This guide details the operation of a View Server and all of its components.

Since the View Server is a Docker based software and all of its components are distributed and executed in context of Docker images and containers, basic knowledge of Docker and Docker Swarm is a prerequisite.


View Server Architecture

Figure 1 View Server Architecture

The View Server consists of the following service components (with their respective Docker image in parenthesis):

  • Reverse proxy (traefik)

  • Web Client (client)

  • Cache (cache)

  • Renderer (core)

  • Registrar (core)

  • Seeder (cache)

  • Preprocessor (preprocessor)

  • Ingestor (ingestor)

  • Database (postgis)

  • Queue Manager (redis)

  • Log collector (fluentd)

  • Kibana (kibana)

  • Elasticsearch (elasticsearch)

  • Shibboleth SP3 (shibauth)

These services are bundled and managed together in a Docker Swarm via Docker Compose configuration files.

Docker Images

The software is distributed as Docker images, which can be instantiated and run in their intended role. Some images are hosted on docker hub, the official and default repository for Docker images. Other images reside on an EOX hosted registry. Images from the official registry are only identified via their name, whereas images from the EOX registry conventionally use the full URL, including the domain name. Below is a list of the used images:

  • mdillon/postgis:10

  • redis

  • traefik

  • elasticsearch

  • kibana

  • atmoz/sftp








Configuration Files

The following configuration files impact the behavior of the View Server:

  • index.html: This is the main file to configure the client. In this file, the viewing layer, search and download endpoints are configured. Usually this is associated with additional backdrop and overlay layers.

  • preprocessor.yml: This file configures the preprocessing steps.

  • mapcache.xml: This file defines the input sources of the cache and its published layers.

  • This file sets up the registrar and renderer side of the VS.

Initialization and Setup

In order to help with the initial setup of a VS, the pvs_starter package described in the section Initialization allows to quickly establish the required structure of configuration files.

The section Setup describes how to deploy a Docker Swarm stack using the configuration files generated in the initialization step.