Skip to content

Latest commit

 

History

History
60 lines (43 loc) · 2.85 KB

geoserver.md

File metadata and controls

60 lines (43 loc) · 2.85 KB

GeoServer

Planmap uses an instance of GeoServer version 2.14.3 dedicated to the project at the address https://geoserver.planmap.eu.

Software environment

Planmap's GeoServer runs from inside a Docker container to allow a stable setup for the development and test of new datasets without compromising the production service.

To allow the backup of the ingested data metadata GeoServer's GEOSERVER_DATA_DIR is split from the main software directory structure. For instance, in the Planmap environment, the data settings are stored in /var/lib/geoserver_data.

The Dockerfile used can be found at https://github.com/chbrandt/docker-geoserver. The Docker container is available at https://cloud.docker.com/repository/docker/chbrandt/geoserver.

In practice, GeoServer sees two data volumes/directories:

  • input (or native) data -- Planmap's raster/vector maps and global mosaics;
  • data base settings and metadata.

The native data is a replica of the data found in Planmap's fileserver, https://data.planmap.eu/pub; Both data storages are kept separated to avoid a single point of failure.

Data

Besides the data produced and published by Planmap -- the maps from specific regions of the Moon, Mars, and Mercury -- global basemaps are used for visualization purposes. The global mosaics are provided by USGS Astrogeology service, https://astrogeology.usgs.gov. In particular, we the following basemaps were chosen:

Projection

Planmap' maps are ingested and published in their native projection, global basemaps though are reprojected to EPSG:4326 before adding them to GeoServer. The reason is mere visual: the webapp displays data in latlong; While GeoServer is perfectly capable of reprojecting the individual (Planmap) maps on demand, the convertion of global rasters in realtime are expensive and prone to fault.

Settings

Planmap data packages are structured in raster and vector files. The vector data is packaged in a GeoPackage, which is a composition of the different features of the corresponding map. GeoPackage is composed by multiple tables -- or layers --, each one storing different features. The number of layers and label used by each map/package is defined by the respective author, there is one layer that is common to all geopackages: layer_styles, which is the table providing the map' SLD-/QML-format styles.