Skip to content


Now that you have Invenio-CLI installed, we will use it to scaffold a new instance of InvenioRDM.

Initialize project

First, we need to create the project - the necessary files and folders for your InvenioRDM instance.

The CLI will require the following data:

  • Project name: Title of your project with space allowed (project name for humans)
  • Project short name: Hyphenated and lowercased title (project name for machines)
  • Project website: URL where the project will be deployed
  • GitHub repository: Repository in format <owner>/<code repository>
  • Description: Short description of project
  • Author name: Your name or that of your organization
  • Author email: Email for communication
  • Year: The current year
  • Python version: 3.8 (default), 3.6, 3.7, or 3.9.
  • Database: PostgreSQL (default) - we hope to support MySQL in the future.
  • Elasticsearch version: 7 (default) or 6
  • Storage backend: Local file system (default) or in a S3-like backend. If S3 is chosen a MinIO container is provided.

It will also generate a test private key which is needed for SSL support in the development server.

Let's do it! Pressing [Enter] selects the option in brackets [].

invenio-cli init rdm -c <version>

Where <version> should be replaced with he version you want to install, e.g.:

  • LTS release (for production systems): v9.0
  • STS release (for feature previews): v9.0
Initializing RDM application...
Running cookiecutter...
project_name [My Site]:
project_shortname [my-site]:
project_site []:
github_repo [my-site/my-site]:
description [My Site InvenioRDM Instance]:
author_name [CERN]:
author_email []:
year [2021]:
Select python_version:
1 - 3.8
2 - 3.7
3 - 3.6
4 - 3.9
Choose from 1, 2, 3, 4 [1]:
Select database:
1 - postgresql
Choose from 1, 2 [1]:
Select elasticsearch:
1 - 7
2 - 6
Choose from 1, 2 [1]:
Select file_storage:
1 - local
2 - S3
Choose from 1, 2 [1]:
Select development_tools:
1 - yes
2 - no
Choose from 1, 2 [1]:

Generating SSL certificate and private key for testing....
Generating a 4096 bit RSA private key
writing new private key to 'docker/nginx/test.key'
Writing invenio-invenio_cli config file...
Creating logs directory...

Project structure

You can now inspect the generated project structure:

cd my-site
ls -a1

Following is an overview of the generated files and folders:

Name Description
Dockerfile Dockerfile used to build your application image.
Pipfile Python requirements installed via pipenv
Pipfile.lock Locked requirements (generated on first install).
app_data Application data for e.g. vocabularies.
assets Web assets (CSS, JavaScript, LESS, JSX templates) used in the Webpack build.
docker Example configuration for NGINX and uWSGI for running InvenioRDM.
docker-compose.full.yml Example of a full infrastructure stack (DO NOT use in production!)
docker-compose.yml Backend services needed for local development.
docker-services.yml Common services for the Docker Compose files.
invenio.cfg The Invenio application configuration.
logs Log files.
static Static files that need to be served as-is (e.g. images).
templates Folder for your Jinja templates.
.invenio Common file used by Invenio-CLI to be version controlled.
.invenio.private Private file used by Invenio-CLI not to be version controlled.

Notes and known issues

  • You may be prompted with You've downloaded /home/<username>/.cookiecutters/cookiecutter-invenio-rdm before. Is it okay to delete and re-download it? [yes]:. Press [Enter] in that case. This will download the latest cookiecutter template.

  • Some OpenSSL versions display an error message when obtaining random numbers, but this has no incidence (as far as we can tell) on functionality. We are investigating a possible solution to raise less eyebrows for appearance sake.