labtest 0.2.3

Overview

Easily create a semi-public, semi-independent deployment of a version of a web site or app.

Semi-public. Allow access via a URL and user credentials. It would be available from the public internet, but not to the general public.

Semi-independent. While the code will be independent, some of the other parts of the infrastructure will be shared. The amount of sharing depends on needs.

Version. Typically this is a branch of development.

In short, it deploys branch foo onto a server that others can reach at foo.test.example.com.

Goals

• Asynchronous evaluation of tickets

• Quick evaluation of new ideas

• Open evaluation to a greater audience

• To make the best test instances, you need:

  • A Machine User for your VCS (for private repos)

  • A Test server configuration in AWS

  • A Dockerfile that works

  • A script to build the container

  • A script to run commands in the container

  • Environment configuration capability

  • Access to a database backup, if necessary

labtest create <branchname> [--name <name>]

labtest update <instancename>

labtest remove <instancename>

labtest stop [<instancename>]

labtest start [<instancename>]

labtest setsecret KEY=value

Two phases: Bulid phase and install phase

• Build phase

  • Git checkout of the repo

  • App build phase:

    • This phase does whatever is necessary to build the application before the docker container is built

    • Runs build commands in a specified docker container

    • The container maps a specified directory on the container to the git checkout

  • Container build phase

    • Builds using a specified Docker file and any other build params

    • Tags and uploads the image to repository

• Install phase

  • setup service

  • setup environment vars

  • other parts of the docker stuff

• Specify a docker repo/image to use to build

• Code is checked out on main server, mapped to docker container

• Follow travis’ or shippable’s method for install and build

Features

• TODO

Credits

This package was created with Cookiecutter and the lgiordani/cookiecutter-pypackage project template.