Jackrabbit's main configuration is stored in the repository.xml file, which contains global settings for login/access management, versioning, and clustering. It also defines how data is stored for each workspace using a persistence manager and search index. For each created workspace, a workspace.xml file is generated containing workspace-specific settings based on a template from repository.xml. Jackrabbit can be run as a standalone JAR or deployed as a WAR file for administration via a web interface.