Hibernate Search is a library that integrates Hibernate ORM with Apache Lucene or Elasticsearch by automatically indexing entities, enabling advanced search functionality: full-text, geospatial, aggregations and more. For more information, see Hibernate Search on hibernate.org.

We just published Hibernate Search 6.0.0.Beta5.

This release mainly changes the way the Elasticsearch backend accesses indexes, introduces delayed commits and near-real-time queries for the Lucene backend, and renames a few methods in the Search DSL. It also includes upgrades to Lucene 8.4, Elasticsearch 7.6 and Hibernate ORM 5.4.12.Final.

Getting started with Hibernate Search 6

If you want to dive right into the new, shiny Hibernate Search 6, a good starting point is the getting started guide included in the reference documentation.

Hibernate Search 6 APIs differ significantly from Search 5.

For more information about migration and what we intend to do to help you, see the migration guide.

What’s new

Aliases for Elasticsearch indexes

As of HSEARCH-3791, Elasticsearch indexes are now accessed through aliases only. Hibernate Search relies on two aliases for each index: one for writes (<myindex>-write) and one for reads (<myindex>-read).

This setup is a first step towards introducing, in a future version of Hibernate Search, a "zero-downtime" reindexing where search queries still return results while reindexing. See this section of the reference documentation for more information, in particular regarding how to change the name of the aliases.

As a result of this change, you will need to either:

  • manually create the two aliases for each index;

  • or drop your indexes and have Hibernate Search re-create empty indexes on startup: it will create the aliases. In this case, you will also need to reindex.

Delayed commits and near-real-time for the Lucene backend

As of HSEARCH-3775, the Lucene backend can now be configured to commit the writer or refresh the reader periodically (e.g. every second), instead of doing it as soon as possible.

This feature, which was already available but exposed differently in Search 5, can lead to substantial performance improvements. See this section of the documentation for details.

Fixed an indexing bug for nested documents in the Lucene backend

HSEARCH-3834 fixed a bug that caused nested documents to be left in the index indefinitely after their root document was deleted. This bug did not affect the results of search queries: the nested documents were not visible after deletion of the root document. However, this could lead to indexes taking more space than necessary.

In order to clean up your index, you should reindex.

More generic wording in the Search DSL

As of HSEARCH-3807, a few methods in the Search DSL have been renamed:

  • searchSession.search(…​).predicate(…​) becomes searchSession.search(…​).where(…​).

  • searchSession.search(…​).asProjection(…​) becomes searchSession.search(…​).select(…​).

  • searchSession.search(…​).asProjections(…​) becomes searchSession.search(…​).select(…​).

  • searchSession.search(…​).asEntity(…​) becomes searchSession.search(…​).selectEntity().

  • searchSession.search(…​).asEntityReference(…​) becomes searchSession.search(…​).selectEntityReference().

The methods with the old name are still present, though deprecated. They will be removed before Hibernate Search 6.0.0.Final is released.

Version upgrades

Hibernate Search 6 requires ORM 5.4.4.Final or later to work correctly. Earlier 5.4.x versions will not work correctly.

Backward-incompatible changes

  • HSEARCH-3808: The names of synchronization strategies for automatic indexing have changed:

    • queuedasync.

    • committed (the default) ⇒ write-sync (still the default).

    • searchable (commonly used for tests) ⇒ sync (still recommended for tests only).

    You will need to update the value of the property hibernate.search.automatic_indexing.synchronization.strategy in your configuration.

  • HSEARCH-3460: an explicit flush() no longer implicitly refreshes the index; you will need to use the new refresh() operation.

  • Full reindexing is necessary if you use the Lucene backend and take advantage of nested documents (@IndexedEmbedded(storage = NESTED)). See above for details.

  • Dropping and re-creating your indexes is necessary if you use the Elasticsearch backend. See above for details or workarounds.


Other improvements and bug fixes

  • HSEARCH-3833: Writing on different local-heap indexes will provoke LockObtainFailedException

  • HSEARCH-3801: Hibernate Search blocks loading of non hibernate batch jobs

  • HSEARCH-3815: Stop publishing the mapper-javabean module

  • HSEARCH-3819: JsonSyntaxException when validating or migrating an Elasticsearch schema with a geo_point field with null_value

  • HSEARCH-3824: Automatically filter search results based on provided routing keys

  • HSEARCH-3825: Purging a document from the POJO mapper will lead to NPE

  • HSEARCH-3836: Allow explicit routing for purges

  • HSEARCH-3778: Remove "index_metadata_complete" configuration property

  • HSEARCH-3779: Remove automatic index optimization

  • HSEARCH-3826: Drop support for non-exclusive index use

  • HSEARCH-3831: Allow setting the automatic indexing sync strategy to a custom bean in configuration

And more. For a full list of changes since the previous releases, please see the release notes.

How to get this release

All details are available and up to date on the dedicated page on hibernate.org.

Feedback, issues, ideas?

To get in touch, use the following channels:

Back to top