nuke-cluster.sh
systemctl stop pve-cluster corosync
pmxcfs -l
rm -rf /etc/corosync/*
rm /etc/pve/corosync.conf
killall pmxcfs
systemctl start pve-cluster...and more from this article
Start with OKD and build around that Don't exclude virting in favor of containers - recipe for bad architecture
S4E
storage: estimate starting at 24TB options:
- storage and compute on same machine
- storage on 1 machine compute on the other
go with second option
hyperconverge rook configue ceph with