Closed zachkinstner closed 11 years ago
The same exception is reported twice -- so that could be improved, too.
The relevant failure response from Rexster/Titan is shown below. The request seems to be a typical batch Class request for adding 50 new classes.
{
"request":"20ce544b-6fca-45ed-bb1e-b44d74edb1cd",
"success":false,
"queryTime":15492,
"exception":'An error occurred while processing the script for language [groovy]. All transactions across all graphs in the session have been concluded with failure: javax.script.ScriptException: com.thinkaurelius.titan.core.TitanException: Could not read from storage'
}
The RexConnect request paradigm and Factor batch functionality has changed considerably. This shouldn't be an issue anymore -- closing for now.
During the most recent WordNet export attempt (batch upload of 260,000+ FabClass objects):
Five total (of over a thousand) batch requests failed. Four were due to database errors, and their errors were handled/reported correctly. The other failure lead to data not added to the database, but still added the Class data to the local cache anyway. This leads to a scenario where the API always requests retries for this data (since it is present in the cache), but that data isn't actually present in the database.
Relevant API logs: