Containers – A Look Under the Hood
Talkat ContainerCon @ Open Source Summit Europe 2022 · on Tuesday, 2022-09-13· at 09:00· Liffey Hall 1 (Level 1), Convention Centre Dublin, Spencer Dock, N Wall Quay, North Wall Dublin 1, D01 T1·
summary
Explains exactly which Linux kernel features make containers work and how to build one step‑by‑step using only standard tools: uses unshare / nsenter to create UTS PID IPC mount and network namespaces, creates a veth pair bridge to connect the container network, mounts a 50MB tmpfs and does pivot_root to tmpfs for a minimal root filesystem, demonstrates PID=1 semantics (zombie reaping, signals), shows persistent namespaces via bind mounting, and inspects capabilities from /proc to map to Docker flags like --pid, --network, --cap-add; the example succeeds.
AI-written summary · about
chapters (14)
- 0:00Start
- 1:23Conventional Unix Approach
- 3:45The Containerized Approach
- 10:11Proc File System
- 12:39Closer Look at Namespaces
- 17:53Start a Containerized Process via Docker
- 25:26Uts Namespace for Hostname Isolation
- 27:33Build a New Container
- 31:33Ipc Namespace
- 55:58Mount Namespace
- 1:09:05Influence the Root Directory for a Particular Process
- 1:12:52Change Root Command
- 1:22:36Bind Mounting
- 1:24:10Traditional Unix Privilege Scheme
related talks
33mLinux Container Internals
01h 31mTutorial: Using Linux Primitives to Build Your Own Containers
23mLife of a Container
39mContainers are Linux - but what does it mean?
01h 15mLinux containers in (less than) 100 lines of shell
50mUnderstanding Linux Containers by Building One from Scratch
37mDocker without docker- understanding the basics
49mThings I Wish I Had Learned Earlier About Containers: Lessons from a Seasoned Linux...