The dumps running on the snapshot servers use rsync to publish the data files from the intermediate storage (dumpsdata) servers to the distribution (clouddumps) servers.
Once we have the dumps running on Kubernetes, we will also need to have a means to publish them in the same way.
This might be rsync, but we might also want to experiment with some other tools, such as lsyncd or parsyncfp2 (which is only available in trixie).
To begin with, we will need an image that includes these sync-utils, so that we can experiment.