The freva-deployment software

The freva-deployment software is used to deploy Freva in different computing environments. The general strategy is to split the deployment into different steps, these are : - Deploy a MySQL DB server - Deploy a HashiCorp Vault service for storing and retrieving passwords and other sensitive data (this step get automatically activated once the MySQL DB service is set) - Deploy the Freva-Rest Server The Freva Rest deployment consists of three mandatory and two optional parts:

• The actual databrowser rest API

• Apache solr search backend

• Mongodb to store search statistics

• Redis server acting as broker (optional)

• Data-Loader server that provisions data (netCDF, grb, HDF5 etc) via zarr streams over http (optional)

• Deploy command line interface and python library (freva)

• Deploy web front end (freva_web) The web front end deployment is sub divided into three parts:

  • Deployment of the django web application

  • Deployment of a redis instance acting as database cache

  • Deployment as a apache httpd service as a reverse proxy server for connections from the client to the django web application.

Note

The vault server is auto deployed once the database server is deployed. The vault centrally stores all passwords and other sensitive data.

• Installing freva-deployment
  • 1. Using pre-built binaries.
  • 2. Installation via pip.
  • 3. Using docker
  • Commands after installation:
• Configuring and running the deployment
  • Container or Conda-Forge Deployment
  • Running the deployment
  • Setting the python
  • Setting up the deployment without root-privileges
  • Version checking
  • Using environment variables
• Usage of the text user interface
  • General usage
  • Core setup
  • Web ui setup
  • Freva rest setup
  • Database server setup
  • The run screen.
• Web UI for Freva Deployment
  • Quick Installation
  • Systemd Unit
  • Configuration
  • Deployment Hooks
  • Using Prefect
• Example deployment configuration
  • Typical deployment