Skip to main content

KPM cli

Project description

[![Build Status](https://travis-ci.org/kubespray/kpm.svg?branch=master)](https://travis-ci.org/kubespray/kpm) [![Code Climate](https://codeclimate.com/github/kubespray/kpm/badges/gpa.svg)](https://codeclimate.com/github/kubespray/kpm) [![Coverage Status](https://coveralls.io/repos/github/kubespray/kpm/badge.svg?branch=master)](https://coveralls.io/github/kubespray/kpm?branch=master)


# KPM

KPM is a tool to deploy and manage applications stack on kubernetes.

KPM provides the glue between kubernetes resources (ReplicatSet, DaemonSet, Secrets...). it defines a package has a composition of kubernetes resources and dependencies to other packages.

## Key concepts

### Simplicity

Keep it simple for both users and packagers.

### Declarative and idempotent

Declare what should be deployed and let the system apply only required changes.

### Reproducible deployment

For production and QA, control of each components is fundamental.
Both, dependency declaration and version control in KPM allows to have a predictable application stack deployment.

### Dependencies management

To propose 'ready-to-deploy' applications, the 'packager' can set a dependency list.

### Patch/Paramaterization

KPM encourages to use and reuse directly the 'upstream' package of a component and adapt it to its own requirement:
- <b>Templates:</b> quickly replace values in a resource via jinja2 templates.
- <b>Patch:</b> While template is good, it's not enough. Sometime the value to be edited is not parametrized. In such case, KPM proposes to apply a 'patch' on the resource

### Packages Hub
Quickly compose an application by searching and pick existing components from a registry.

## Quick start

### Deploy application

In this example we will deploy [rocket.chat](https://github.com/RocketChat/Rocket.Chat) an opensource webchat platform:

```
$ pip install kpm
$ kpm deploy ant31/rocketchat --namespace rocket-chat
-> Deploying ant31/rocketchat
app version type name namespace status
------------------ ------- ---- ------------ ----------- --------
ant31/mongodb 1.0.0 svc mongodb rktchat changed
ant31/mongodb 1.0.0 rc mongodb rktchat changed
ant31/rocketchat 1.2.0 svc rocketchat rktchat changed
ant31/rocketchat 1.2.0 rc rocketchat rktchat changed
```

## Install kpm

##### From Pypi

kpm is a python2 package and available on pypi
```
$ sudo pip install kpm -U
````

##### From git

```
git clone https://github.com/kubespray/kpm.git kpm-cli
cd kpm-cli
sudo make install
```

### Configuration

KPM uses `kubectl` to communicate with the kubernetes cluster.
Check if the cluster is accessible:
```bash
$ kubectl version
Client Version: version.Info{Major:"1", Minor:"1", GitVersion:"v1.1.4", GitCommit:"a5949fea3a91d6a50f40a5684e05879080a4c61d", GitTreeState:"clean"}
Server Version: version.Info{Major:"1", Minor:"1", GitVersion:"v1.1.4", GitCommit:"a5949fea3a91d6a50f40a5684e05879080a4c61d", GitTreeState:"clean"}

```

### Customize an existing package
1. Create a new package
2. Edit manifest
3. Change variables
4. Patch custom metric
5. Upload package
6. login
7. push
8.Test

## Account registration
### Signup
### Login/Logout

## Search and deploy a package
### List a user package
#### Show/Pull
### deploy

## Create a new package
The quickest way to get started is to use the command `new`:

```
kpm new namespace/packagename [--with-comments]
```
It creates the directory namespace/packagename

#### Directory structure
A package is composed of a `templates` directory and a `manifest.yaml`.
```
.
├── manifest.yaml
└── templates
├── heapster-rc.yaml
└── heapster-svc.yaml
```
Optionaly, it's possible to add a `README.md` and a `LICENSE`.

#### Templates
The `templates` directory contains the kubernetes resources to deploy.
It accepts every kind of resources (rc,secrets,pods,svc...).

Resources can be templated with Jinja2.

> We recommend to parametrize only values that should be overrided.
> Having a very light templated resources improve readability and quickly point to users which values are
> important to look at and change. User can use 'patch' to add their custom values.

You can declare the deploy order inside the `manifest.yaml`

#### Manifest
The `manifest.yaml` contains the following keys:

- package: metadata around the package and the packager
- variables: map jinja2 variables to default value
- resources: the list of resources, `file` refers to a filename inside the 'template' directory
- deploy: list the dependencies, a special keyword `$self` indicate to deploy current package.

```yaml
package:
name: ant31/heapster
author: "Antoine Legrand <2t.antoine@gmail.com>"
version: 0.18.2
description: Kubernetes data
license: MIT

variables:
namespace: kube-system
replicas: 1
image: "gcr.io/google_containers/heapster:v0.18.2"
svc_type: "NodePort"

resources:
- file: heapster-svc.yaml
name: heapster
type: svc

- file: heapster-rc.yaml
name: heapster
type: rc

deploy:
- name: $self
```
#### Publish

## Compose a package
### Dependency
#### Show manifest
#### variables
#### Patch
#### Shards

## Clustered applications/Shards
### Introduction
### Sharded: yes
### Shard list


=======
History
=======

0.1.0 (2016-2-22)
------------------

* First release on PyPI.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

kpm-0.12.0.tar.gz (27.2 kB view hashes)

Uploaded Source

Built Distribution

kpm-0.12.0-py2.py3-none-any.whl (18.6 kB view hashes)

Uploaded Python 2 Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page