The challenge
The client's VMware renewal carried a steep cost increase under the post-Broadcom subscription model, with no path to scale capacity without paying for tiers they didn't use. The vSAN cluster ran the business (domain controllers, ERP databases, file services), so any migration had to preserve high availability and avoid weekend outage windows.
What we built
We designed a six-node Proxmox VE cluster with Ceph for shared storage on the existing Dell PowerEdge R740 hardware, sized to match the prior vSAN performance envelope and headroom. Roughly 150 VMs were converted and migrated in waves, non-critical workloads first, then production after parity was proven, using live migration where possible and short, scheduled cutovers for the rest. We documented operational procedures (snapshots, backups, HA failover) so the in-house team could run the platform without changing how they work day-to-day.
What changed
All workloads now run on Proxmox + Ceph with HA equivalent to the prior vSphere/vSAN setup and performance at parity across compute and storage. The client owns the platform end-to-end, with no per-core licensing exposure on next year's budget, and storage and compute can scale independently as the business grows.
Stack & partners
- Proxmox VE
- Ceph
- Dell PowerEdge R740 (reused)
- Live migration tooling