Cluster becomes unstable during duplicate vCenter adapter cleanup

Products

• VCF Operations
• VMware Cloud Foundation

Issue/Introduction

• After applying the resolution in KB Integrations UI becomes unresponsive or times out due to high number of vCenter adapter instances, the cluster may successfully come online but become unstable while attempting to clean up duplicate vCenter adapter instances.

• Unstable cluster pattern includes:

  • Cluster online, both nodes show online, objects/metrics shown

  • Cluster online, HA degraded, primary node waiting for analytics, replica node online with objects/metrics shown

  • Cluster online, HA degraded, primary node analytics starting, replica node online with objects/metrics shown

  • Cluster online, HA healthy, both nodes show online, objects/metrics shown

Environment

Aria Ops: 9.0.2.0

VCF: 9.0

Cause

The cluster requires additional compute resources to process the duplicate adapter removal, and standard data retention policies prevent leftover objects from being deleted immediately.

 

Resolution

1. Increase the CPU and memory for each analytic node to give the cluster sufficient resources to remove duplicate adapter instances while staying online.

   • Extra Large node configuration: 24 vCPU / 256 GB RAM

2. Objects associated with duplicate vCenter adapters remain in the environment due to standard retention policies. Adjust these settings to purge unwanted objects within 24 hours:

   1. Navigate to Operate > Global Settings > Data Retention.

   2. Set Deleted Objects: 24 hours

   3. Set Object History: 30 days