c1971aa30b
Initially we were using Weave because it was super easy to install (single kubectl apply -f command), but we now have multiple problems: - it is not maintained anymore - its network policy controller seems to have subtle bugs (I had multiple classes in 2026 where we ran into policies that didn't work as intended, but worked fine when trying on another cluster that used Cilium) - it doesn't support IPV6 We did a test run with Cilium on a few IPv6-only clusters. It worked fine. Then a test run on dual stack clusters. Now we're using Cilium on all kubeadm clusters (removing support for Weave) and we've updated all materials to either remove or update mentions to Weave.
MarkMaker
General principles:
- each slides deck is described in a YAML manifest;
- the YAML manifest lists a number of Markdown files that compose the slides deck;
- a Python script "compiles" the YAML manifest into a HTML file;
- that HTML file can be displayed in your browser (you don't need to host it), or you can publish it (along with a few static assets) if you want.
Getting started
Look at the YAML file corresponding to the deck that you want to edit. The format should be self-explanatory.
I (Jérôme) am still in the process of fine-tuning that format. Once I settle for something, I will add better documentation.
Make changes in the YAML file, and/or in the referenced Markdown files. If you have never used Remark before:
- use
---to separate slides, - use
.foo[bla]if you wantblato have CSS classfoo, - define (or edit) CSS classes in workshop.css.
After making changes, run ./build.sh once; it will
compile each foo.yml file into foo.yml.html.
You can also run ./build.sh forever: it will monitor the current
directory and rebuild slides automatically when files are modified.
If you have problems running ./build.sh (because of
Python dependencies or whatever),
you can also run docker-compose up in this directory.
It will start the ./build.sh forever script in a container.
It will also start a web server exposing the slides
(but the slides should also work if you load them from your
local filesystem).
Publishing pipeline
Each time we push to master, a webhook pings
Netlify, which will pull
the repo, build the slides (by running build.sh once),
and publish them to http://container.training/.
Pull requests are automatically deployed to testing subdomains. I had no idea that I would ever say this about a static page hosting service, but it is seriously awesome. ⚡️💥
Extra bells and whistles
You can run ./slidechecker foo.yml.html to check for
missing images and show the number of slides in that deck.
It requires phantomjs to be installed. It takes some
time to run so it is not yet integrated with the publishing
pipeline.