1 · Cluster & Node Configuration
Advanced: per-node reservations (defaults per Red Hat docs + field practice)
System-reserved memory is auto-computed with the OpenShift auto-sizing tier formula (25% of first 4 GiB, 20% of next 4, 10% of next 8, 6% of next 112, 2% above 128).
2 · VM Templates & Quantities
| Template | vCPU | RAM GiB | Disk GiB | Dedicated CPU | Storage | Qty | Mem OH GiB | Eff RAM GiB | Threads | Eff Disk GiB |
|---|
Unit metrics include the Red Hat per-VM memory overhead formula: 0.002×RAM + 218 MiB + 8 MiB×vCPU + 16 MiB (graphics). Dedicated-CPU VMs cost vCPU+1 full threads.
Storage modes: Block RWX = raw SAN LUN, live-migratable, no overhead · File RWX (NFS/CephFS) = live-migratable, +FS overhead · File RWO = node-bound, blocks live migration. Object (S3) is not a valid VM disk type — Kubernetes PVCs support only Block or Filesystem modes; object storage is sized at cluster level in section 3 for backup/DR copies.
Storage modes: Block RWX = raw SAN LUN, live-migratable, no overhead · File RWX (NFS/CephFS) = live-migratable, +FS overhead · File RWO = node-bound, blocks live migration. Object (S3) is not a valid VM disk type — Kubernetes PVCs support only Block or Filesystem modes; object storage is sized at cluster level in section 3 for backup/DR copies.
3 · Object Storage (Backup & Import)
Object (S3) storage cannot host running VM disks — it sizes the backup/DR target (OADP/Velero) and MTV/CDI import staging. Required ≈ protected disk × fulls × factor + protected disk × change rate × incrementals × factor. Set capacity to 0 to exclude.
4 · Capacity Result
Expert Service
Need a validated design, not an estimate?
This tool gives indicative sizing from Red Hat published overheads. NCINGA delivers measured pilots, MTV migration waves, storage & live-migration network design, and production-ready OpenShift Virtualization platforms.
Methodology & references
- Red Hat OCP 4.20 Virtualization — cluster resource requirements (per-VM memory overhead formula, 2 CPU + ~0.4 GiB per worker for virtualization management, 10 GiB node disk overhead).
- Red Hat KCS — OpenShift Virtualization Cluster Sizing Guide (access.redhat.com/articles/7107457).
- Red Hat KCS 5843241 — node system reservations; auto-sizing tier formula.
- OCP Virtualization docs — live migration defaults (5 parallel/cluster, 2 outbound/node, 800 s/GiB) and RWX storage requirement.
- KubeVirt vmiCPUAllocationRatio (default 10:1); this tool defaults to a conservative 4:1 for VMware-equivalent performance.
- Storage roles: Block/File (RWX) host VM disks; File RWO volumes are node-bound and block live migration; Object (S3) serves backup (OADP/Velero), MTV/CDI import staging and archive only — it cannot back a running VM disk.