CAMP (Configuration AMPlification) takes as input a sample testing configuration and generates automatically a number of diverse configurations. The generation is guided by predefined features and constraints, and utilizes a set of reusable pieces. The current version of CAMP is focused on the Docker environment, and the input and output configurations are specified as Dockerfiles or docker-compose files.
If you try CAMP, feel free to:
The fastest solution is to use docker to run CAMP and all its dependencies in one container, using the following command:
fchauvel@debian$ docker run --name camp \ -it \ -v /var/run/docker.sock:/var/run/docker.sock \ -t fchauvel/camp:latest bash root@9dd7e1f061ce:/camp# cd samples/java root@9dd7e1f061ce:/camp# camp generate -d .
The Docker image include all samples available on Github.
Here are the three main CAMP commands:
camp generate ...to generate alternative test configurations from a seeds
camp realize ...to concretize variables into these test configurations
camp execute ...to run some of these new test configurations
CAMP requires two inputs:
The CAMP Model, which is a YAML file that describes the pieces that can vary in the orchestrations. These may be service providers, feature providers, version or other numerical configurations.
A template orchestration that illustrates how the components listed in the CAMP model are bound and configured together. So far, CAMP only supports the Docker technologies and the orchestration must be a docker-compose file.
From there, CAMP will generate, realize and execute alternative configurations, each with alternative variations.
In XWiki, we have set up CAMP to test XWiki, a wiki platform to facilitate collaborative process inside any organization. XWiki can be set up in various environments and various configurations. CAMP provides means to capture this variations and generate those different environments and configurations. CAMP also facilitates testing against those generated configurations.
In CityGo, CAMP is set up against the CityGo application by ATOS. CityGo can be set up in various environments and configurations. In this example, we demonstrate how CAMP can vary not just elements which map to docker images and services, but also arbitrary parameters and commands in docker files.
CAMP extracts from the input Docker specifications an abstract configuration model, and tries to synthesize new models based on the features, variables and constraints. The figure below illustrates the approach: The new models will then be translated back into Docker specifications. These specifications can be executed in the same way as the original input, and therefore to replace the original testing configuration during either the manual testing or in a continuous integration pipeline.