My first experience with Proxmox came from my homelab. I brought that experience into STI-TEC, where I initiated and led a migration from Nutanix/VMware and Veeam to Proxmox VE, Ceph, and Proxmox Backup Server. Reducing operating costs was a major business driver, but I also wanted more direct control over the platform, its maintenance, and the useful life of our existing hardware.
Why I changed the platform #
In our Nutanix environment, I had encountered problems with automatic updates and limitations on the controls available to me. Licensing and hardware costs were substantial, and hardware lifecycle requirements were pushing us toward replacing equipment that was still useful. I developed a migration plan around retaining that hardware and building a platform I could operate and maintain more directly.
Platform design and migration #
I designed and personally implemented a highly tuned Proxmox/Ceph platform across two eight-node clusters. The storage design included mirrored NVMe boot drives, a mixed-drive Ceph layout in one cluster, and all-SSD Ceph storage in the other. I also configured LACP-bonded dual 25GbE uplinks to increase aggregate bandwidth and provide link redundancy.
I handled all technical planning and implementation for STI-TEC’s move into a Tulsa data center. I migrated customer VDI, employee VDI, and production server workloads from Ohio to Tulsa as part of the consolidation. I also remotely migrated North Carolina VMware workloads to Proxmox VE without data loss.
For the broader platform migration, I retained overall technical ownership while training teammates and supervising the implementation tasks I delegated to them.
Automating rolling maintenance #
I also wanted to retain the convenience of automated rolling maintenance. I wrote custom Ansible playbooks to carry out that workflow on Proxmox, with checks built into the process before and after updates. The automation performs preflight checks, live-migrates VMs away from the host being serviced, updates hosts sequentially, and runs post-update health validation. It gave us a repeatable maintenance workflow that I could adapt to the environment.
Operating the platform #
To make the platform easier to operate, I built dashboards for infrastructure metrics and service uptime, alongside observability and alerting using Grafana, Prometheus, Loki, and InfluxDB. I continued operating the platform and its client-facing Citrix VDI services, including high availability, backup and recovery, maintenance, and service monitoring. I also created operational runbooks and repository documentation and trained teammates in documented procedures.
Results and ownership #
The migration extended the useful life of existing hardware while giving us control over how the platform was configured, monitored, and maintained. Across the virtualization and backup migrations and retirement of Duo MFA, STI-TEC reduced annual licensing and support costs by an estimated $200,000.
For me, this project brought together the parts of infrastructure work I enjoy most: learning through experimentation, making architecture decisions, implementing the systems, and building the tooling that makes them practical to operate over time.