Fusion Server

Version 4.1
How To
    Learn More

      Fusion Server 4.2.2 Release Notes

      Release date: 17 May 2019

      Component versions:

      Solr 7.5

      ZooKeeper 3.4.13

      Spark 2.3.2

      Jetty 9.4.12.v20180830

      Ignite 2.6.0

      New features

      No new features were introduced in this release.


      • Web connector improvements:

        • A new f.allowCircularRedirects/Allow circular redirects configuration key allows a request to be redirected to the same URL multiple times.

        • Fixed an issue that prevented crawling sitemaps that include <lastmod> tags.

      • MongoDB connector improvements:

        • The connector now supports SSL

        • The Java driver was updated to eliminate a driver-related error when crawling.

      • The Box connector now includes an option to keep security cache local to a connector node.

      • It is no longer necessary to repeat the entire upgrade procedure to recover from an error. During upgrades to later versions of Fusion Server, the migrator script now permits recovery from errors by correcting the error, repeating the step that failed, and continuing with the upgrade procedure.

      • Jobs API performance is improved when working with large numbers of datasources.

      • Fusion now supports running multiple Parallel Bulk Loader (PBL), Script, or Custom jobs in parallel.

        This feature can be enabled via the Configurations API, as follows:

        curl -X POST -H "Content-type: application/json" -d "2" http://localhost:8765/api/v1/configurations/fusion.spark.max.launcher.jobs
        This is an experimental feature. Note that running multiple jobs in parallel is memory-intensive.
      • Performance is improved in v2 and custom SDK-based connectors.

      Other changes

      • In the SQL service, basic select SQL queries on {collection}_signals, {collection}_query_rewrite, or {collection}_query_rewrite_staging no longer fail when an ORDER BY clause is not included.

      • The install.sh script for systemd no longer starts services after installing them. Services must be started manually.

      • Fixed a document sampling issue which prevented sample documents from appearing in the Index Workbench when previewing a pipeline that used a v2 connector or custom SDK-based connector.

      • Fixed an issue that allowed users to delete pipelines without the correct permissions.

      • Fixed an issue that affected the accuracy of the start times displayed in the Scheduler.

      • Fixed the ownership settings on files in the Fusion tarball so that both the UID and GID are 0 when the package is unpacked by a user with root privileges.

      • Fixed a bug that caused an infinite loop after logging out of Fusion on Chrome.

      • Fixed a bug that prevented the RPC service from starting after connectors were deleted from the blob store.

      • Query Workbench bug fixes:

        • Fixed an issue that changed the display in both panels instead of only the working panel when using Compare mode.

        • Fixed an issue that prevented some document fields from appearing in the dropdown list of fields for faceting.

        • Fixed a bug that caused a "Search Error Unknown Pipeline" error after logging out and logging in again.

        • A few display issues were fixed.

      • Confluence connector bug fixes:

        • Added parent permission ACL values.

        • Fixed a bug that caused a null pointer exception when f.includePrivateContent/Include private content and f.enableSecurityTrimming/Enable Security Trimming were both set to "false".

      • Fixed an issue in the Local Filesystem connector which cause crawls to continue after the configured Maximum Output Limit/maximumItemLimitConfig had been reached.

      Known issues

      • When under load, the Fusion proxy service can occasionally become stuck, causing user authentication to fail. This is the result of the proxy InputStream failing to close properly.

        An upgrade to Fusion 4.2.4 is required to fix this issue. See Upgrade Fusion.

      • Fusion Admin may become sluggish with extended use over a period of several hours. To alleviate this issue, reload the page in the browser while bypassing the cache using the Shift+F5 shortcut.

      • Ommitting the port from the Origin and Host headers from a cURL request incorrectly triggers the CORS filter, resulting in a 400 bad request error.

        For example:

        $ curl -s -o /dev/null -w "%{http_code}\n" -H "Origin: https://example.com:443" -H "Host: example.com:443" http://localhost:8764/api
        $ curl -s -o /dev/null -w "%{http_code}\n" -H "Origin: https://example.com" -H "Host: example.com" http://localhost:8764/api
      • If a user has no assigned groups, conducting SharePoint security trimming on that user will result in a NullPointerException error.

      • An IOException error is sometimes logged while attempting to fetch documents from SharePoint.

      • Entering a number value only in the Query Workbench finds stages as a result. No result should be found.

      • A Job execution error is produced while trying to run a Phrase Detection job with an empty output collection.

      • When creating a new collection within a Fusion AI app, collections previously deleted from other apps may appear in the new app as collections or sub-collections.

      • After starting and stopping a job in a new Fusion AI app, the job’s start time value may be listed as being older than the job’s stop time value. Example:

        start: Wed, April 17, 2019 at 06:15:55 PM +0300
        stop: Wed, April 17, 2019 at 06:14:57 PM +0300