Skip to main content

Highly available Proxmox cluster with Ceph

Client environment: Technology company · approximately 4,500 employees · worldwide locations

Starting point
#

An international technology company headquartered in Thuringia operated several VMware vSphere clusters with ESXi hosts, some of which were ageing. The need to refresh hardware and virtualisation software, together with changes to VMware licensing following Broadcom’s acquisition, prompted a review of the platform strategy.

The client chose a gradual transition to Proxmox VE. The first step was to establish a new, highly available target platform.

My contribution
#

I planned the cluster hardware and implemented the new environment:

  • Designed the x86-64 enterprise server foundation.
  • Installed and configured the Proxmox VE nodes and cluster.
  • Set up Ceph as the distributed storage backend.
  • Hardened the cluster and its individual nodes against unauthorised access.
  • Introduced a role-based access model with clearly defined operational permissions.

Technical approach
#

The solution combines multiple enterprise servers—at least three nodes—in a hyperconverged infrastructure. Each node provides compute capacity, memory and local storage devices. Ceph combines the storage resources into a distributed backend, while Proxmox VE provides virtualisation and central management.

Compute capacity and memory remain local to each host. High availability depends on the combination of cluster operation, redundant data storage and the ability to run virtual machines on other nodes.

Technologies: Proxmox VE, Ceph, x86-64 enterprise servers, HCI, role-based access control.

Outcome
#

The client gained a modern, high-performance and highly available virtualisation platform with low operating costs, providing the foundation for gradually replacing the existing VMware environments.

Reference
#

Reference contact available on request.

← All projects · Discuss a project