For the backup of Persistent Volumes develop procedures using the tools most suited (and potentially developed for) each storage solution. Include in-host backups for nodes in your solution that host path persistent volumes. If you use these, develop a standard volume on each node that can easily be identified and quarantined from the rest of the node. In the case of vSphereVolume plugin storage, verify with your VMware administrator that these vmdk's are indeed part of their usual backp / restore process (as they may be created dynamically and around typical processes). For GlusterFS, not only should you include backup of infrastructure and storage volumes, take care to ensure you have regular backups of the Heketi database (this should be part of best-practice Heketi / GlusterFS setup). Your enterprise likely has a solution suited for the NFS server, take care not to create a process / environment that is outside of these processes.