Lab: Scale VVF Cluster — Add and Remove ESXi Host with vSAN Data Migration
Objectives
- Add an ESXi host to an existing VVF cluster
- Claim the new host's devices into vSAN (ESA storage pool or OSA disk group)
- Remove an ESXi host with proper vSAN data evacuation
- Understand cluster scaling impact on vSAN fault domains and storage policies
- Import existing vSphere environment as VVF cluster (conceptual)
Prerequisites
VVF lab with vCenter and vSAN cluster operational. At least one unconfigured ESXi host available for addition.
Prior labs: vvf-support-01
Required skills:
- vCenter cluster management
- vSAN basics
- Networking concepts
Lab Environment
Holodeck vSphere Foundation 9.0 instance (New-HoloDeckInstance -VVF) with a three-host vSAN cluster; add a fourth nested host for the scale-out exercise with New-HoloDeckESXiNodes (Holodeck is not limited to four hosts). vSAN mode follows -vSANMode (OSA default, ESA where supported).
Credentials
| System | Username | Password |
|---|---|---|
| vCenter | administrator@vsphere.local | Holodeck documented default (Post Deployment table) |
| Nested ESX hosts | root | Holodeck documented default (Post Deployment table) |
Tasks
Task 1 Add ESXi Host to VVF Cluster
availabilityHost addition is a common scale-out operation. Understanding networking, vSAN, and DRS implications is essential for support troubleshooting.
Pre-check: Verify standalone host (esxi-04) has: same ESXi version as cluster, vSAN-compatible disks, VMkernel adapters configured for vSAN and vMotion traffic.
vCenter UI → Navigate to target cluster → Right-click → Add Hosts. Enter esxi-04 FQDN/IP and root credentials.
Review host summary: CPU sockets, memory, disk groups. Accept and add host to cluster.
Verify vSAN: Cluster → Configure → vSAN → Disk Management. ESA: claim the new host's NVMe devices into the single-tier storage pool. OSA: claim cache and capacity devices into a disk group (VCF 9.0.1 docs p.506).
Verify networking: Check that new host has correct vDS uplink assignments. Monitor → vSAN → Skyline Health → Network checks (host connectivity, MTU) pass for the new host.
Check rebalance: Monitor → vSAN → Resyncing Objects. vSAN rebalances automatically by default when disk usage variance passes the threshold (30%, VCF 9.0.1 docs p.1804).
Validation Gate
Check: 4th host added to cluster, vSAN healthy, DRS operational
Expected: Cluster shows 4 hosts. vSAN health all green. DRS rebalances VMs.
Common Errors
Task 2 Remove ESXi Host from VVF Cluster with vSAN Data Evacuation
recoverabilityHost removal requires careful vSAN data evacuation. Incorrect removal causes data loss. This is a critical support scenario.
Pre-check: Verify cluster can sustain host removal — check vSAN capacity after removal. Cluster → Monitor → vSAN → Capacity. Ensure remaining hosts have sufficient capacity.
Place target host in maintenance mode: Right-click esxi-04 → Maintenance Mode → Enter. Select vSAN data migration: 'Full data migration' (moves ALL vSAN objects off this host).
Monitor vSAN data migration: Cluster → Monitor → vSAN → Resyncing Objects (VCF 9.0.1 docs p.1759). Wait until no objects remain to resync.
After migration completes, verify: esxi-04 shows 0 vSAN components. Cluster → Monitor → vSAN → Health → all green.
Remove host: Right-click esxi-04 (in maintenance mode) → Remove from Inventory. Confirm removal.
Post-removal verification: Check storage policies still satisfied. Cluster → Configure → vSAN → Storage Policies → verify FTT=1 can still be met with 3 hosts.
Validation Gate
Check: Host removed cleanly, vSAN healthy with 3 hosts, no policy violations
Expected: Cluster back to 3 hosts. vSAN health green. All VM storage policies compliant.
Common Errors
Task 3 Import Existing vSphere Environment as VVF Cluster (Conceptual)
manageabilityObj 5.3 includes importing existing vSphere environments. This task covers the process conceptually since Holodeck has limited standalone vSphere environments.
Review convergence requirements (VCF 9.0.1 docs p.1344): converging to VVF 9.0.0 needs vCenter 9.0 and ESX 9.0; converging to VVF 9.0.1 or 9.0.2 accepts vCenter 8.0 Update 3 and ESX 8.0 Update 1 or later without upgrading.
Document the convergence process: (1) deploy the VCF Installer appliance, (2) run the vSphere Foundation convergence workflow with the existing vCenter and hosts (or a JSON specification), which deploys VCF Operations, (3) add and assign the vSphere Foundation license in VCF Operations. Existing storage is kept — vSAN is not required for convergence.
List pre-import checklist items: DNS/NTP configured, ESXi version compatible, vDS preferred over VSS, vSAN disk groups identifiable, network MTU consistent end to end (jumbo frames optional).
Document post-import tasks: Assign VVF licenses, configure vSAN storage policies, enable DRS/HA if not configured, integrate with VCF Operations for monitoring.
Note limitations: VSS-based clusters should be migrated to vDS before import. Mixed-version clusters (ESXi 7.x + 8.x) must be homogenized first.
Validation Gate
Check: Import process documented with pre-checks and post-tasks
Expected: Written checklist covering requirements, process, and post-import configuration.
Common Errors
Final Validation
Cluster scaling (add/remove host) completed with vSAN data integrity maintained; import process documented
✓ Host successfully added and integrated with vSAN → 4-node cluster operational with vSAN healthy
✓ Host removed with full data evacuation → 3-node cluster, all storage policies compliant
✓ Import process documented → Checklist covers pre-reqs, process, post-tasks
Cleanup / Restore
• Re-add esxi-04 if removed for subsequent labs
• Revert to 3-node snapshot if needed
Design Reflection (VCDX)
Cluster scaling decisions directly impact availability and performance. Architects must plan for N+1+1 capacity (maintenance + failure) and vSAN fault domain awareness during scaling operations.
Requirements
- Scale cluster without VM downtime
- Maintain vSAN data availability during all operations
Constraints
- Nested hosts add capacity slowly — budget physical host resources for each extra nested ESX host
- vSAN requires minimum 3 hosts for FTT=1
Assumptions
- New host has compatible hardware and ESXi version
- Sufficient network bandwidth for vSAN data migration
Risks
- Simultaneous host failure during data evacuation causes data loss
- vSAN rebalance during business hours impacts VM performance
Self-Assessment Discussion Prompts
- When is 'Ensure accessibility' appropriate vs 'Full data migration' for maintenance mode?
- How does adding a 4th host change vSAN fault domain calculations for FTT=1?
- What happens if you remove a host without entering maintenance mode first?
Extensions
Scale cluster to 4 hosts, then back to 3 — measure vSAN rebalance time
Test host removal with 'No data migration' option — observe policy violations
Create a stretched cluster scenario (conceptual) with 2 fault domains + witness
⚠ Known Pitfalls
References
- vSphere Cluster ManagementTier 1 — Official
- vSAN Host MaintenanceTier 1 — Official
- VVF Cluster Import GuideTier 1 — Official