, ,

How to Add ESXi Hosts to a Cluster in VCF 9.1 (VCF 9.1 Day-2 Operations Step by Step Guide, Part 12)

Step-by-step how-to for adding ESX hosts to a VCF 9.1 cluster: commission the host in the vSphere Client, add it to the cluster, assign uplinks, and remediate the vLCM image.

VCF 9.1 Day-2 Operations · Part 12 of 20
Task type: Scenario. Run this when a cluster needs more compute or storage capacity and you have spare ESX hosts ready to join it. It depends on a healthy workload domain, a commissioned host that matches the cluster hardware, and a network pool that supports the cluster storage type.
Before you begin. Confirm the fleet is healthy and fully backed up before you change cluster membership. Work through the VCF 9.x pre-installation checklist first, then take a fresh SDDC Manager and vCenter backup so you can roll back if an add operation stalls.
Quick summary
  • Console: commission and add both run in the vSphere Client, with the SDDC Manager UI as a fallback.
  • Commission the new host into the Unassigned Hosts inventory before you add it anywhere.
  • Match the new host to the existing cluster hardware, storage type, and physical NIC layout to keep the cluster balanced.
  • Size the NSX host overlay TEP pool for the extra host, and confirm a network pool exists for the storage type.
  • Add hosts from the cluster Actions menu, assign uplinks on the cluster vDS, then finish and watch the task.
  • Run a vSphere Lifecycle Manager image compliance check and remediate the new host so it matches the cluster image.

Adding ESX hosts to a cluster gives a workload domain more compute and storage without building a new cluster. In VCF 9.1 you run this work in the vSphere Client. You first commission the host into the shared Unassigned Hosts inventory, then add it to the target cluster, where it inherits the cluster vDS, the NSX host overlay, and the vSphere Lifecycle Manager image. This guide covers a single cluster expansion. To add a whole new cluster to a domain, follow the neighbouring part on adding a cluster to a workload domain, and to swap out a host that has failed, follow the host replacement part, which is a remediation flow rather than an expansion.

Match the new host to the hosts already in the cluster before you start. A mismatch in CPU, memory, storage, or NIC layout does not block the add, but it leaves the cluster unbalanced and raises a warning you can dismiss. Plan the storage type as well, because a host commissioned for vSAN ESA can only join a vSAN ESA cluster, and a host commissioned for NFS can only join an NFS cluster. For background on how the domain was built and where these consoles come from, see the Day-2 operations overview and the VCF 9.1 Deployment guide.

Prerequisites

Confirm each item below before you commission or add a host. Most failures during expansion trace back to a missed prerequisite, usually DNS, certificates, or a network pool that does not match the storage type.

RequirementDetailWhere to confirm
Fleet health and backupCluster and domain healthy, with a fresh SDDC Manager and vCenter backup on handVCF Operations, SDDC Manager
Supported ESX buildHost runs an ESX version listed for VCF 9.1, with matching drivers and firmwareBroadcom release notes
DNS recordsForward and reverse lookup for the host FQDN, with the host name equal to the FQDNYour DNS server
Host certificateSelf-signed certificate regenerated from the host FQDN, or an external CA certificate if the fleet uses oneESX host
Physical NICsNICs of at least 10 Gbps matching the cluster uplink layout, with spare VMNICs for the vDS uplinksHost hardware
Storage matchSame principal storage type as the cluster, with all disk partitions clearedvSphere Client
Network poolA network pool that supports the cluster storage type, with spare management and vMotion addressesSDDC Manager network pools
NSX TEP poolStatic host overlay TEP pool has enough free addresses for the new hostNSX Manager
PermissionsAccount is a member of the vsphere.local sddcadmins groupVCF SSO

Step 1, Confirm the cluster is ready to expand

Start by confirming that the cluster actually needs capacity and that it is in a clean state. Fleet-wide health and capacity trend views live in VCF Operations, which a later part covers in detail.

  1. In VCF Operations, open the Capacity view for the workload domain and confirm the cluster is running short on CPU, memory, or storage headroom.
  2. In the vSphere Client, browse to the target vSphere cluster and note its exact name.
  3. On the cluster Summary tab, confirm there are no active alarms and that all existing hosts are connected.
  4. Note the cluster storage type, for example vSAN ESA or NFS, so you commission the new host to match.

Step 2, Commission the new host into inventory

Commissioning adds the host to the shared Unassigned Hosts inventory so it becomes available to clusters. In VCF 9.1 you run this from the vSphere Client under Global Inventory Lists.

  1. In the vSphere Client, click Global Inventory Lists, then Hosts, then Unassigned Hosts.
  2. Click Commission Hosts.
  3. On the Checklist page, review each prerequisite, then click Select All.
  4. Click Proceed.
  5. Select Add new to enter one host, or Import to upload a JSON file for several hosts at once.
  6. Type the Host FQDN, matching the DNS name including capitalization.
  7. Select the Storage Type that matches the cluster, for example vSAN with ESA enabled for a vSAN ESA cluster.
  8. Select the Network Pool Name that supports that storage type from the drop-down menu.
  9. Type the ESX host root User Name and Password, then click Add.
  10. Activate the Confirm All Finger Prints toggle, then click Validate All.
  11. Click Next, leave Skip failed hosts during commissioning enabled, then click Commission.

