From VMware to Proxmox VE: Migrating 3,000 Cores and Building a Custom Orchestrator
For years, VMware was my go-to virtualization platform. It’s stable, feature-rich, and widely trusted. But when Broadcom acquired VMware, licensing models changed, prices skyrocketed, and flexibility decreased. It was time to explore alternatives.
That’s when I discovered Proxmox VE – an open-source virtualization platform built on KVM. It combines power, transparency, and flexibility, giving you enterprise capabilities without enterprise-level costs.
Migration at Scale
This wasn’t a small lab test. Over three months, I migrated:
3,000 CPU cores
8 TB RAM
30 TB NVMe storage for production workloads
40 TB archival storage
Here’s what made the process manageable:
1. NFS as Middleware
VMware’s VMFS and Proxmox speak different languages. NFS bridged the gap, allowing VMs to move seamlessly without complex manual conversions.
2. Ceph for Storage
I implemented Ceph for scalable, resilient shared storage. It required tuning, but it handled high workloads and delivered excellent performance.
3. My Own Orchestrator
To streamline management, I built a custom CLI orchestrator on top of Proxmox:
Connect to all clusters from a single command
No need to remember multiple credentials
Perform almost any operation—provisioning, scaling, backups—directly from the terminal
Interestingly, many hosting providers, including Hostinger, rely on Proxmox as their primary hypervisor, building custom infrastructure on top of it. This validates Proxmox’s reliability and scalability.
Why Proxmox VE Stands Out
Proxmox offers transparency and flexibility, and now supports NVIDIA vGPU, enabling GPU-intensive workloads such as AI, VDI, and advanced computing.
KVM, the underlying hypervisor, powers Google Cloud, AWS, Oracle Cloud, Nutanix, OpenStack, and many other cloud platforms. With Proxmox, you get the same robust foundation with a simpler, user-friendly interface.
Key Takeaways
If VMware’s new pricing or complexity is pushing you to reconsider, Proxmox VE is a viable, cost-effective alternative.
Rising costs pushed me to explore Proxmox
NFS and Ceph made migration and storage scalable
Custom orchestrator made cluster management seamless
Even major providers trust Proxmox for their infrastructure
Scale handled: 3,000 cores, 8 TB RAM, 70 TB storage
This is just the beginning. Future posts will dive into migration workflows, performance tuning, orchestrator design, and storage architecture. Stay tuned!