The SPARQLStore communicates with an external repository (triple-store) using the cURL method for transferring data from and to an endpoint.
A connection to an external endpoint can be severed caused by network lag or any other form of data loss. In those cases where the system (namely SMW) cannot make a judgement about how to handle that sort of error (e.g. connection issue as indicated by a HTTP/cURL response code), the
SPARQStore will raise a
HttpDatabaseConnectionException to indicate an erroneous event that was not resolvable.
For an analysis as to what went wrong
HttpDatabaseConnectionException reports the error code, the endpoint, and a possible textual description of the error.
There are no additional information available besides what was reported by the
The example below identifies a communication issue with the Endpoint and needs to be fixed (externally).
Unexpected non-MediaWiki exception encountered, of type "SMW\SPARQLStore\Exception\HttpDatabaseConnectionException" [ab6501bf] /mw-23-01/index.php/Property:Has_area Exception from line 63 of ...\src\SPARQLStore\BadHttpResponseMapper.php: Failed to communicate to Endpoint: http://192.168.1.101:8080/openrdf-sesame/repositories/smw-memory due to curl error: 28 (Connection timed out after 10000 milliseconds).
- Network lag
- Unresponsive server (could no respond to the request the SPARQLStore has made)
- Time-out caused by an extended connection duration
Errors marked or raised as
HttpDatabaseConnectionException are beyond the direct control of Semantic MediaWiki.
- curl error: 56 (Problem (2) in the Chunked-Encoded data)
$smwgSparqlRepositoryConnectorForcedHttpVersionexplicitly force a CURLOPT_HTTP_VERSION for the endpoint communication