Homelabbing

In the homelab I have a bunch of commodity hardware running mostly open source software. This includes several misc x86 hardware and several Rasbperry Pis, clusters of Pis, and other small ARM machines. In total I usually have about 20 nodes in my cluster. This provides interesting challenges with regard to managing workloads that run in such a dynamic environment. Some challenges I have faced and will document here include:

• Assigning hardware to workloads on arbitrary nodes
• Handling multi-arch images, identifying issues with upstream providers, and remediating
• Provisioning and maintaining such infrastructure
• Dealing with persistance (databases and solutions to DB management)