Learn how the VCF Automation 9.0.x to 9.1.0.0 upgrade works, from the five precheck validation categories to the six-stage side-by-side migration model orchestrated by Fleet Lifecycle.
Overview
This post is part of the VCF Support infrastructure modernization series. It focuses exclusively on upgrading from VMware Cloud Foundation (VCF) Automation 9.0.x to 9.1.0.0, setting aside older 8.18.x migration paths to examine the current supported upgrade path in detail.
When planning a move to VCF Automation 9.1.0.0, preparation and predictability are critical. A successful upgrade depends on VCF’s Fleet Lifecycle management capabilities, and the central gatekeeper to this entire process is the Precheck Phase. This post explains what occurs during these prechecks and maps out the execution stages the system proceeds through to complete the upgrade.
Video walkthrough: A video overview of the Fleet Lifecycle upgrade process is available for reference.
Environment Details
- Source version: VCF Automation 9.0.x
- Target version: VCF Automation 9.1.0.0
Prerequisites
Before initiating the upgrade, confirm the following:
- The source appliance is running a supported VCF Automation 9.0.x build.
- The Primary FQDN resolves correctly to an IP address.
- SSH (port 22) is accessible and authentication is configured.
- Kubernetes nodes are in a “Ready” state.
- Sufficient disk I/O capacity is available on Postgres pods. This is validated by a 30-second, 250 MB latency test during prechecks.
- DNS resolves correctly for all allocated node IPs.
- Allocated node IPs are not already in use on the network.
Understanding the Precheck Phase
Upgrading an automation platform requires orchestrating new configurations and migrating data at scale. The precheck phase serves as a non-destructive safety net, running a series of validations in a strict dependency order before any destructive operations are triggered.
The Five Precheck Categories
The Fleet Lifecycle evaluates the environment across five major categories to confirm readiness before the upgrade begins:
- Connectivity Checks: The Fleet Lifecycle verifies communication channels with the source environment. This includes confirming that the Primary FQDN resolves to an IP address, verifying that SSH (port 22) is open, ensuring cipher negotiation matches expectations, and validating SSH authentication.
- Source Validation: The system verifies that the upgrade proceeds from a supported starting point. It runs a Product Identification Check to confirm the source is a VCF Automation 9.0.x product, alongside a strict Version Validation Check.
- Health Checks: The system confirms that Kubernetes nodes are in a “Ready” state, verifies that application category services are healthy, checks for database split-brain conditions, validates the TLS certificate chain, and runs a 30-second, 250 MB disk I/O latency test on Postgres pods.
- Topology Checks: This phase confirms that the physical and logical infrastructure maps correctly to the upgrade configuration. It validates datastore consistency, checks that node IP pools and Platform VIPs fall within the target network CIDR block, confirms DNS resolution locally, and pings allocated node IPs to verify they are not already in use.
- Application Prechecks: The system verifies application-level compliance, specifically the FIPS Endpoint Compatibility check, which assesses the current FIPS mode status on the source VCF Automation 9.x environment.
Troubleshooting note: If a precheck fails, the 9.x precheck execution logs are written via SSH to the source appliance at the following path:
“/tmp/vcfa-upgrade/logs/upgrade-precheck<script> <timestamp>.log”
Executing the VCF Automation 9.1 Upgrade
VCF Automation does not perform an in-place software update. Instead, it uses a side-by-side migration approach. The Fleet Lifecycle deploys a new 9.1 cluster alongside the existing 9.0.x nodes, migrates data to the new cluster, and then powers down the legacy environment. This approach reduces risk and preserves data integrity throughout the transition.
The Complete Upgrade Workflow
The upgrade process proceeds through six coordinated stages:
Stage 1: Initialization and Staging
The system retrieves source component details, initializes required parameters, and validates the IP and FQDN specifications provided during configuration. It then stages the necessary VCF Automation packages and imports the source configuration.
Stage 2: Create Automation Cluster
The Fleet Lifecycle creates a dedicated vcf-automation folder on the management vCenter, deploys a runtime template (vcf-services-runtime-template-9.1.0.0), and provisions a temporary bootstrap virtual machine. The bootstrap VM clones and deploys three new VCF Automation 9.1 nodes, assigning them IP addresses from the provided CIDR block.
Stage 3: Discover Automation Cluster
Once the new virtual machines are online, an initialization process discovers the newly created VCF services runtime cluster and verifies it is ready to receive data.
Stage 4: Perform Component Upgrade
This is the core data migration phase. The orchestrator syncs legacy 9.0.x data to the new 9.1 nodes through the following sequence:
- A full backup of the source 9.0.x cluster is initiated.
- The Postgres database is prepared for migration.
- VCF component services on the source cluster are scaled down to prevent new writes.
- A final component backup is taken and data is transferred to the new cluster.
- Services and configurations are restored on the target 9.1 nodes.
- The old 9.0.x nodes are powered off.
- Component services are installed on the new cluster, the Postgres upgrade is finalized, and the target component version is patched to 9.1.0.0.
Stage 5: Install Backend Services
The system installs the required VCD Migrator Backend Service on the new 9.1 cluster.
Stage 6: Inventory Sync and Cleanup
The Fleet Lifecycle runs an inventory sync to register the new 9.1 infrastructure in the Fleet Lifecycle manager. It then unstages downloaded binaries and removes all temporary files.
Summary
The upgrade from VCF Automation 9.0.x to 9.1 is designed to be automated and low-risk. The precheck phase validates network topology, credentials, and database performance before any configuration changes occur. The side-by-side migration model ensures continuity by standing up the new environment before decommissioning the old one. Understanding these validation gates and execution stages enables administrators to plan and execute infrastructure modernization with confidence.
Acknowledgements
Special thanks to Arun Nukula, Iyaz Khan, and Owen Sheehy.
Discover more from VMware Cloud Foundation (VCF) Blog
Subscribe to get the latest posts sent to your email.
