|
|
|
# Helm Glossary
|
|
|
|
|
|
|
|
Helm uses a few special terms to describe components of the
|
|
|
|
architecture.
|
|
|
|
|
|
|
|
## Chart
|
|
|
|
|
|
|
|
A Helm package that contains information sufficient for installing a set
|
|
|
|
of Kubernetes resources into a Kubernetes cluster.
|
|
|
|
|
|
|
|
Charts contain a `Chart.yaml` file as well as templates, default values
|
|
|
|
(`values.yaml`), and dependencies.
|
|
|
|
|
|
|
|
Charts are developed in a well-defined directory structure, and then
|
|
|
|
packaged into an archive format called a _chart archive_.
|
|
|
|
|
|
|
|
## Chart Archive
|
|
|
|
|
|
|
|
A _chart archive_ is a tarred and gzipped (and optionally signed) chart.
|
|
|
|
|
|
|
|
## Chart Dependency (Subcharts)
|
|
|
|
|
|
|
|
Charts may depend upon other charts. There are two ways a dependency may
|
|
|
|
occur:
|
|
|
|
|
|
|
|
- Soft dependency: A chart may simply not function without another chart
|
|
|
|
being installed in a cluster. Helm does not provide tooling for this
|
|
|
|
case. In this case, dependencies may be managed separately.
|
|
|
|
- Hard dependency: A chart may contain (inside of its `charts/`
|
|
|
|
directory) another chart upon which it depends. In this case,
|
|
|
|
installing the chart will install all of its dependencies. In this
|
|
|
|
case, a chart and its dependencies are managed as a collection.
|
|
|
|
|
|
|
|
When a chart is packaged (via `helm package`) all of its hard dependencies
|
|
|
|
are bundled with it.
|
|
|
|
|
|
|
|
## Chart Version
|
|
|
|
|
|
|
|
Charts are versioned according to the [SemVer 2
|
|
|
|
spec](http://semver.org). A version number is required on every chart.
|
|
|
|
|
|
|
|
## Chart.yaml
|
|
|
|
|
|
|
|
Information about a chart is stored in a special file called
|
|
|
|
`Chart.yaml`. Every chart must have this file.
|
|
|
|
|
|
|
|
## Helm (and helm)
|
|
|
|
|
|
|
|
Helm is the package manager for Kubernetes. As an operating system
|
|
|
|
package manager makes it easy to install tools on an OS, Helm makes it
|
|
|
|
easy to install applications and resources into Kubernetes clusters.
|
|
|
|
|
|
|
|
While _Helm_ is the name of the project, the command line client is also
|
|
|
|
named `helm`. By convention, when speaking of the project, _Helm_ is
|
|
|
|
capitalized. When speaking of the client, _helm_ is in lowercase.
|
|
|
|
|
|
|
|
## Helm Home (HELM_HOME)
|
|
|
|
|
|
|
|
The Helm client stores information in a local directory referred to as
|
|
|
|
_helm home_. By default, this is in the `$HOME/.helm` directory.
|
|
|
|
|
|
|
|
This directory contains configuration and cache data, and is created by
|
|
|
|
`helm init`.
|
|
|
|
|
|
|
|
## Kube Config (KUBECONFIG)
|
|
|
|
|
|
|
|
The Helm client learns about Kubernetes clusters by using files in the _Kube
|
|
|
|
config_ file format. By default, Helm attempts to find this file in the
|
|
|
|
place where `kubectl` creates it (`$HOME/.kube/config`).
|
|
|
|
|
|
|
|
## Lint (Linting)
|
|
|
|
|
|
|
|
To _lint_ a chart is to validate that it follows the conventions and
|
|
|
|
requirements of the Helm chart standard. Helm provides tools to do this,
|
|
|
|
notably the `helm lint` command.
|
|
|
|
|
|
|
|
## Provenance (Provenance file)
|
|
|
|
|
|
|
|
Helm charts may be accompanied by a _provenance file_ which provides
|
|
|
|
information about where the chart came from and what it contains.
