Migration Gotchas
VCF 9 Migration Gotchas & Lessons Learned
Pre-Migration Pitfalls
Licensing Changes: VCF 9 eliminates traditional license keys. Products must be licensed via VCF Operations and vCenter. VMware1! passwords no longer work—admin passwords must meet new complexity requirements or onboarding fails.
Mandatory Components: VCF Operations (formerly Aria Operations), VCF Operations Fleet Management (Aria Suite Lifecycle), and VCF Automation (Aria Automation) are now mandatory. Plan to install these if not already present.
Version Prerequisites: Aria Automation and Aria Operations must be on version 8.18 before upgrading to VCF 9.0. Multiple sequential upgrades may be required from older versions.
Cluster Image Migration: Clusters using baselines must be migrated to cluster images before starting the upgrade.
VCF Automation NSX Mode: NSX clusters must be in Policy Mode. Manager Mode breaks the upgrade.
vSphere Version Requirement: Converting existing vSphere environments requires vSphere 8.0 Update 1 or later and removal of Enhanced Linked Mode (ELM).
During Migration Issues
vCenter Upgrade Failures: Upgrading from vCenter 8.0 Update 3+ to 9.0 may fail with “Failed to get sso server certificate for validation.” Reduced Downtime Upgrade (RDU) fails at switchover if file-based backup/restore was run before migration.
NSX Download Issues: download.vmware.com is no longer available. NSX release notifications and automatic binary downloads are unavailable. Pre-upgrade check bundles won’t download automatically.
VCF Operations Network Adapter Issues: Custom network adapters don’t start after upgrading VCF Operations. VCF Operations for Networks upgrade may fail with errors.
VCF Automation Disk Issues: Upgrade from Aria Automation 8.18 to VCF Automation 9.0 can fail due to disk dismount issues.
Orchestrator Authentication: If using vSphere authentication with special characters (spaces in domain names), the re-registration command fails during upgrade.
Post-Migration Surprises
No Aria Operations for Logs Migration Path: There is no import/upgrade path for Log Insight. Deploy a new VCF Operations Logs instance, redirect all log sources, and use VCF Operations Log Data Transfer to migrate up to 90 days of existing logs.
vSAN Adapter Warnings: After workload domain redeployment, the vCenter/vSAN adapter may enter Warning state.
VMware Cloud Director Incompatibility: VCD is not supported with VCF 9.0, with no official migration paths available.
Isolated Domain Limitations: If isolated workload domains haven’t been upgraded to 9.0, you cannot add hosts or clusters via vSphere UI.
Community Reported Issues
IP Conflicts: VCF Operations Collector deployment fails if SDDC Manager IP conflicts with another device. Requires full cleanup and redeployment.
NSX Image Validation: “NSX_T_MANAGER install image not found” error with VCF Installer 9.0.1—requires workaround before deployment starts.
vSAN ESA Disk Problems: Avoid using create_custom_vsan_esa_hcl_json.ps1 script—causes “No storage pool eligible disks found” errors during vCenter deployment. Use the vSAN ESA HCL mock VIB method instead.
NFS MTU Issues: Configure correct MTU (9000) on vSwitch and VMKernel port before deployment for NFS success.
Session Timeout During Deployment: 401 UNAUTHORIZED errors occur if session times out during deployment. Keep UI active and don’t log out during deployments.
Troubleshooting Tips
Primary Log Location: /var/log/vmware/vcf/domainmanager/domainmanager.log—tail in real-time during deployment.
Retry Mechanism: Use VCF Installer UI “Retry and Proceed with Deployment” button for recoverable errors like DNS timeouts or vSAN disk claiming failures.
Manual Intervention Points: Log into vCenter to manually claim unused disks for vSAN ESA datastore. Log into NSX Manager to resolve registration failures.
Certificate Regeneration: For 9.0 deployments, regenerate vCenter certificate with IP in SAN. For 8.x upgrades, regenerate both vCenter and NSX certificates, then re-validate cloud accounts.
No Easy Rollback: No single rollback for entire VCF. Assess failures at component level and rollback to pre-failure state or contact support.