0

I have a service that I want to port to Helm. It's a django application that requires memcached and postgresql and a forward proxy like nginx.

I would ideally like to have one repository with all that stuff (dockerfiles/buildah scripts/charts). I think I can that easily with Helm repository if the images were already created. But they're not.

Is there any convention to include Dockerfile's or buildah scripts inside the Helm repository? It seems weird to have to create another repository to build the images, and then to have a helm repository to tie them together. Is this the way it's done currently?

1

Well that's a legit question IMHO, but there are multiple way of doing things (and no one is worth than other).

1- Everything from dev to prod at the same place

If you want to grab everything from dev to production in a single project, it's possible, but you cannot build the docker image from hell command.

Helm is only here to organize, use template in yaml and sync releases following kubernetes file structure.

You can follows helm repository structure and add other scripts, but you have to execute steps in the right order: build, push, deploy

2- The "regular" way to do

I more often saw a separation of production configuration files and application files.

What I do is:

  1. One repository which carry my application source code + everything for the CI (test, build, docker file)
  2. One repository which carry my helm.charts for all my infrastructure.

Choose what suits the best for you, but keep in mind that helm is a templating system for kubernetes yaml files, so it will not build the docker image.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.