SSP : FCD folder in vmfs volume are not getting reclaimed due to stale PVCs

Products

• VMware vDefend Firewall with Advanced Threat Prevention

Issue/Introduction

1. The number of container volumes is significantly higher than expected.This can be verified from:

    Cluster > Monitor > Cloud Native Storage > Container Volumes

2. The aggregate size of PVCs (in TB) appears lower.However, associated datastores show full capacity or higher-than-expected utilization.

3. PVCs are considered stale when:

• They are not associated with any Pod or Namespace
• They do not have any labels

4. These stale PVCs do not appear in the standard PVC listing:

• k get pvc -A

Environment

SSP 5.1

Cause

The issue occurs when underlying FCD (First Class Disk) objects remain in the VMFS datastore even after their corresponding Kubernetes PVCs are deleted. These orphaned volumes are not automatically reclaimed, leading to storage consumption discrepancies.

Resolution

 We have attached a script (cleanup_stale_fcd.sh) designed to identify inactive volumes. To use it, simply copy the file to the SSPI VM and execute it. 

• Copy the script cleanup_stale_fcd.sh to the SSPI VM.

• Execute the script:

  chmod +x cleanup_stale_fcd.sh
  ./cleanup_stale_fcd.sh --detect

  Note :The script is designed only to identify inactive volumes, not automatically delete them. Manual verification is strongly recommended before performing any deletion to avoid data loss.

  To accurately map and validate which PVCs/FCDs are safe to delete, it is recommended to engage the Broadcom Support team.

  They can assist with:

  • Correlating FCDs to Kubernetes objects
  • Validating orphaned volumes
  • Performing safe cleanup procedures