dstelljes / container-project Goto Github PK

For consistency, we'll say http://containers.devshawn.com is the canonical version.

• virtual machines vs. containers (not virtualization vs. containers)
• diagram VM and container (emphasize shared kernel/filesystem diffs)
• add slide for system containers vs. app containers (LX[CD] are for system, Docker/rkt are for app)
• split lab services and course stuff into different slides; diagram stuff?
• mention what's up with images: registries store images; containers are instances of images with diffs on top
• list of demos at end
• appendix slide: picture of cactus
• appendix slide: Kyle Hakala or Seth Meyers?
• distributed research systems

I figure we can use issues to track progress on things we are working on.

What's Needed:

• Pick 4 computers to have Docker installed on (I say 2 dev lab, 2 teaching lab)
• Send e-mail to dungeon masters
• Have DM's add us to docker group on those machines
• See that Docker is added on the requested machines

Why Docker Group:
From what I've been reading and trying, there is no way to spin up docker without being apart of this docker group or root. Let me know if you can find a way around it.

Eventually, I'd like to look into container management systems (mainly for Docker as they are most popular)

• Docker Swarm: Native clustering capabilities to turn group of Docker engines into a single, virtual Docker Engine. Easy to scale applications
• Shipyard: Built on top of Docker Swarm, UI to manage Docker images, containers, resources, and more within the swarm system
• Kubernetes: Open-source system developed by Google for automating deployment, scaling, and management of containerized applications
• Docker UI: Web interface for working with Docker as a UI rather than command-line. Could be useful / easy to manage commonly used containers / deploying in a lab setting to multiple computers, worth looking into
• Rancher: Minimalist system (and operating system!) designed solely for running Docker containers. Not necessary applicable in the lab but very cool nonetheless

Feel free to add any suggestions on other things to look into.

The links in your ToC to the sections on Docker and LXC/LXD both are broken. Maybe you need to rebuild the book? Or just rebuild the ToC?

I took the following as indicating that if docker -v returned something appropriate, you had a daemon running, but that's not true. You get the same response from docker -v regardless of whether you have a daemon running.

Before you can spin up your first Docker container, the Docker daemon has to be installed on the workstation you are using:

$ docker -v
Docker version 1.12.1, build 23cf638

It would be good to at a minimum change the wording to clarify things. Having a way to test whether you have a docker daemon running would also be cool.

Should cover stuff like dependencies (emoji, currently) and pushing to GitHub Pages.

Somehow missed rkt when doing initial research. It doesn't look like it's heavily used yet, but the philosophy is neat and it might be worth mentioning.

Have you two considering making a lightweight blog around this project? At some level it might be kind of redundent since the content is all in the book, but more than a few books have started life as a blog, so it's not clear that this is a problem.

I'm definitely not "expecting" or "requiring" it, but if you were interested it would be a cool way to share (updates on) your work "as you go".

Apache Mesos is a distributed kernel that powers DC/OS. Marathon is an orchestration tool that looks neat.

This maybe has more to do with distributed systems than containers, but it might be a thing worth looking into for the lab.

By the end of Week 2, I would like to have a nice write-up in out GitBook of the following things, feel free to add more:

• Containers vs Virtual Machines
• Docker Information (what is it, how does it work, etc) (Docker-Information.md)
• chroot information (maybe cover this as a step below containers... chroot -> container -> VM?)
• LXC Information
• Write up of spinning up containers in the dungeon on lab computers