Skip to content

Quetzal (February 2018)

Change Summary

Title Git Issue Description
New parallelism query parameter for sstream & bulk-consume 338 Allows controlling the number of parallel processing threads during streaming. Default is 10.
Play upgrade 402 Play minor version upgrade from 2.6.7 to 2.6.11
Akka upgrade 402 Akka libraries minor version upgrade from 2.5.6 to 2.5.9
Full-text search improvements 414 Infrastructure for improving FTS performance. Solves some cases of timeouts in Elastic Search client layer.
The SPARQL Trigger Processor Engine now uses consume rather than bulk-consume 427 This improves STP's robustness.
New fields in STP Dashboard 454 See /proc/stp.md to view the dashboard, and see Using the SPARQL Trigger Processor to learn more about the details displayed.
Activated cluster sniffing for Elastic Search clients 448 Previously full-text search might fail if the default ES instance didn't respond. Now the client will look for alternative ES instances.
Optional source configuration for STP 480 Previously materialized views created by the SPARQL Trigger Processor were always from and to the same environment. This is still the default, but you can now specify a different source environment, using the host-updates-source parameter.
Bug fix: searching for content and link fields 399 Fixed bug that prevented for searching for content and link metadata fields in file and link infotons respectively
Bug fix: Cassandra status in health dashboard 411 The health dashboard failed to report Cassandra status in some cases
Bug fix: STP sensors were written to the wrong folder 386 Didn't affect sensor operation, only the sensor file storage path.
Bug fix: health-detailed dashboard showed wrong metrics for Kafka partition status 449 Showed the status of the (irrelevant) batch process rather than the bg process

Changes to API

Known Issues

Queries on values of all fields, using the _all wildcard, currently do not work. To be fixed.