About a year we
wrote about an interesting online service for self-education in the world of containers -
Play with Docker . Its authors, supported by Docker Inc., did not stand still and followed the trends that we observe not only in the “narrow circle” of Docker, but also in the industry as a whole: they launched a new similar service, already dedicated to Kubernetes. So, meet Play with Kubernetes (PWK)!
Playground PWK
PWK completely repeats the idea (and even the interface) of its “progenitor” Play with Docker: its
main site is the so-called “playground”, which provides access to the Linux virtual machine in a web browser to enable experiments with Kubernetes clusters . In essence, this is a Minaube SaaS equivalent available for free with its own facilities (working directly in the browser) and restrictions (see below).
Technically, this uses the capabilities of the technology
Docker-in-Docker (DIND), which allows you to run Docker containers inside other Docker containers.
(This approach to working with Docker is not generally recommended, but it has the right to life in particular applications, such as developing Docker itself and related technologies.) Software versions used in the service:
- "Host" operating system (on the node) - CoreOS 7.4.1708,
- Kubernetes server - 1.8.11,
- Kubernetes client - 1.8.14,
- Docker - 03/18/1-ce.
For terminal emulation, a JavaScript implementation of
xterm.js is still offered in the web browser, for which a small number of settings are available (font size selection, the ability to use keyboard shortcuts from macOS). On the site are the same, ie familiar with Play with Docker, restrictions: no more than 4 hours of uninterrupted work within one session and no more than 5 cluster nodes (4 GB RAM each).

If the entrance to the Play with Docker platform is currently available only via Docker ID, then in the case of Kubernetes there is a second option - via an account on GitHub.
The actual launch of the main Play with Kubernetes site, the playground, was "quiet and inconspicuous" last summer, but in itself it is not as interesting as laboratory work. And only a month ago there
was a public announcement of the first such work available on the
Play with Kubernetes Classroom resource.
Kubernetes Lab
It is based on the works of
Jérôme Petazzoni , who devoted 7 years of his life to Docker Inc., where he
went from SRE-engineer to evangelist and conducted
many workshops on working with containers and Kubernetes.

Interesting offtopic: Why did Jérôme leave Docker?Jérôme, which you may also know from the
repository of the DIND already mentioned, began its journey at Docker with a small dotCloud startup competing with Heroku (~ 2011). In 2013, the company joined Docker Inc. and the author, being SRE, made a presentation on containers for the first time. He has become an active speaker since 2014, making an average of 2 performances per week.
By 2016, container orchestration workshops were added to these regular performances ... and at the same time, Jérôme realized that it had exhausted its resources. In October, his feelings were confirmed by the diagnosis of depression. Antidepressants and therapy followed, leading to a visible improvement. However, in the summer of 2017, Jérôme, taking part in a study on mental health, passed the
Maslach Burnout Inventory test for burnout and found that it is in the “red zone”. After talking with people knowledgeable in this matter, he came to the decision to take a break in 2018, and before that he put maximum of his efforts into the implementation of the Docker strategy aimed at supporting Kubernetes.
“In 2018, I'm going to devote time to myself. I study Rust. I am writing a small [musical instrument] Ableton clone to connect a grid controller (like Monome or LaunchPad) to a Raspberry Pi to play live music. I am going to go through a vipassana meditation retreat. I hope to be a mentor for those who have not received such happiness and privileges as I myself, and become the best ally for them. Leaving Docker is the first step and the most difficult, but the upcoming path looks great. ”
But back to the actual Play with Kubernetes ...
The laboratory work offered in the online service is aimed at beginners and is devoted to the main concepts and capabilities of Kubernetes:
- what this system allows you to do: launching containers, load balancing, rolling out new versions of images, autoscaling ...;
- Kubernetes architecture;
- Kubernetes resources: nodes, scams, services, namespaces, secrets;
- declarative approach;
- network model Kubernetes;
- etc.
The main thing is that this theory is supported by practice, for which everything was started. As part of the practical part, it is proposed to launch my first containers and learn how to forward ports for them, create a deployment and scale it, work with resources in YAML, get acquainted with labels and selectors ... I will not list all the stages in detail - it’s better to see (and try) them yourself page
Kubernetes Hands-on Workshop .
It also looks like going through laboratory work in the same way as in Play with Docker: on the left you have a document instruction (including commands for input), and on the right is a terminal (more precisely, there are two for two Kubernetes nodes), which allows “Play” in the K8s cluster administrator and see what really happens and how. The latter, of course, contributes to the ability to perform arbitrary clarifying commands at any stages of the work.

Instead of conclusion
Now only one introductory laboratory work is available in Play with Kubernetes, but the authors promise to create new ones. Their
development in Play with Docker allows us to hope that this will really happen. Those who wish to share their experience can do this in a
special project
repository .
PS
Read also in our blog: