Protecting OpenShift ImageStreams

Image streams offer a method for continuously creating and updating container images. Whenever improvements are made to an image, tags can be used to assign new version numbers and monitor changes. Read the official documentation to learn more about OpenShift ImageStreams.

Backing up an ImageStream

Since Veeam Kasten can discover the ImageStreams present in a namespace, they can be easily backed up. To protect ImageStreams in a namespace, create and run a policy and specify a destination location. The local images referenced in the backed-up ImageStream will then be stored to this location. After a successful backup, the exported images will appear in the restore point as a Kanister artifact.

Note

If OpenShift image pruning has removed historical image digests from the internal registry, Veeam Kasten will automatically skip the missing images during backup and proceed with the remaining accessible images.

Restoring the Image Streams

The process of restoring backed-up ImageStreams is similar to restoring any other application using the Veeam Kasten dashboard. Restoring Applications can be followed for the detailed steps.

Note

A Location Profile with the same name as the one used on the exporting cluster must be present in the importing cluster and will be referenced by the restore action.

Note

When restoring an ImageStream into a namespace that already contains a tag with the same name, Veeam Kasten will overwrite the existing tag with the backed-up image. Tags that are identical to the backed-up version are skipped automatically.

Restoring an ImageStream that references an image from another ImageStream in the same namespace may require prioritizing the restoration of the referenced ImageStream first.

During the export process, a temporary intermediate persistent volume will be created using the default storage class. In some cases, the allocated size of this temporary intermediate volume set up for exporting images may be insufficient for the intended image size. This can result in a no space left on device error during this process. To resolve this issue, it is recommended to adjust the Helm parameter ephemeralPVCOverhead, which is set to a default value of 0.1 to increase the storage overhead for this volume.