts.
  4. On the Checklist page, review each item, click Select All, then click Proceed.
  5. Choose Add new for a single host, or Import to upload a JSON template for many hosts.
  6. Enter the Host FQDN exactly as it appears in DNS, including capitalization.
  7. Set Storage Type to vSAN, then set vSAN Type to vSAN HCI and select the Express Storage Architecture option for a vSAN ESA host.
  8. Select the Network Pool Name from the drop down menu.
  9. Enter the root User Name and Password for the host, then click Add.
  10. Activate the Confirm All Finger Prints toggle after checking each server fingerprint.
  11. Click Validate All, correct any host marked Invalid, then click Next.
  12. Leave Skip failed hosts during commissioning enabled, then click Commission.

Step 2 open the workload domain wizard in VCF Operations

All of the deployment now lives in VCF Operations rather than a separate SDDC Manager screen. Start from the inventory and target the VCF instance that will own the new domain.

  1. Log in to VCF Operations.
  2. Select Inventory, then Detailed View.
  3. Expand VCF Instances and browse to the instance for the new domain.
  4. Click Add Workload Domain, then Create New.
  5. Review the prerequisites, click Select All, then click Proceed.

Step 3 enter the general information

Name the domain and decide whether it will host vSphere Supervisor. A name between 3 and 20 characters is required, and it is good practice to include a location code because generated object names derive from it.

  1. Enter a Workload Domain Name such as sfo01.
  2. Leave Enable vSphere Supervisor on to prepare the domain for Supervisor, or deactivate it for a plain compute domain.
  3. Enter the SSO Domain Name, for example vsphere.local, to keep the domain isolated from other domains.
  4. Select Auto generate my passwords for newly installed appliances, or set and confirm the SSO administrator password.
  5. Click Next.

Step 4 enter the vCenter details

A workload domain gets its own vCenter, which is deployed into the management domain. Supply an FQDN that already resolves in DNS.

  1. Enter the vCenter FQDN, for example wld01-vc.corp.local.
  2. Enter and confirm a vCenter Root Password of 15 to 20 characters, unless you chose to auto generate passwords.
  3. Click Next.

Step 5 name the cluster and select a cluster image

Define the first vSphere cluster in the domain, then pick the vSphere Lifecycle Manager image that every host in the cluster will run. Hosts that carry a different image get remediated to match during cluster creation.

  1. Enter a Cluster Name for the vSphere cluster.
  2. Enter a vSphere Zone Name if you enabled vSphere Supervisor.
  3. Click Next.
  4. Select a cluster image from the list.
  5. Click Next.

Step 6 configure the NSX Manager instance

Choose whether to stand up a new NSX Manager instance or reuse an existing one from another workload domain. For production, pick High Availability so a three node cluster carries the control plane. Appliance size follows scale, as shown below.

SettingOptionsNotes
Deployment SizeStandard or High AvailabilityStandard is one node, High Availability is three nodes
Appliance SizeMedium, Large, Extra LargeMatch size to host and workload count
ConnectivityCentralized or DistributedvSphere Supervisor requires Centralized
  1. Select an existing NSX Manager instance, or leave the create new option selected.
  2. Set Deployment Size to High Availability for production.
  3. Set Appliance Size to Medium, Large or Extra Large.
  4. Enter Appliance 1 FQDN, and Appliance 2 and 3 FQDNs if you chose High Availability.
  5. Enter the Appliance Cluster FQDN.
  6. Enter and confirm the Administrator Password and Auditor Password, unless you auto generate passwords.
  7. Select Centralized Connectivity for a Supervisor ready domain, or Distributed Connectivity to make the domain VPC ready with a VLAN ID, gateway CIDR and external IP block.
  8. Click Next.

Step 7 choose the principal storage type

Select one principal storage type for the primary cluster. Pick the same type your hosts were commissioned with. If you are unsure which vSAN variant to run, review the vSAN ESA and OSA storage design comparison before committing.

  1. Select vSAN to build a vSAN cluster, and select Enable vSAN ESA for an ESA cluster.
  2. Select NFS to use an NFS datastore instead.
  3. Select VMFS on FC for a Fibre Channel datastore.
  4. Select vVol only if you already added a VASA provider, noting that vVols is deprecated in 9.x.
  5. Click Next.

Step 8 set the vSAN storage options

This screen appears only when you chose vSAN, and the fields differ between OSA and ESA. For vSAN ESA the storage type and policy are fixed, and you mainly pick the cluster type.

  1. For vSAN OSA, set Failures to Tolerate, which sets the minimum host count.
  2. For vSAN OSA, select Space efficiency to enable deduplication and compression.
  3. For vSAN ESA, set vSAN Cluster Type to vSAN HCI for storage and compute, or vSAN Storage for storage only.
  4. Leave Storage Type as Local vSAN datastore, which cannot be changed.
  5. Leave Storage Policy as Auto policy management, which cannot be changed.
  6. Click Next.

Step 9 select the ESX hosts

Pick the hosts that will form the primary cluster. Only hosts commissioned for the matching storage type appear, so a vSAN ESA domain lists only vSAN ESA hosts.

