How To Handle Node Shutdown In Kubernetes
summary
Explains concrete behavior and hacks: kubelet uses systemd inhibitor locks for graceful shutdown, introduced in 1.20; non‑graceful handling is an alpha in 1.24 targeting beta in 1.26 and uses the taint node.kubernetes.io/out-of-service plus the feature gate node-out-of-service-volume-detach. Demo stops kubelet, observes pods enter terminating after ~5 minutes and volumes detach after a default 6 minutes, versus immediate failover when the taint triggers force delete and force detach. Shows commands, timing numbers, and manual vs planned automation paths.
AI-written summary · about
chapters (6)
- 0:00Node shutdown in Kubernetes
- 1:48Graceful node shutdown
- 4:47Non-graceful node shutdown
- 13:24Non-graceful shutdown demo
- 24:40Next steps
- 28:42Q&A
related talks
27mØ Kubernetes Failover Improvement: Non-Graceful Node Shutdown
19mGraceful Node Shutdown – A Beginner’s Journey of Kubernetes Development
33mNot so graceful: K8s node failure and stateful workloads
31mBeyond the Start Button: The Hidden Life of Pod Termination in Kubernetes
43mKubernetes SIG Node Intro and Deep Dive
38mCaulking the deployment gap: absolutely zero downtime deployments in Kubernetes
33mThe Gotchas of Zero-Downtime Traffic /w Kubernetes
35mKubernetes, Lessons Learnt