Explore an in-depth overview of Mesos containerization and the default executor in this 42-minute Linux Foundation talk. Delve into the history of Mesos containerization, from process-based systems to the adoption of container standards. Learn about Mesos' native container runtime, support for major container image formats, and integration with OCI, Docker, CSI, and CNI. Discover the advantages of the default executor for running pods in production, including nested containers and debug capabilities. Gain insights on best practices for running sidecar containers, implementing health checks and probes, and understanding termination policies. Examine the Mesos Agent API for nested containers, and get a glimpse of upcoming features on the roadmap. Benefit from the expertise of Mesosphere engineers as they guide you through the intricacies of Mesos containerization and its evolving ecosystem.
An Overview of Mesos Containerization and the Default Executor
Overview
Syllabus
Intro
History of Meses Containerization
Process based (pre 0.10.0, 2011)
Containerizer and isolators (0.18, 2014)
Docker engine integration (0.20, 2014)
Native Docker image support (0.28, 2016)
Adopting container standards
De facto container standard
We need true container standards!
Ideal world
Standards we need for containers
OCI: Open Container Initiative
Container networking spec
CNI: Container Networking Interface
Mesos supports CNI
Container storage spec
Latest new features
Nested Container
Debug Container
What is an Executor?
Executor API
Types of Executor
Why need Task Groups aka Pods?
Limitations of old APE
Default Executor Features
WORKFLOW
Task Group Lifecycle wrt Default Executar
Health Checks
Probes aka Checks
Executor AuthN
Future Work
Summary
Taught by
Linux Foundation
