Advanced Kubernetes (304) — Lab Files

Hands-on lab assets for Advanced Kubernetes: Production-Scale Internals & the AI-Native Platform by School of DevOps & AI.

Corporate firewall image workaround

If docker pull works on your laptop but Pods fail with ImagePullBackOff and an x509: certificate signed by unknown authority error, preload the course images through host Docker:

bash labs/tools/preload-course-images.sh --pull-only
export COURSE_IMAGE_CACHE=1

Then create or reuse the course cluster normally. The cluster scripts import the required images into containerd on every kind node:

bash labs/clusters/core-internals/create.sh

For an already-running core cluster, you can pull and load all course images in one run:

export COURSE_IMAGE_CACHE=1
bash labs/tools/preload-course-images.sh \
  --cluster kubeadv-core \
  --scope all

This workaround requires host docker pull to succeed. It avoids registry access from inside kind; it does not install a corporate CA or proxy configuration.

Course content lives at https://schoolofdevops.github.io/304-kubeadv/ — start with the Setup section there. This repo is the only thing you clone:

git clone https://github.com/schoolofdevops/304-labs.git
cd 304-labs

All lab commands in the course run from this directory.

Layout

labs/
├── clusters/<profile>/   # cluster profiles — create.sh / verify.sh / teardown.sh
│   └── core-internals/   # M1–M2: kind, 1 control-plane + 1 worker, k8s v1.35
└── m<N>/                 # per-module manifests + checks.json (automated lab checks)

Rules

• One cluster profile at a time. Every profile ships create/verify/teardown.
• Requirements: 8 GB RAM min (16 GB recommended), Rancher Desktop (Moby/dockerd engine), kind ≥ 0.32, kubectl ≥ 1.35, helm, kwok, cilium-cli, hubble, git — full install guide in the course Setup section.
• All images are multi-arch (Apple Silicon + x86-64). No GPU, no cloud account.