Hi, I’m working in an incus orchestrator, so I need to know why you don’t mapped and mount instances images in all cluster nodes?
Because you can only safely mount a filesystem on exactly one machine at a time, doing it on multiple machines concurently leads to corruption.