  1. Select each host you want in the cluster from the list of available hosts.
  2. Confirm the count meets the minimum for your Failures to Tolerate choice.
  3. Click Next.

Step 10 configure the distributed switch

Choose how traffic maps onto vSphere Distributed Switches. The Default profile puts all traffic on a single switch, and the separation profiles split storage or NSX traffic onto their own switches and NICs.

  1. Select Default for a single switch that carries all traffic types.
  2. Select Storage Traffic Separation, NSX Traffic Separation, or both, if you want dedicated NICs per traffic type.
  3. Select Custom Switch Configuration to build multiple switches by hand, keeping Management, vMotion, vSAN and NSX each configured once.
  4. Click Next.

Step 11 configure vSphere Supervisor or skip it

This screen appears only when Enable vSphere Supervisor is on. Supply the Kubernetes networking values, or reuse the ESX management VMK settings to save time. If Supervisor is off, the wizard skips straight to review.

  1. Enter a Supervisor Name.
  2. Enter a Service CIDR that does not overlap any Supervisor routable network.
  3. Select Use ESXi Management VMK settings to reuse the management VLAN, mask, gateway and switch.
  4. Enter a Control Plane IP Range with at least five addresses if you did not reuse the management settings.
  5. Enter Workload DNS and Workload NTP server addresses.
  6. Click Next.

Step 12 validate and finish

VCF Operations runs a system check against every input on the Validation page. Fix any failed check, revalidate, then start the build.

  1. Review the summary of the workload domain settings.
  2. Wait for validation to complete and clear any failed test.
  3. Click Finish to start workload domain creation.
  4. Click Fleet Management, then Tasks, to monitor progress.
Storage decision path inside the wizard.
flowchart TD
  A[Start storage step] --> B{Which storage type}
  B -->|vSAN| C[Set vSAN ESA or OSA]
  B -->|NFS| D[Enter datastore, folder, server IP]
  B -->|VMFS on FC| E[Enter datastore name]
  C --> F[Select matching hosts]
  D --> F
  E --> F
  F --> G[Validate and Finish]

Verify the workload domain

After the task reaches a successful state, confirm the pieces are in place. Open Fleet Management, Tasks, and check that the create workload domain task shows a green successful status with no subtasks in error. In VCF Operations, Inventory, Detailed View, the new domain appears under the VCF instance with its vCenter, NSX Manager and cluster listed. Log in to the new vCenter and confirm the cluster is present, the hosts are connected, and the datastore is mounted and reporting capacity. For a vSAN domain, open the cluster, then Monitor, then vSAN, then Skyline Health, and confirm the checks are green. If you enabled vSphere Supervisor, the config status stays pending until you deploy an NSX Edge cluster, which is covered in the next part of this series.

Common errors and fixes

SymptomCauseFix
Validation fails on vCenter or NSX FQDNMissing forward or reverse DNS recordAdd both A and PTR records, confirm with nslookup, then revalidate
No hosts appear on the host selection screenHosts commissioned on a different storage type, powered off, or not reachable by SSHRecommission on the correct storage type and confirm the hosts are healthy
Host overlay TEP allocation failsNo static IP pool or DHCP on the HOST TEP VLANProvide a static IP pool or a DHCP scope on the overlay VLAN, then retry
vMotion network validation failsMAC address filtering blocks the temporary test VMkernelsRelax MAC filtering on the physical fabric during commissioning

Common questions

Is a workload domain mandatory in VCF 9.1
No. A VCF instance runs on the management domain alone. You add a workload domain when you want tenant workloads on a separate vCenter and NSX Manager.

How many hosts do I need for a vSAN ESA domain
The minimum depends on the failure tolerance you set, and a common starting point is four hosts for a single failure to tolerate with good rebuild headroom.

Can I reuse an existing NSX Manager instance
Yes. On the NSX screen you can select an existing NSX Manager from another workload domain instead of creating a new one.

What happens when vSphere Supervisor is enabled
The domain is prepared for Supervisor, and you finish activation later by deploying an NSX Edge cluster with an active standby tier 0 gateway.

Can I deploy a workload domain with the API instead
Yes. VCF Operations exposes a create domain API, and LACP based distributed switches must be configured through the SDDC Manager API.

About The Author


Discover more from Journal of Intelligent Infrastructure

Subscribe to get the latest posts sent to your email.

Leave a Reply

Your email address will not be published. Required fields are marked *

Architect’s Toolkit

About the Author

Dr. Pranay Jha is a Cloud and AI Consultant with 18+ years of experience in hybrid cloud, virtualization, and enterprise infrastructure transformation. He specializes in VMware technologies, multi-cloud strategy, and Generative AI solutions. He holds a PhD in Computer Applications with research focused on Cloud and AI, has published multiple research papers, and has been a VMware vExpert since 2016 and a VMUG Community Leader.

Discover more from Journal of Intelligent Infrastructure

Subscribe now to keep reading and get access to the full archive.

Continue reading