Docker & Infrastructure Tutorials¶

Containerization & Orchestration¶

• Multi-Stage Docker: Conda Build → scratch - Start from Conda, ship a zero-OS scratch image. Covers shared lib collection, CA certs, non-root users, and the dynamic loader trap.
• Slim Geospatial + GPU Containers (GDAL + CuPy) - Build a GDAL + PROJ + GPU Python stack with Conda and slim it from 12 GB to under 4 GB using micromamba, multi-stage builds, and a ruthless bloat checklist.
• Slimming GPU Docker Images - Optimize GPU-based containers for production
• Slimming TensorFlow GPU Images - Cut down TensorFlow GPU images to fighting weight
• TensorFlow GPU Slim Images - Repository Skeleton - Complete production-ready example for building slim TensorFlow GPU containers
• Compose Profiles Polyglot Stack - Toggle complex subsystems (realtime, batch, gpu, observability) with surgical Compose profiles

Kubernetes & Clustering¶

• RKE2 on Raspberry Pi Farm - Set up a Raspberry Pi cluster running RKE2
• ZFS Tank with OS on NVMe - Build enterprise-grade storage with ZFS

Infrastructure Automation¶

• Deploy SLURM with Ansible on Raspberry Pi Cluster - Build a HPC cluster with SLURM job scheduler
• Building a 2×8 RKE2 Cluster with Rancher, PGO, and Prefect - Automated Kubernetes cluster provisioning with Ansible, RKE2, Rancher, Postgres operator, and Prefect
• Harbor Container Registry Setup - Enterprise container registry with network storage
• Dask with Ansible (CPU + GPU) - Heterogeneous distributed computing cluster