Lab: Linux Container Internals
summary
Hands-on lab traces exactly how containers become Linux processes: DockerD → containerd → runc calls clone and creates namespaces, with cgroups, SELinux and seccomp enforcing isolation; Kubernetes/OpenShift add API, nodes and etcd; images are layered repositories with tags and manifests, turned into a single filesystem by overlay2 and device-mapper graph drivers, creating a copy‑on‑write writable layer that hurts metadata-heavy workloads versus bind mounts. Exercises run on Katacoda lab environment which at scale shows slow startups and intermittent failures.
AI-written summary · about
chapters (8)
- 0:00Start
- 9:29CONTAINERS ARE LINUX
- 12:04THE USER SPACE TOOL CHAIN
- 14:13THE ORCHESTRATION TOOLCHAIN
- 17:49TYPICAL ARCHITECTURE
- 41:21Container Images
- 43:28Fancy Files
- 54:40Fancy File Servers
related talks
30mWhat’s Under the Hood of Docker? Process Separation in the Linux kernel
31mWhat’s under the hood of Docker?
59mBeyond Docker Buzzwords
40mHow are Docker and such are implemented in the Linux Kernel?
05mIgnite: The History of Linux Containers
43mIntro: containerd
37mDocker without docker- understanding the basics
26mContainerization crossroads: Experiences-Experiments-Expectations