Red Hat

Today we’re proud to announce the first Alpha release sporting experimental integration with Elasticsearch!

We also updated to use Apache Lucene 5.5 - the latest stable release of our favourite search engine - as of course we’re not abandoning our traditional users!

What is the Hibernate Search / Elasticsearch integration?

Both Hibernate Search and Elasticsearch can do much more, but for the purpose of explaining this integration at an high level I’ll simplify their definitions as:

Hibernate Search

is a popular extension of the super popular Hibernate ORM framework, which makes it easy to index and query your entities using Apache Lucene.

Elasticsearch

is a very popular REST server which wraps the capabilities of Apache Lucene and makes it easier to scale this service horizontally.

Until today when using Hibernate Search you’d be using Apache Lucene directly, in what we will now call "embedded mode". In this mode a query is executed by the same process of your application, and while indexing happens in background still the overhead of such processing is happening within the same server and within the same JVM process as your Hibernate powered application.

With the Elasticsearch integration, rather than indexing your entities directly by managing the Lucene resources, we will be sending RPCs to an Elasticsearch cluster to achieve a very similar purpose: after all it is also Lucene based, so the feature match is extremely close!

This means that we’re able to transparently map all the current features to this new alternative backend, and by so doing give you more architectural choices at minimum required changes in your applications: the goal is that for most users the differences will be mostly in configuration details.

When using Elasticsearch we will need to send RPCs over the network to run queries and index updates, but on the other hand you benefit from Microservices - style decoupling and all the nice features that Elasticsearch can provide in terms of running and managing an horizontally scalable cluster.

In a nutshell…​

Elasticsearch will manage the index for you, Hibernate ORM will manage your objects and the transactions, Hibernate Search will transparently keep Elasticsearch synchronized to your database transactions and let you query your Domain Model returning managed objects, but moving the heavy lifting to an independent Elasticsearch service.

Elasticsearch integration status

This is literally being developed right now, so do not expect this to be feature complete nor reliable enough to run in a production system. Still, we already have a great set of features working so it’s a nice time to start playing with it and hopefully provide some feedback.

The documentation sports a new Elasticsearch chapter, which is a good place to start.

Configuring the Elasticsearch backend

Dependencies

You’ll need a new dependency, named org.hibernate:hibernate-search-backend-elasticsearch.

The Maven dependencies for Hibernate Search with Elasticsearch:

<dependency>
   <groupId>org.hibernate</groupId>
   <artifactId>hibernate-search-backend-elasticsearch</artifactId>
   <version>5.6.0.Alpha2</version>
</dependency>
<dependency>
   <groupId>org.hibernate</groupId>
   <artifactId>hibernate-search-orm</artifactId>
   <version>5.6.0.Alpha2</version>
</dependency>

Configuration options

Point it to your Elasticsearch node

hibernate.search.elasticsearch.host http://127.0.0.1:9200

Enable the backend to be used for all your indexed entities

hibernate.search.default.indexmanager elasticsearch

Allow it to destroy and create indexes

hibernate.search.elasticsearch.index_management_strategy CREATE_DELETE

For details about these configuration options see the Reference documentation.

Updating the indexes

As with Lucene in embedded mode, the indexes are updated automatically when you create or update entities which are mapped to Hibernate Search using the same annotations already familiar from our traditional index mapping (see Mapping entities to the index structure).

Running a query on an Elasticsearch mapped entity

In many cases the existing way (see Querying) of running queries should work: we do automatically translate the most common types of Apache Lucene queries and many of the queries generated by the Hibernate Search DSL.

On top of translating Lucene queries, you can directly create Elasticsearch queries by using either its String format or a JSON format:

Creating an Elasticsearch native query from a string:

FullTextSession fullTextSession = Search.getFullTextSession(session);
QueryDescriptor query = ElasticsearchQueries.fromQueryString("title:tales");
List<?> result = fullTextSession.createFullTextQuery(query, ComicBook.class).list();

Creating an Elasticsearch native query from JSON:

FullTextSession fullTextSession = Search.getFullTextSession(session);
QueryDescriptor query = ElasticsearchQueries.fromJson(
      "{ 'query': { 'match' : { 'lastName' : 'Brand' } } }");
List<?> result = session.createFullTextQuery(query, GolfPlayer.class).list();

Remaining work ahead

This is an early preview, and while we’re proud of some of the progress there are several areas which still need much coding. On the other hand, implementing some of these is not very hard: this might be the perfect time to join the project.

Please check with JIRA and the mailing lists for updates, but at the time of writing this at least the following features are known to not work yet:

  • Analyzer definitions are not being applied

  • Spatial queries need more work

  • Filters can’t be applied yet

  • Faceting is mostly implemented

  • Scheduled index optimisation is not applied

  • Query timeouts

  • Delete by queries

  • Resolution for Date type mapping is ignored

  • Scrolling on large results won’t work yet

  • MoreLikeThis queries

  • Mixing Lucene based indexes and Elasticsearch based indexes

Any aspect related to performance and efficiency will also be looked at only at the end of basic feature development.

API Changes

In the 5.x series we will keep backward compatibility.

That might come at a cost of not perfect Hibernate Search / Elasticsearch integration API wise. This is something we will address in the 6.x series. But our focus is on offering the right set of features and get feedback in 5.x before improving the APIs.

In a nutshell, 6.x will depend on how you use this feature in 5.6.

How to get this release

Everything you need is available on Hibernate Search’s web site.

Get it from Maven Central using the above coordinates.

Downloads from Sourceforge are available as well, but these don’t contain the Elasticsearch integration components yet. Similarly the WildFly modules also are not including the new Elasticsearch extensions yet.

Feedback

Feedback always welcome!

Please let us know of any problem or suggestion by creating an issue on JIRA, or by sending an email to the developer’s developer’s mailing lists, or posting on the forums.

We also monitor Stack Overflow; when posting on SO please use the tag hibernate-search.

back to top