Well not exactly a “solution”.
We have a cluster of one HW-Appliance and one Virtual.
I broke up the cluster, resetted the faulty Appliance and let it resync from the healthy appliance.
That fixed it completely, but it’s not a “solution” in my opinion.