Addendum:
I delved deep into my memory and documentation:
For "historical reasons," we need (among other things) a "file-based storage" for VM images and, for example, also for install images. So, we first created a dataset for this, let's call it "Pool/files." Also for "historical reasons,"...