Adam Reese
f44f6780fc
|
8 years ago | |
---|---|---|
_proto | 8 years ago | |
cmd | 8 years ago | |
docs | 8 years ago | |
pkg | 8 years ago | |
rootfs | 9 years ago | |
scripts | 8 years ago | |
.gitignore | 9 years ago | |
CONTRIBUTING.md | 9 years ago | |
LICENSE | 9 years ago | |
Makefile | 8 years ago | |
README.md | 8 years ago | |
circle.yml | 8 years ago | |
code_of_conduct.md | 9 years ago | |
glide.lock | 8 years ago | |
glide.yaml | 8 years ago | |
versioning.mk | 9 years ago |
README.md
Kubernetes Helm
Helm is a tool for managing Kubernetes charts. Charts are packages of pre-configured Kubernetes resources.
Use Helm to...
- Find and use popular software packaged as Kubernetes charts
- Share your own applications as Kubernetes charts
- Create reproducible builds of your Kubernetes applications
- Intelligently manage your Kubernetes manifest files
- Manage releases of Helm packages
Helm in a Handbasket
Helm is a tool that streamlines installing and managing Kubernetes applications. Think of it like apt/yum/homebrew for Kubernetes.
- Helm has two parts: a client (
helm
) and a server (tiller
) - Tiller runs inside of your Kubernetes cluster, and manages releases (installations) of your charts.
- Helm runs on your laptop, CI/CD, or wherever you want it to run.
- Charts are Helm packages that contain at least two things:
- A description of the package (
Chart.yaml
) - One or more templates, which contain Kubernetes manifest files
- A description of the package (
- Charts can be stored on disk, or fetched from remote chart repositories (like Debian or RedHat packages)
Using Helm is as easy as this:
$ helm init # Initialize Helm as well as the Tiller server
$ helm install docs/examples/alpine # Install the example Alpine chart
happy-panda # <-- That's the name of your release
$ helm list # List all releases
happy-panda
quiet-kitten
Install
Helm is in its early stages of development. At this time there are no binary releases.
To install Helm from source, follow this process:
Make sure you have the prerequisites:
- Go 1.6
- A running Kubernetes cluster
kubectl
properly configured to talk to your cluster- Glide 0.10 or greater with both git and mercurial installed.
- Properly set your $GOPATH
- Clone (or otherwise download) this repository into $GOPATH/src/k8s.io/helm
- Run
make bootstrap build
You will now have two binaries built:
bin/helm
is the clientbin/tiller
is the server
From here, you can run bin/helm
and use it to install a recent snapshot of
Tiller. Helm will use your kubectl
config to learn about your cluster.
For development on Tiller, you can locally run Tiller, or you build a Docker
image (make docker-build
) and then deploy it (helm init -i IMAGE_NAME
).
The documentation folder contains more information about the architecture and usage of Helm/Tiller.
The History of the Project
Kubernetes Helm is the merged result of Helm Classic and the Kubernetes port of GCS Deployment Manager. The project was jointly started by Google and Deis, though it is now part of the CNCF.
Differences from Helm Classic:
- Helm now has both a client (
helm
) and a server (tiller
). The server runs inside of Kubernetes, and manages your resources. - Helm's chart format has changed for the better:
- Dependencies are immutable and stored inside of a chart's
charts/
directory. - Charts are strongly versioned using SemVer 2
- Charts can be loaded from directories or from chart archive files
- Helm supports Go templates without requiring you to run
generate
ortemplate
commands. - Helm makes it easy to configure your releases -- and share the configuration with the rest of your team.
- Dependencies are immutable and stored inside of a chart's
- Helm chart repositories now use plain HTTP instead of Git/GitHub.
There is no longer any GitHub dependency.
- A chart server is a simple HTTP server
- Charts are referenced by version
- The
helm serve
command will run a local chart server, though you can easily use object storage (S3, GCS) or a regular web server. - And you can still load charts from a local directory.
- The Helm workspace is gone. You can now work anywhere on your filesystem that you want to work.