Snapshot and Backup Data Retention in MinIO with CSI Snapshot Data Movement

[ehemmerlin]

Hello,
We’re using Velero for on-prem backups in combination with Longhorn and MinIO, and we’ve enabled CSI Snapshot Data Movement.
From what we observe, when a backup is triggered, Velero requests Longhorn to create a snapshot, which is then uploaded to a MinIO bucket.
We’re currently seeing several terabytes of data in the Longhorn bucket within MinIO. Are these expected to be retained snapshots, or could they be remnants of deleted backups?
Additionally, the MinIO bucket used by Velero still contains around 250 GB of data, even though all Velero backups were deleted several days ago. Could this be stale or orphaned data that wasn’t properly cleaned up? Is there a recommended way to ensure that all associated data is removed when backups are deleted?
We’d appreciate any guidance on how Velero manages snapshot lifecycle and cleanup in this setup, especially when using CSI Snapshot Data Movement.
Thanks for your support!

[blackpiglet]

#9044
Some related issues were created for this topic.
Please refer to this issue and check whether maintenance jobs were running correctly.

Please also check the uploader type. We already deprecated the support for Restic. Please use Kopia if not yet.
And please try to use the latest Velero version. We just released v1.17.0.
There were some performance tuning changes included in the v1.17 and the v1.16 branches.