I’m trying to create a Docker Compose setup, which is fully compatible with the foreman-installer (say, with the default options).
The foreman-docker-compose project is pretty close, but trying to use the foreman-rake command I noticed that the setup resulting from a foreman-installer run differs. Looking at the “foreman-rake” script I noticed that a foreman user is assumed to be installed and the application (obviously) doesn’t run as root (as in Docker, currently).
Can someone help me find the appropriate pieces of the foreman-installer and/or the foreman that help me derive a default setup that is matching? Any hints and constructive ideas are appreciated!
P.S.: You may want to consult my feature branch to get an idea of the current setup.
So, this setup is clearly meant for developing The Foreman, correct?
Should we split up the Compose configuration into separate files, e.g. by moving the development pieces into a docker-compose.override.yml file (as suggested by the Docker documentation)?
This way we could make the picture clearer, separate the concerns of the application architecture + dependencies and development. This could help to keep a suggested reference implementation for a base setup in a single file. We could then add a Kubernetes configuration mirroring that setup (that we would have to keep aligned).
What do you think?
Full disclosure and background information: My motivation here is to have a single file or small set of configuration files that are “production ready” (labeled “The Foreman certified”) that one could curl or wget w/o having to clone the entire repository, for setup automation freed of any manual tasks.