Structuring configuration data in a hierarchy

In the previous section, we reduced the data problem to a simple need for key/value pairs that are specific to each node under Puppet management. Puppet and its manifests then serve as the engine that generates actual configuration from these minimalistic bits of information.

A simplistic approach to this problem is an ini style configuration file that has a section for each node that sets values for all configurable keys. Shared values will be declared in one or more general sections:

[mysql]
buffer_pool=15G
log_file_size=500M
...
[xndp12-sql01.example.net]
psk=xneFGl%23ndfAWLN34a0t9w30.zges4
server_id=1

Rails applications customarily do something similar and store their configuration in a YAML format. The user can define different environments, such as production, staging, and testing. The values that are defined per environment override the global setting values.

This is quite close to the type of hierarchical configuration that Puppet allows...