A lot of works happened for this release. More than 800 commits since version 2.6.
FSCrawler can now send documents to Workplace Search, meaning that users can benefit from a powerful and centralized interface to search for local documents in addition to enterprise documents like Dropbox, Google Drive…
This version is mainly meant to work with Elasticsearch 7.x but you might be able to use it with 6.8 version.
The mapping for folders have changed and is more aligned with the mapping for documents.
Docker images are now generated from the build.
- FSCrawler comes now with an elasticsearch 7.x implementation.
- FSCrawler supports Workplace Search 7.x.
- FSCrawler also supports YAML format for jobs (default).
- The elasticsearch 6.x implementation does not support elasticsearch versions prior to 6.7. If you are using an older version, it’s better to upgrade or you need to “hack” the distribution and replace all elasticsearch/lucene jars to the 6.6 version.
- FSCrawler does not follow symbolic links anymore. You need to set explicitly
trueif you wish revert to the previous behavior.
- The mapping for elasticsearch 6.x can not contain anymore the type name.
- We removed the Elasticsearch V5 compatibility as it’s not maintained anymore by elastic.
- You need to use a recent JVM to run FSCrawler (Java 11 as a minimum. Java 15+ recommended)
- The mapping for the folders changed and is now consistent with the mapping for documents. If you are already using
FSCrawler, you will need to first remove the existing
*_folderindices and remove or edit the default settings files in
~/_default/6/_settings_folder.jsonor any job specific setting file like
Erwan Arzur and
fco-at-801217851326 for this release!