Installation

It’s strongly recommended to use a python virtualenv

pip install qbertconfig

Usage

qc [-h] [-k KUBECONFIG] <operation> [--name cluster_name] [--uuid cluster_uuid] [-c]

Supported Operations

• fetch - get a kubeconfig for a PMK cluster

• help - show this message

• list-clusters - list available PMK clusters in the target Platform9 Managed Cloud

Providing Credentials

Qbertconfig uses the Openstack SDK to perform authentication against a Platform9 Cloud. Credentials can be provided in either a clouds.yaml file, environment variables, or by using the --os command-line arguments. For more information, please refer to the official documentation

Example

source ~/openstack.rc
qc fetch --name dev-cluster -k dev-cluster.kcfg.yml
export KUBECONFIG=$(pwd)/dev-cluster.kcfg.yml
kubeconfig get nodes --context dev-cluster
kubeconfig get pods -n foo

For more information on openstack rc files and how to generate them, see Installing Openstack CLI Clients.

Testing

Running Tests

pip install -r requirements.txt
nosetests -v -d tests/

Linting

flake8 --exclude versioneer.py

How it works

Here is the basic structure of a Kubeconfig:

apiVersion: v1
kind: Config
preferences: {}
current-context: default
clusters: []
contexts: []
users: []

Each of cluster, context, or user, has a name associated with it. This is the unique identifier for each object, and each context uses these names to tie it all together.

Each of these sections can be managed with the kubectl config command. [Documentation]

This utility will fetch a fresh kubeconfig from the Qbert API, and merge it’s details into the specified kubeconfig.

With the fresh kubeconfig, the following sections are renamed to resolve common collisions when managing many PMK clouds.

• user is renamed to fqdn-username to align with unique keystone environments

• context is renamed to the cluster_name

• cluster is renamed to the cluster_uuid