These are the notes I took on the GeoServer workshop at the last gvSIG Conference.
It is important to change the settings of the geoserver_data_dir in the web.xml file to keep the data each time you restart the application container (like Tomcat). It is also good to check out the other settings as it contains interesting facts such as the type of projections to be used or the size of the cache. There are configurable data on the fly and data not configurable on the fly.
Let’s try adding some data sources to generate the layers. To add a shapefile, you have to copy the file in the same physical server machine. To include the shapefile in GeoServer, look for the option of adding a new shapefile datastore type. If you use the location “file: data /…” you use a relative uri to geoserver. You can also search using the “Browse” and use absolute paths.
Warning: Do not you give permission to any user on the configuration interface because they can see all the physical hard disk in this type of dialog.
Memory Mapped Buffers
It is best to use memory mapped buffers (unless you use Windows) if you have enough RAM, then you will avoid continuous access to physical disk. Also, it is best to reproject from native to declared projection. If the shapefile is very big, calculate the bounding box take some time. This does not happen in real databases where spatial indexes.
GeoServer allows you to insert watermarks in your data (for example to use OpenStreetMap).
When you use database connections, you should check the validate connection option, because you never know when the connection is going to crash.
GeoWebCache allows, on the latest version of GeoServer, to set up easily which layers will be cached.
You can use data which varies through time, for example for storms and hurricanes. On the database table, you will have a column for time. You can also use elevation data, but then you have to use it with Google Earth. The interesting thing is that, once you have the data, some hipothetical free visor can be used to support it…