Server Connections

Before you can publish data with GeoCat Bridge, you will have to configure one or more server connections. This can be achieved in the Connections section of the Bridge dialog:

Adding connections

Click Add… and choose one of the supported server types to create a new server connection:

• GeoServer

• MapServer

• PostGIS

• GeoNetwork

Fill in the required parameters and click the Save button to store the connection details.

You can edit the parameters of a server at any given time. Just select the server connection in the left panel, and edit the fields you wish to change in the right panel.

If you wish to remove a server connection, select it from the left panel and click the Remove button. The Duplicate button may come in handy if you wish to try out an alternative configuration based on an existing server connection.

Bridge will automatically store all server connections in the QGIS settings for later use.

However, sometimes you might want to backup or share all the configured server connections. The Export button will allow you to save a JSON file of the current configuration.

The Import button will do the opposite and load server connections from a JSON file. Note that the imported connections will be added to the list (i.e. no existing connections will be removed). Bridge will append a numeric suffix to server connections with a name that already exists.

Below you can find more info about the different types of services and connections supported by Bridge.

GeoNetwork

Configure a GeoNetwork connection to publish your metadata to an online catalog (CSW).

You can set a (unique) name to identify the connection, the server URL, and the corresponding credentials (user with publish permissions).

The default GeoNetwork node is srv, which will be used if you leave the Node field empty.

Click Test Connection to verify that the connection can be established.

Note

In order to convert the QGIS layer metadata into a format that GeoNetwork understands, Bridge uses the lxml Python library. This library used to be installed by default, but this no longer applies to QGIS 3.22 (LTR) and up, for example.

If Bridge cannot find lxml, it will issue a warning. Therefore, if you wish to publish metadata, please install this missing Python package (e.g. using the OSGeo4W setup or Python’s pip tool).

Users who only have to publish geodata should never have to install this library.

GeoServer

Configure a GeoServer connection to publish your geodata (layers and symbology) as a map service.

When you add a new GeoServer connection, Bridge will automatically add the corresponding WMS and/or WFS endpoints to QGIS, so you can easily add to the server layers to your project, if needed.

You can set a (unique) name to identify the connection, the server URL, and the corresponding credentials (user with publish permissions).

The URL can point to the GeoServer base URL (e.g. http://localhost:8080/geoserver) or the REST API URL (e.g. http://localhost:8080/geoserver/rest). Both URLs will work.

If required, you can create and publish Mapbox styles for vector tiles. Mapbox styles and sprite sheets will be processed after all layers were successfully published to GeoServer.

Click Test Connection to verify that the connection can be established.

You also need to specify how the layer data should be stored. Three different storage methods are available:

1. File-based storage

2. Import into PostGIS database (direct connect) (vector layers only)

3. Import into PostGIS database (managed by GeoServer) (vector layers only)

File-based storage

The layer data is uploaded and stored as files in the GeoServer target workspace datastore.

Bridge prefers to upload vector data as GeoPackages and raster data as GeoTIFF files.

This method is the most straight-forward and should always work, as there are no dependencies. However, depending on the size of the data, this method may be slower and could potentially fail due to network timeouts or server-side file upload size limitations.

Import into PostGIS database (direct connect)

The layer data is imported by Bridge into a PostGIS database directly.

GeoServer will merely establish a link to the imported database table for the created layer(s). This will require a direct PostGIS connection, which you can set up in the Connections section of Bridge as well (see below).

Note that GeoServer should also have read access to that same PostGIS connection.

Note

This method is only available for vector layers.

For raster layers, Bridge will resort to file-based storage instead.

Import into PostGIS database (managed by GeoServer)

The layer data is packaged and uploaded to GeoServer, and a job is scheduled on the server to import the data into a PostGIS database.

This is a good approach if you do not have direct access to the PostGIS database from your machine, but GeoServer does.

Note

This method is only available for vector layers.

For raster layers, Bridge will resort to file-based storage instead.

Please note that some limitations apply:

1. This method requires the GeoServer Importer extension to be installed and enabled.

2. Because Importer can only import vector data from Shapefiles, attribute names may be shortened on export to 10 characters (.dbf limitation). Those names will also be applied to the PostGIS table attributes. However, Bridge will try to make sure that attribute names referenced in styles are renamed accordingly, so the styles won’t break.

3. Created PostGIS tables must be deleted manually. Please read this warning for more information.

Datastore selection

In the Import into PostGIS database (managed by GeoServer) mode, you must either:

• select an existing PostGIS datastore connection on GeoServer that can act as a template;

• or define a new PostGIS connection to use on the GeoServer side.

Bridge will automatically try to list all available PostGIS datastores on GeoServer. You can select any datastore as the template that Bridge should use to create a new datastore with the same connection details as the selected one.

Note

The listed datastores will be prefixed by the workspace name to which they belong.

This does not mean that the layers will be published to that workspace.

Alternatively, you can click the button to create a new datastore yourself, if you know the connection details. Note that this is the connection that GeoServer should use, and that the connection might not work locally on your machine.

For more fine-grained control, or if you wish to specify a JNDI connection pool instead, please create the datastore in the GeoServer Web UI instead. Bridge should then pick up the new datastore automatically, so you can select that as a template.

Warning

Neither the GeoServer REST API nor the Importer extension is authorized to delete underlying PostGIS layer data tables.

This means that Bridge also won’t be able to clean up these data tables and that each publication will create new tables in the database, adding a numeric suffix to the table name and GeoServer feature type to avoid conflicts. If you (re)publish frequently, this may lead to a lot of redundant tables!

Therefore, we recommend creating a clean-up script on the database side that will remove any “orphaned” tables that were created by GeoServer, but are no longer used by any feature type.

Update August 2022

Starting from Bridge 4.3 and GeoServer 2.21.1, existing PostGIS tables can now actually be overwritten by the Importer extension. This should lead to a much cleaner publication process, as there will be less “orphaned” tables.

However, please be aware that the Importer extension still cannot delete any PostGIS tables: if a workspace is completely removed, this will not delete the underlying feature type tables.

PostGIS

Configure a PostGIS connection if you wish to upload layer data to PostGIS directly. Setting up this connection is required when using the Import into PostGIS database (direct connect) in a GeoServer connection.

Note

JNDI connection pool support is currently unavailable.

MapServer

Configure a MapServer connection to publish your map (layer data and symbology). A MapServer endpoint (Mapfile) is created for each QGIS project.

Under Storage, configure how Bridge should save the MapServer data. Choose Local path to store all files on the local file system. Choose FTP service to transmit the files over FTP to the online MapServer instance. Depending on which option is selected, you will see some additional parameters that must be configured:

Proxies & certificates

The GeoNetwork and GeoServer connections mentioned above communicate with the server over HTTP(S) and all their outgoing traffic is handled by the Python requests library. Currently, GeoCat Bridge does not provide a way to set up a proxy server for these connections, and any QGIS proxy settings that may have been defined are ignored. There are plans to add proxy support in future QGIS Bridge releases.

Any QGIS SSL certificate settings are also ignored, which means that Bridge will not be able to connect to servers that use self-signed certificates, for example. However, there is a workaround to this problem. You could set a system environment variable called REQUESTS_CA_BUNDLE that points to a .pem certificate bundle. This bundle should include the root certificate and any intermediate certificates that are required to verify the authenticity of the server certificate.