This will be brief, and I’ll start by saying two things:
It’s beside the goal of this blog, but I need everyone to get context on the second point. I’ve been, according to docker’s image pull logs, trying to create a sane Jenkins setup with ephemeral docker build agents for 4 months now.
But regardless, I eventually gave up, and instead opted for a more modern solution: Github Actions
The reason for this post being brief, is that all I need to highlight is Flant’s own actions repo
Thanks to this, the trivial implementation ended up looking siple enough:
name: Docker Image CI on: push: branches: [ master ] jobs: build: runs-on: ubuntu-latest steps: - uses: actions/checkout@v2 - uses: werf/actions/build-and-publish@master env: WERF_LOG_VERBOSE: "on"
An important thing to note, is that by default the built-in publish option pushes images to Github’s own repository, and no inputs are exposed to change this. So if you want to use an external repository, your pipeline would end up looking different.
Closing up, I guess what’s next is trying deployment via the converge action, and migrating to 1.2. I noticed just today that Flant published a migration guide covering this exact concern of mine.
Stay tuned for the next blog where Kubernetes earns my eternal hatred as well.tags: