Fusion Server 4.2.3 Release Notes

Release date: 28 June 2019

Component versions:

Solr 7.7.2

ZooKeeper 3.4.13

Spark 2.3.2

Jetty 9.4.12.v20180830

Ignite 2.6.0

New features

  • SQL aggregations aggregations now support the use of the NOT operator inside AND or OR clauses. For example:

    FROM ${inputCollection} c
    WHERE
          ((c.type IN ('click') AND !(c.query <> '*:*')))
          OR
          (c.type IN ('cart','purchase'))

Improvements

  • Wait times at the end of V2 connector jobs have been reduced.

  • Users can now enable or disable spell correction and tail rewrites separately in the Text Tagger stage.

    TextTagger Stage Options^

  • Recently created jobs are now available in the Scheduler immediately.

  • Changed how the Fusion UI handles proxied requests to Solr in order to reduce errors.

  • Images sent to the Apache Tika parser with the Include images option unchecked are no longer discarded.

  • Added a new field, unsafe, to the index pipeline schema.

Other changes

  • Fixed an issue with the SharePoint connector, which was skipping some documents in the index pipeline.

  • Fixed an error with the SharePoint Online connector’s security trimming feature, which failed to filter by the "Everyone except external users" group.

  • Fixed an issue which caused new information added to fusion.properties to be appended to an existing value instead of creating a new line.

  • Fixed an issue which caused the misspelling, phrase, and synonym detection features to publish corrections which had low confidence scores.

  • Fixed an issue which caused file-extension filtering in the SharePoint connector to be case-sensitive.

  • Fixed an issue which produced a Null Pointer Exception (NPE) error while using SharePoint security trimming on a user with no user groups.

  • Fixed how Cross-Origin Resource Sharing (CORS) is handled when the port is missing from the host/origin headers.

  • Fixed an issue that caused the SmartForm login process in the Web connector to fail.

  • Fixed an issue with the Slack connector that caused only non-public information to appear when security trimming was enabled.

  • Fixed an issue which caused multivalued fields that were being sent to the signals collection to be assigned the single value string field type _s instead of the multivalued string field type _ss.

  • Fixed minor issues with the drag-to-reorder feature in the Fusion UI.

  • Fixed an issue that caused the Confluence connector to not create the title field or update the last modified date value.

  • Fixed an error that could result from the Connectors Classic Service failing to start while attempting to migrate from 3.1.5 to 4.2.1.

  • Fixed an error that caused the SharePoint Online datasources to disappear while upgrading Fusion.

  • Fixed an issue which caused a target collection to not be found while attempting to transfer documents from a Solr collection in Fusion 4.1.x to a Solr collection in Fusion 3.1.x.

Addressed Security Vulnerabilities

  • Fusion 4.2.3 upgrades Solr to version 7.7.2, which uses Jackson Release 2.9.8. This resolves multiple Common Vulnerabilities and Exposures (CVE).

  • Fusion 4.2.3 upgrades Hibernate Validator to version 4.1.0, which resolves multiple CVEs.

Known issues

  • The JDBC connector ignores the max_docs field value when crawling a datasource.