To commission several hosts at once, download the JSON template from the Import option and fill in one block per host. Valid storage type values are VSAN, VSAN_REMOTE, VSAN_ESA, VSAN_MAX, NFS, VMFS_FC, and VVOL.

[
  {
    "hostname": "esx07.wld01.example.local",
    "username": "root",
    "password": "REPLACE_WITH_ROOT_PASSWORD",
    "storageType": "VSAN_ESA",
    "networkPoolName": "wld01-network-pool"
  }
]

Step 3, Confirm the host is unassigned and active

Wait for commissioning to finish, then confirm the host is healthy and free before you attach it to a cluster.

  1. In the vSphere Client, open Global Inventory Lists, then Hosts, then Unassigned Hosts.
  2. Confirm the new host shows a configuration status of Active.
  3. Confirm the storage type and network pool columns match the target cluster.

Step 4, Add the host to the cluster

Adding the host runs from the cluster itself. It reuses the cluster vDS for management, vMotion, vSAN, and host overlay traffic, so you do not rebuild networking by hand.

  1. In the vSphere Client, browse to the target vSphere cluster in the inventory.
  2. Select the cluster, then click Actions, Add Hosts, Add Unassigned Hosts.
  3. Select the host or hosts to add, then click Next.

Step 5, Assign uplinks and review networking

Each cluster vDS carries two uplinks. Map free physical NICs on the new host to those uplinks so the host joins the same networks as its peers.

  1. Review the vDS and NSX networking shown for the cluster.
  2. Assign a free VMNIC on the new host to each existing uplink on the cluster vDS.
  3. Click Next.

Step 6, Finish and monitor the task

Review the summary, commit the add, and watch the task complete. You can add hosts to several clusters in parallel, so a long-running add on one cluster does not block work on another.

  1. Review the host, switch configuration, and license details.
  2. Click Finish.
  3. Open the Recent Tasks pane and watch the add hosts task run to completion.

Step 7, Check image compliance and remediate

A freshly added host can carry a different ESX build than the cluster. Bring it in line with the cluster vSphere Lifecycle Manager image so the whole cluster stays on one image.

  1. Select the cluster, then open the Updates tab.
  2. Click Check Compliance to compare the new host against the cluster image.
  3. If the host reports out of compliance, click Run Pre-Check to surface blockers such as a disconnected host or incompatible hardware.
  4. Click Remediate All to bring the host to the cluster image, then confirm the host returns to a compliant state.
CommissionhostVerifyunassignedAdd toclusterAssignuplinksFinishtaskCheckcompliance
Order of operations for a single cluster expansion in the vSphere Client.

How to verify the expansion

Run these checks after the add hosts task reports success. Confirm the host is a full cluster member, carries the right networking, and sits on the cluster image.

  1. In the vSphere Client, select the cluster and confirm the new host appears under Hosts as connected.
  2. On the host Configure tab, confirm the cluster vDS is attached and both uplinks show assigned VMNICs.
  3. For a vSAN cluster, open Monitor, then vSAN, then Skyline Health, and confirm disk groups from the new host are contributing capacity.
  4. On the cluster Updates tab, confirm the new host shows as compliant with the cluster image.
  5. Confirm no unbalanced cluster warning remains, or that any remaining warning is expected and dismissed.

Common errors and fixes

SymptomLikely causeFix
Unbalanced cluster warning after the addNew host CPU, memory, or storage does not match the existing hostsMatch the hardware where you can, or dismiss the warning if the imbalance is acceptable
Commission fails on validationDNS, certificate, or fingerprint mismatch on the hostCorrect DNS and the self-signed certificate, re-confirm the fingerprint, then run Validate All again
vMotion network validation failsMAC address filtering on the physical switch blocks the temporary VMkernel testDisable MAC filtering on the host ports, or allow the extra MAC addresses, then retry
Host lands but reports out of complianceHost image does not match the cluster vLCM imageOpen the Updates tab, run a compliance check, then remediate the host to the cluster image
Add host stalls on the overlay stepNSX host overlay TEP pool has no free addressesExpand the static TEP IP pool in NSX Manager, then resume the task
UnassignedHosts poolNew hostTarget clusterExisting hostExisting hostNew hostCluster vDS uplinksvSAN datastoreAdd to cluster
The new host moves from the shared pool into the cluster and joins the existing vDS and vSAN datastore.

Common questions

Do I still use SDDC Manager to add a host?
You can. In VCF 9.1 both commissioning and adding hosts run in the vSphere Client, but the SDDC Manager UI remains available as a fallback for the same tasks and for custom automation that calls its API.

Can the new host have a different NIC count than the others?
Yes. Different NIC enumeration is allowed, and you assign any free VMNICs to the existing uplinks during the add. Aim to match speed and layout so the cluster stays balanced.

Can I add hosts to more than one cluster at the same time?
Yes. Add and remove host tasks can run in parallel across different clusters, subject to the parallel task maximums in the VMware configuration maximums tool.

Does the host need the same storage type as the cluster?
Yes. A host commissioned for one storage type can only join a cluster of that type, so commission a vSAN ESA host for a vSAN ESA cluster and an NFS host for an NFS cluster.

References

VCF 9.1 Day-2 Operations · Part 12 of 20
« Previous: Part 11  |  Complete Guide  |  Next: Part 13 »

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