|
|
|
|
|
|
|
|
Provenance files are one part of the Helm security story. A provenance contains
|
|
|
|
a cryptographic hash of the chart archive file, the Chart.yaml data, and
|
|
|
|
a signature block (an OpenPGP "clearsign" block). When coupled with a
|
|
|
|
keychain, this provides chart users with the ability to:
|
|
|
|
|
|
|
|
- Validate that a chart was signed by a trusted party
|
|
|
|
- Validate that the chart file has not been tampered with
|
|
|
|
- Validate the contents of a chart metadata (`Chart.yaml`)
|
|
|
|
- Quickly match a chart to its provenance data
|
|
|
|
|
|
|
|
Provenance files have the `.prov` extension, and can be served from a
|
|
|
|
chart repository server or any other HTTP server.
|
|
|
|
|
|
|
|
## Release
|
|
|
|
|
|
|
|
When a chart is installed, Tiller (the Helm server) creates a _release_
|
|
|
|
to track that installation.
|
|
|
|
|
|
|
|
A single chart may be installed many times into the same cluster, and
|
|
|
|
create many different releases. For example, one can install three
|
|
|
|
PostgreSQL databases by running `helm install` three times with a
|
|
|
|
different release name.
|
|
|
|
|
|
|
|
(Prior to 2.0.0-Alpha.1, releases were called _deployments_. But this
|
|
|
|
caused confusion with the Kubernetes _Deployment_ kind.)
|
|
|
|
|
|
|
|
## Release Number (Release Version)
|
|
|
|
|
|
|
|
A single release can be updated multiple times. A sequential counter is
|
|
|
|
used to track releases as they change. After a first `helm install`, a
|
|
|
|
release will have _release number_ 1. Each time a release is upgraded or
|
|
|
|
rolled back, the release number will be incremented.
|
|
|
|
|
|
|
|
## Rollback
|
|
|
|
|
|
|
|
A release can be upgraded to a newer chart or configuration. But since
|
|
|
|
release history is stored, a release can also be _rolled back_ to a
|
|
|
|
previous release number. This is done with the `helm rollback` command.
|
|
|
|
|
|
|
|
Importantly, a rolled back release will receive a new release number.
|
|
|
|
|
|
|
|
Operation | Release Number
|
|
|
|
----------|---------------
|
|
|
|
install | release 1
|
|
|
|
upgrade | release 2
|
|
|
|
upgrade | release 3
|
|
|
|
rollback 1| release 4 (but running the same config as release 1)
|
|
|
|
|
|
|
|
The above table illustrates how release numbers increment across
|
|
|
|
install, upgrade, and rollback.
|
|
|
|
|
|
|
|
## Tiller
|
|
|
|
|
|
|
|
Tiller is the in-cluster component of Helm. It interacts directly with
|
|
|
|
the Kubernetes API server to install, upgrade, query, and remove
|
|
|
|
Kubernetes resources. It also stores the objects that represent
|
|
|
|
releases.
|
|
|
|
|
|
|
|
## Repository (Repo, Chart Repository)
|
|
|
|
|
|
|
|
Helm charts may be stored on dedicated HTTP servers called _chart
|
|
|
|
repositories_ (_repositories_, or just _repos_).
|
|
|
|
|
|
|
|
A chart repository server is a simple HTTP server that can serve an
|
|
|
|
`index.yaml` file that describes a batch of charts, and provides
|
|
|
|
information on where each chart can be downloaded from. (Many chart
|
|
|
|
repositories serve the charts as well as the `index.yaml` file.)
|
|
|
|
|
|
|
|
A Helm client can point to zero or more chart repositories. By default,
|
|
|
|
Helm clients point to the `stable` official Kubernetes chart
|
|
|
|
repository.
|
|
|
|
|
|
|
|
## Values (Values Files, values.yaml)
|
|
|
|
|
|
|
|
Values provide a way to override template defaults with your own
|
|
|
|
information.
|
|
|
|
|
|
|
|
Helm Charts are "parameterized", which means the chart developer may
|
|
|
|
expose configuration that can be overridden at installation time. For
|
|
|
|
example, a chart may expose a `username` field that allows setting a
|
|
|
|
user name for a service.
|
|
|
|
|
|
|
|
These exposed variables are called _values_ in Helm parlance.
|
|
|
|
|
|
|
|
Values can be set during `helm install` and `helm upgrade` operations,
|
|
|
|
either by passing them in directly, or by uploading a `values.yaml`
|
|
|
|
file.
|
|
|
|
|
|
|
|
|