Support complex capacity queries

[WanzenBug]

In piraeusdatastore/piraeus-operator#440 LINSTOR CSI tried to create a volume with replication with replicasOnSame set in such a way, that it could never succeed. In this case the problem was that the preferred node was the only one in a zone, so there where not enough nodes to fulfill the autoplace criteria.

Ideally, the volume should have never been scheduled there, as the reported storage capacity in this case should have been zero.

LINSTOR recently added a new API to query storage capacity, but it does not account for there advanced auto place setting as far as I know: LINBIT/linstor-server@8d323fa