Duplicate IP/MAC Address Conflict (F3083) Post-VCF 9.1 Upgrade

Products

• VMware SDDC Manager / VCF Installer

Issue/Introduction

Following an upgrade of the Management domain to VCF 9.1, the Cisco ACI fabric reports fault F3083 (duplicate IP detected on multiple MACs) for an IP address (e.g., ##.##.###.##). The environment may experience intermittent network connectivity issues or traffic flapping, which persists until the legacy Fleet Management appliance is powered off.

Environment

VMware Cloud Foundation (VCF) 9.1, SDDC Manager, Cisco ACI Networking

Cause

In VCF 9.1, the Fleet Management appliance functionality is transitioned to a containerized service (vcf-runtime). During the upgrade, the new containerized service assumes the IP address previously assigned to the legacy Fleet Management VM. If the legacy VM is not automatically powered off or decommissioned by the installer process, both the old VM and the new container will advertise the same IP address simultaneously, causing a network-layer collision.

Resolution

1. Isolate the Conflict: Power down the legacy Fleet Management VM. If the ACI fault F3083 clears and network connectivity stabilizes, this confirms the legacy VM is the conflicting device.
2. Purge Stale Entries: Clear the MAC address and ARP tables on the ACI leaf switches associated with the impacted Endpoint Group (EPG) to remove lingering entries for the legacy VM’s MAC address.
3. Verify New Configuration: Confirm that the new vcf-runtime container is correctly advertising the expected IP address and that it is the sole source for that IP in the fabric.
4. Decommission: Once network stability is confirmed with the new containerized service active, the legacy Fleet Management VM should be decommissioned and removed from the inventory to prevent future recurrence.

Additional Information

• If the new containerized runtime fails to pull an IP from the expected IP pool, ensure that the pool configuration and DHCP/static assignments in SDDC Manager are correctly aligned with the VCF 9.1 deployment requirements.
• If the issue persists with the legacy VM powered off, verify for any other orphaned network interfaces or duplicate static IP assignments within the management domain.