Harvester is an operating system designed for large-scale operations and home servers, but it comes with a steep price tag in terms of hardware requirements. To build a home lab centered around Harvester, one needs to invest in high-end hardware, including an 8-core x86 CPU, 200GB storage, and 32GB RAM.
While Harvester’s minimum requirements are higher than those of popular home lab operating systems like Proxmox and Xen-based XCP-ng, the platform offers a unique set of features that make it an attractive option for advanced users. Built on top of Kubernetes, Harvester leverages Longhorn and KubeVirt to grant additional functionalities.
During my testing, I found the installation process to be relatively easy, although configuring two IP addresses during setup can be confusing. The web UI is clean and organized, making it a pleasure to navigate.
Harvester’s performance was solid, with KVM-based VMs providing excellent performance. However, users need to switch the boot order to a virtual disk after installing an operating system, as Harvester defaults to using the ISO file. Enabling monitoring and logging facilities created Prometheus and Grafana instances, which provided valuable insights into my virtual machines.
One notable feature of Harvester is its ability to pass PCI devices, including GPUs, to virtual machines. This was a welcome surprise during testing. Overall, I’m impressed with Harvester’s capabilities and would recommend it to home lab enthusiasts who require a high-end platform for their virtualization projects.
Source: https://www.xda-developers.com/i-tried-building-a-home-lab-with-harvester