Mapping can be configured via config file and environment variables. The configuration is loaded on startup in this order:
- /etc/mapping.conf
- mapping.conf relative to executable
- from environment variables
Configurations can be overriden by specifying them in later steps of the configuration loading process.
The CGI mode is specified via an environment variable. If the variable FCGI_WEB_SERVER_ADDRS is set, Mapping will assume it is run in FCGI mode. Otherwise it assumes CGI mode.
| Key | Values | Default | Description |
|---|---|---|---|
| fcgi.threads | <integer> | the number of threads to spawn in FCGI mode | |
| userdb.backend | sqlite | The backend to use for the user db | |
| userdb.sqlite.location | <path-to-the-sqlite-file> | The file path where the sqlite database is stored | |
| cache.enabled | true | false | false | Enable/Disable cache |
| cache.type | local | remote | Cache either inside (F)CGI process or use remote cache | |
| cache.replacement | lru | The replacement strategy of the cache | |
| cache.<type>.size | <integer> | Size of <type> in bytes. <type> can be raster, points, lines, polygons, plots, provenance | |
| cache.strategy | always | never | When to cache | |
| global.debug | 0 | 1 | Global debug flag e.g. used in services | |
| global.opencl.preferredplatform | <string> | The preferred platform for OpenCL | |
| global.opencl.forcecpu | 0 | 1 | Force OpenCL to use the CPU instead of GPU | |
| rasterdb.backend | local | remote | local | Remote specifies to use a tileserver to fetch raster tiles instead of loading them from disk |
| rasterdb.tileserver.port | <integer> | Specify the port for starting the tileserver. | |
| rasterdb.remote.host | <string> | Specify the host of the tileserver to connect to. | |
| rasterdb.remote.port | <integer> | Specify the port of the tileserver to connect to. | |
| rasterdb.local.location | <string> | Specify the location for the local rasterdb to use for storing data. | |
| featurecollectiondb.backend | postgres | The backend for the featurecollectiondb | |
| featurecollectiondb.postgres.location | <string> | The SQL connection string e.g. user = 'user' host = 'localhost' password = 'pass' dbname = 'featurecollectiondb_test'. Note that the corresponding database needs to have the POSTGIS extension installed |
|
| wms.norasterforgiventimeexception | 0 | 1 | 1 | Configures the handling of NoRasterForGivenTimeException in WMS. If set to 0, a requested tile for a raster where there is no data for the given time results in a blank tile. If it is set to 1, the Exception is thrown. |
| gdalsource.datasets.path | <string> | The path to the JSON data set descriptions for the GDALSource | |
| crsdirectory.location | <string> | The location of the file containing the definitions of the supported CRS | |
| operators.r.location | <string> | The connection string for the R-Operator to use when connecting to the rserver. e.g. tcp:127.0.0.1:20200. |
|
| uploader.directory | <string> | Path to the directory where the uploader stores the files. |
Specific configurations for distributed mode where the CGI connects to index node and workers compute results.
| Key | Values | Default | Description |
|---|---|---|---|
| log.level | off | error | warn | info | debug | trace | info | The log level of the index and nodes in distributed mode. |
| nodeserver.port | <integer> | The port for a worker node to use | |
| nodeserver.threads | <integer> | The number of threads for a worker to use | |
| nodeserver.cache.manager | local | remote | The cache manager to use. | |
| nodeserver.cache.local.replacement | lru | The replacement strategy of the cache | |
| nodeserver.cache.<type>.size | <integer> | Size of <type> in bytes. <type> can be raster, points, lines, polygons, plots, provenance | |
| nodeserver.cache.strategy | always | never | When to cache | |
| indexserver.port | <integer> | The port for the index server to open and for the workers to connect to | |
| indexserver.host | <string> | The host of the index node for the workers to connect to. | |
| indexserver.scheduler | default | bema | default | The scheduler of the indexserver |
| indexserver.reorg.interval | <integer> | The reorganization interval e.g. 500 | |
| indexserver.reorg.strategy | capacity | graph | geo | capacity: redistribute using memory-usage as metric, graph: Cluster entries by similar operator-graphs, cluster entries by spatial locality | |
| indexserver.reord.relevance | lru | costlru | lru: simple lru replacement, costlru: cost-based lru |