Adjust to Longhorn disk config rework

What does this MR do and why?

• Brings sylva-core adjustsment for and served as libvirt-metal testing for sylva-projects/sylva-elements/helm-charts/sylva-capi-cluster!229 (merged) CAPM3 Longhorn rework.
• For the above purpose we've also introduced a capm3-kubeadm-virt variant, to test Longhorn with CABPK also. Atm only Ubuntu-based images are built with Kubeadm, for openSUSE we have sylva-projects/sylva-elements/diskimage-builder#52 (closed), and given that the target is to use the SUSE-maintained Metal3 for BareMetalOperator components, this job is based on capm3-rke2-virt-equinix-deploy-suse, so Ubuntu-basesd OS + SUSE Metal3.
• Reworks docs for mentioned envolution and tries to have the following parts clearly separated:
  • one part explaining only what end users have to do to define Longhorn disks, without any explanation on how this is handled
  • one explaining what the Sylva settings translate into in terms of CAPI things (cloud-init) and Longhorn things (node annotations), but without mentioning the Sylva bits
  • one "behind the scene / for dev only" explaining the cooking we do in Sylva to manage all that, in particular the sylva-capi-cluster bits
• Dumps also the type=infrastructure.cluster.x CAPI Secrets to debug-on-exit archive (their contents is cloud-init metadata and network data), as I found it useful to try debug at some point and may come handy further on. If you're against it, I can drop this change

Works on top of the change proposed in !1542 (merged).

Related reference(s)

Closes #417 (closed)
Closes #739 (closed)

Test coverage