Enabling AVI load balancing in VCF Automation 9.1 VPC fails with "incompatible load balancer" error

Products

• VCF Operations/Automation (formerly VMware Aria Suite)

Issue/Introduction

In VMware Cloud Foundation (VCF) Automation 9.1, attempts to enable AVI load balancing within a Virtual Private Cloud (VPC) fail.

• Users receive an "incompatible load balancer" error in the tenant portal when toggling "Enable load balancing" on a VPC.

• The issue occurs even when Service Engine Groups (SEG) are correctly allocated to the tenant.

• Load balancing can be enabled/disabled successfully from the NSX side, but the status is not correctly manageable via the VCF Automation interface.

Environment

VCF Automation 9.1.0

Cause

The ccs-nsx-eas service persists the Region's LoadBalancerType in a local cache based on an architectural assumption of immutability. This cache becomes stale if a Region is recreated with a new LoadBalancerType- for example, if a Region was initially associated with Supervisors using NSX_LB and subsequently recreated using Supervisors configured with AVI as the LoadBalancerType. Because the cache is not invalidated during this transition, the orchestration engine evaluates new enablement requests against the outdated metadata, resulting in a validation failure.

Resolution

This issue is resolved in a future release of VCF Automation. Subscribe to this article to receive email notifications when updates are published.

Prerequisites

• Ensure you have SSH access to the VCF Automation control-plane node. Follow Establish an SSH Connection to a VCF Services Runtime Instance if you don't already have it
• The below procedure will restart the ccs-nsx-eas service, therefor it is recommended to apply this during a maintenance window.

Workaround:

To resolve this issue, the ccs-nsx-eas cache must be flushed by restarting the service pods within the VMSP platform.

1. Log in to VCF Automation as vmware-system-user and elevate as root:

   ssh vmware-system-user@<vcfa-appliance-ip>
   sudo -i

2. Delete the ccs-nsx-eas-app pods to trigger an automatic restart. Execute the following command:

   kubectl delete pod -n prelude -l app=ccs-nsx-eas-app

3. Monitor the pods until they reach a Running state:

   kubectl get pods -n prelude -l app=ccs-nsx-eas-app

4. Once the service is operational, re-attempt the AVI load balancer enablement within the VPC settings.

Additional Information