It would be helpful if the “Erase nodes’ disks prior to releasing” option was not just system-wide but configurable for each resource pool.
When deploying CAPI-based Kubernetes clusters on MAAS (with Spectrocloud Palette), we often do not want to wipe those nodes upon release if they are running production workloads. Storage solutions such as Portworx and Rook-Ceph can re-use the data on additional disks, so for servers in such a resource pool, we do not want disk erasure on node release.
However when there are other resource pools for e.g. test/dev purposes, the “Erase nodes’ disks prior to releasing” option is very useful to ensure that nodes given back to the pool are fully wiped. Since you can only configure this option system-wide today, the only viable choice is to always disable it.
If the option could be enabled on a per-resource-pool basis, it would be more useful to users.