How the hell does helm work?: Difference between revisions
From Federal Burro of Information
Jump to navigationJump to search
No edit summary |
|||
Line 1: | Line 1: | ||
== Introduction == | == Introduction == | ||
Helm docs | I'm frustrated with Helm docs. | ||
1st I figure it out, then I post a contribution. | 1st I figure it out, then I post a contribution. |
Latest revision as of 12:07, 20 October 2023
Introduction
I'm frustrated with Helm docs.
1st I figure it out, then I post a contribution.
the helm docs tell you what helm does, but not how helm does it.
I've just upgraded my k8s from 1.14 to 1.16 with a small stack of homebrewed + community charts. Here is what I learneded.
1. First someone writes a chart.
a list of variables.
a list of template that use those variable that make object that hopefully hang together.
hopefully not too much if special case X, slightly different behavior Y.
2. some poor sucker thinks that a chart will save them some time. lets call them a "user".
they "just" do this:
Get the "filled-in" templates aka the manifests
helm get manifest <release> -n <namespace> > <release>-manifests.yaml
Now you can k apply that manifest in another k8s cluster.