NSX Backup Retention Settings in SDDC Manager Not Applied in VCF 9.0 and 9.1

Products

• VMware SDDC Manager / VCF Installer

Issue/Introduction

• Backup retention policies configured for NSX within the SDDC Manager UI (e.g., 2-day hourly, 14-day daily) are not being honored in VCF 9.0 or 9.1 environments.
• NSX backups continue to accumulate on the SFTP server or SDDC Manager NFS mount until storage is exhausted.
• Manual intervention or external scripts (such as nsx_backup_cleaner.py) are required to prune old backups despite UI settings.

Environment

• Software: VMware Cloud Foundation (VCF)
• Versions: 9.0.x, 9.1.x
• Component: SDDC Manager, NSX-T / NSX Manager
• Infrastructure: External SFTP Server or SDDC Manager NFS mount

Cause

In VMware Cloud Foundation (VCF) 9.0 and 9.1, NSX does not possess a native Retention Policy API that SDDC Manager can interface with.
While the SDDC Manager UI provides fields to define these retention values, the management plane cannot currently push these instructions to the NSX components.
This results in backups being created according to the schedule, but never being automatically deleted by the system based on the defined retention window.

Resolution

This behavior is a known limitation in the VCF 9.0 and 9.1 releases.

A unified backup and retention policy driven by Fleet LCM is planned for a future version of VMware Cloud Foundation, which will provide the necessary API integration to automate this process for NSX.

Workaround

To mitigate storage exhaustion without using external scripts, you can reduce the backup frequency (cadence) to daily instead of the default hourly setting. This reduces the total volume of data stored while waiting for a future release that includes automated retention.

To override the default hourly backup cadence:

1. Establish an SSH connection to the SDDC Manager VM as the vcf user and switch to root.
2. Update the domain manager properties:

   echo 'seconds.between.backup=86400' >> /etc/vmware/vcf/domainmanager/application-prod.properties

3. Update the operations manager properties:

   echo 'seconds.between.backup=86400' >> /etc/vmware/vcf/operationsmanager/application-prod.properties

4. Restart the services to apply changes:

   systemctl restart domainmanagersystemctl restart operationsmanager

Note: Setting the value to 86400 seconds changes the interval to 24 hours. Valid values range from 300 (5 minutes) to 86400 (24 hours).

Additional Information

Update backup cadence for NSX Manager from SDDC Manager
Backups can accumulate on the backup file server and consume a large amount of storage. You can run a script that comes with NSX to automatically delete old backups.