I use a cluster of 3 nodes with each node 8 TB SSD (8x 1 TB) for ZFS and 8 TB nVME (4x 1,92 TB) for Ceph, Proxmox OS is running on separate disks in ZFS Raid 1. It's connected with a 10 Gbit network. I was wondering if you should consider memory considerations for Ceph and ZFS independently or...