Klustered: Kubernetes Debugging Live
summary
Live debugging of Cluster 2 shows a deliberately sabotaged single-node Kubernetes cluster: missing kubeconfig, no CNI, CoreDNS pending and containerd running; disk capacity alarm appears then clears. Calico installation is attempted and pinned to Calico v0.6.16 to succeed; initial failure reveals a network policy blocking all pod-to-pod traffic that prevents service communication. Investigation also finds the replica set controller disabled and the scheduler removed from static manifests, explaining why deployments don’t create new pods.
AI-written summary · about
chapters (7)
- 0:00Introduction and Overview
- 0:23Setting Up the Challenge
- 1:25Initial Debugging Steps
- 2:21Identifying the Issues
- 2:55Attempting Fixes
- 5:07Dealing with Network Policies
- 14:58Final Attempts and Conclusion
related talks
01h 32mTutorial: Debug Your Kubernetes Apps
55mThe Hacker's Guide to Kubernetes
33mNavigating the Container Jungle
45mWhat I Learnt Fixing 50+ Broken Kubernetes Clusters
34m"My CNI Plugin Did… What?!": Debugging CNI with Style & Aplomb
40mAttacking Kubernetes Clusters Through Your Network Plumbing
36mKubernetes Confessions: Real Mistakes We’ve All Made (and What They Teach Us)
35mThe Art of Debugging Inside K8s Environment