Configuration gView.Server

The gView.Server can be configured via the file _config/mapserver.json:

{
     // Folder, where gView Server stores Services, Logins, etc
     "services-folder": "{repository-path}/server/configuration",

     // Path, where gView Server stores Map Request Images
     "output-path": "{repository-path}/server/web/output",
     // Url to the Path, where gView Server stores Map Request Images
     "output-url": "{server-url}/output",

     // Path, where gView Server stores Tiles
     "tilecache-root": "{repository-path}/server/web/tile-caches",

     // The Task Queue
     "task-queue": {
         // Indicates how many requests can be processed at the same time
         "max-parallel-tasks": 50,
         // The maximum length of the queue
         "max-queue-length": 1000
     },

     // default value for mapserver and services
     "mapserver-defaults": {
         // maximum width in pixel of an map image
         "maxImageWidth": 4096,
         // maximum height in pixel of an map image
         "maxImageHeight": 4096,
         // maximum records that will be returned from a query
         "maxRecordCount": 1000
     },

     // Whether clients are allowed to log in through the web interface
     "allowFormsLogin": true,
     // It can be assumed that all calls are made over HTTPS
     "force-https":  false,

     // the level that causes an error on load and deploy services
     "CriticalErrorLevel": {
         "ErrorType": "error"  // any, warning, error, critical, never (default: error)
     },

     // optional: graphic engines. gdiplus only works on windows systems
     "graphics": {
         "rendering": "skia",  // engine for rendering: skia|gdiplus
         "encoding": "skia"  // engine for encoding images (jpg|png): skia|gdiplus
     },

     // optional: projection engine
     "proj-engine": {
         "engine": "ManageProj4Parallel" // ManagedProj4, NativeProj4
     },

     // Optional Settings
     "Facilities": {
         "MessageQueue": {
             "Type": "messagequeue-net",
             "ConnectionString": "http://localhost:9091",
             "Namespace": "development"
         }
     }
 }

In the first section, paths and URLs are defined. These are initially set up via the gview.deploy tool and can easily be adjusted here.

Under task-queue, it can be specified how many requests can be processed simultaneously. If sufficient RAM is available, this value can be increased arbitrarily. However, one must consider that the server generates map images (bitmaps) which can require a significant amount of RAM. If you are consistently reaching the limit of RAM usage (due to many requests), the value can also be set lower. If all tasks are occupied, a request goes into the queue. The length of this queue can also be adjusted here.

In the section mapserver-defaults, default values for map services can be specified. If these values are not explicitly set for a service, the default value configured here will be used. This refers to the maximum size of a map image in pixels that can be retrieved by a client. It can also be specified how many geo-objects can be retrieved in a query at most. If the section or individual values are not specified, the values mentioned above apply.

Note

The values specified here are default values and can be overridden for each service via the Map Service Setting in gView.Carto. Higher values than the default values can also be set for individual services. The values in this configuration are not global maximum values.

Note

A client can request a map image larger than the maximum values. The map server will not return an error but instead deliver a resized image that matches the maximum values. The aspect ratio and geographic area remain intact. However, the resolution of the image (DPI value) is automatically adjusted so that the scale corresponds to the originally requested map image.

A client can verify whether the image was scaled by checking the size of the resulting image. The values are also included in the GeoServices Response in the result JSON.

In the section graphics, the Graphic Engine can be specified. This can either be skia or gdiplus (on Windows platforms). However, gdiplus is becoming obsolete and only works if the application is run on a Windows system.

In the graphics section, the Graphic Engine can be specified. This can either be skia or gdiplus (on Windows platforms). However, gdiplus is considered deprecated and only works when the application is running on a Windows system.

The CriticalErrorLevel section is optional. Here, you can specify from which error level a service cannot be loaded or deployed. Possible values are:

• any: Any error, whether a warning or a critical error, prevents a service from being loaded and deployed.

• warning: From warnings onwards, a service will not be loaded or deployed.

• error: From errors onwards, a service will not be loaded or deployed.

• critical: Only in the case of critical errors will a service not be loaded or deployed.

• never: A service will always be loaded and deployed, regardless of any errors that occur.

The proj-engine section is optional. Here, you can specify which Proj4 Engine to use:

• ManageProj4Parallel: Uses the Proj4 port in C#, where transformations are parallelized for objects with many vertices (>100) to improve performance.

• ManagedProj4: Uses the Proj4 port in C#.

• NativeProj4: Uses the native C++ Proj4 library.

One of the managed versions is recommended, as they are platform-independent and no longer suffer from performance drawbacks. The native version can be used if the managed version does not produce the correct results.

Under Facilities, further optional settings can be made.

One option is MessageQueue. Currently, only a message queue of the type messagequeue-net can be specified (https://github.com/jugstalt/messagequeuenet). A message queue makes sense when the gView.Server is run in multiple instances simultaneously, for example, when gView.Server is running in Kubernetes with several Replicas. The message queue allows the individual components to communicate with each other, for instance, when a map service is modified or deleted. This keeps all instances up to date. With Namespace, a namespace can be specified here, which defines which instances belong together. Since instances from both Staging and Production can access the same MessageQueue API, the namespace can differentiate which instances really belong together.