Introduction
In a changing IT world where applications move to be containerized, it is important to consider what needs to be restored from backup in case of disaster or accidental delete in a Kubernetes/OpenShift environment.
For this purpose, GBS has investigated how to best design and implement a backup for critical components in the Kubernetes/OpenShift environment and which open-source tools (license free) are already available for it.

Velero
We have found Velero to suit our needs. Velero is a well-known Open-Source tool for (Disaster) recovery, Data Protection and Data Migration in a Kubernetes (K8s) environment. It can be used in a cloud environment, (AWS cluster) or on-premises (OpenShift).
It also covers AWS and GCP as a storage provider for the backup and has scheduling functionality.

For a backup architecture in a Kubernetes environment, we depart from our regular centralized backup server. Instead, the Velero backup will be installed on K8s system itself. However, we can still use our (existing) GBS backup server as “offsite” Object Store for the backup.
Our backup strategy is to store backups off-site at cloud storage or a remote backup server, not in the same environment as the original K8s cluster, to still have the backups available in case of a site disaster.
Velero uses an AWS- S3 compatible object store to store its backups. Therefore, we can use AWS cloud storage or a local S3 Object Store.
For local object storage, we have integrated a MinIO Object Store that can be deployed on a centralized (existing) GBS backup server.
In this case, the data will be stored in a subdirectory on the regular /backup partition.