cf-deploy is a tool designed to deploy configuration files for configuration management tools (CFEngine policies, Puppet manifests...). The script itself is part of a wider strategy 7B0B where the code repository is designed to support multiple projects at once, sharing common libraries and keeping project-specific code separated.
You can read more about this approach in my blog post "Git repository and deployment procedures for CFEngine policies". You can also take a look at the slides of my presentation at Config Management Camp 2015.
The code is perfectly functional but doesn't support an external configuration. All the configuration information is inside the cf-deploy script and the Makefile and you'll need to edit both to suit them to your environment. Or, better, you could fork this project and modify cf-deploy so that it reads its defaults from a configuration file and command line options -- I'll gladly accept pull requests to implement that. I am personally fond of AppConfig, but feel free to use any tool that fits.
The components of the tool are currently the following:
- a Makefile, the real workhorse of the system;
- cf-deploy, a Perl frontend to the Makefile;
- a configuration file called projects.db that establishes a correspondance between a project's name, the directory in which the files for those projects reside, and the type of project: remote (deploy the files to a remote server) or local (deploy the files in a local directory)
- a configuration file called hub.db that associates each hub with its location, the project it is assigned to and the environment it serves: production (prod), preproduction (preprod), staging, test...
A sample for hub.db and projects.db is provided.
- you are using git; you can use a different tool (e.g.: svn, hg...) by changing the Makefile;
- your local copy of the git repository is checked out in
/var/cfengine/git, as set in the Makefile variableLOCALDIR; - cf-deploy and its database files are in
/var/cfengine/git/common/tools/deploy, the name of cf-deploy's "databases" is also hardcoded in the script; you are supposed to place/link cf-deploy in a directory in your PATH - the master branch is deployed by default; it is set as a variable
BRANCHin the Makefile but overridden in cf-deploy; - the files are deployed using rsync; you can use a different tool/system or different options for rsync by changing the Makefile;
- when deploying to a remote server, the destination directory is
/var/cfengine/masterfilesas set in the Makefile variableMASTERDIR; it is then overridden in cf-deploy; - the options using when diff-ing files are set in the Makefile in the variable
DIFF_OPTS.
Just run the command without operands and it will tell you. This is what it returns on my laptop where it is linked in /var/cfengine/bin:
$ cf-deploy
Usage:
/var/cfengine/bin/cf-deploy PROJECT_NAME
deploys project PROJECT_NAME on all hubs. It's a shortcut for
/var/cfengine/bin/cf-deploy deploy PROJECT_NAME
/var/cfengine/bin/cf-deploy deploy PROJECT_NAME [ branch BRANCH ] [ hub SERVER ]
Deploys PROJECT_NAME with the specified options
/var/cfengine/bin/cf-deploy preview PROJECT_NAME [ branch BRANCH ] [ hub SERVER ]
preview the changes that would be applied
/var/cfengine/bin/cf-deploy diff PROJECT_NAME hub SERVER [ branch BRANCH ]
runs a diff for a project (hub is mandatory)
/var/cfengine/bin/cf-deploy show PROJECT_NAME
Describes the project PROJECT_NAME, other options are ignored
/var/cfengine/bin/cf-deploy list projects
Lists all defined projects, other options are ignored
/var/cfengine/bin/cf-deploy list hubs
Lists all but the test hubs (normally they are the one you are
mostly interested in, but in case you are more curious...)
/var/cfengine/bin/cf-deploy list all_hubs
List absolutely all hubs, skipping none of them.
You can override the default branch of a project by using the keyword
branch. You can also override the project's hub list by specifying an
hub with the hub keyword -- notice that specifying a hub for a local
project is pointless