In my former company I started with an HBSearch project. Then I discovered Elasticsearch and put all that HBSearch stuff away.
In my service layer, basically, each time I was persisting my main entities with merge, save or delete, I:
- Serialize the javabean to JSon with Jackson
- Send it to Elasticsearch
For searches, I add to adapt my HBSearch code to an ES code (using ES Query DSL instead of Criteria).
So my advice here is to use ES only as your search engine and not try to combine both.
My 2 cents
David Pilato | Technical Advocate | Elasticsearch.com
@dadoonet | @elasticsearchfr | @scrutmydocs
Le 14 juin 2013 à 09:13, Zoran Jeremic firstname.lastname@example.org a écrit :
Thank you for your explanation. So you are suggesting that it's not possible to run those two together in the same application?
How about using Elasticsearch instead of Hibernate Search? I'm using Hibernate search to support search of the several different Hibernate entities based the user input in the input box and using a wildcard based search on the couple of the Hibernate entities. As the user type something in the textbox, a possible matches are displayed. Do you think it could be possible to implement search in elasticsearch (without hibernate-search) and then to query hibernate to pull the entity? I mean, I know it can be implemented, but would it be wise to apply this approach in terms of performances?
On Thursday, 13 June 2013 23:25:07 UTC-7, Erik Fäßler wrote:
This is a common problem since Lucene is such a wide-spread technology. I had this problem before and I opted for AlexR's proposal: Don't run two libraries dependent of Lucene in the same JVM. I even wouldn't recommend that if the Lucene versions were currently compatible: It might be that e.g. ElasticSearch rapidly develops forward and it actually is quite quick with adapting new Lucene versions. I don't know about Hibernate, but if you update to a newer ElasticSearch version but Hibernate still uses an older version of Lucene you could run into problems, even if your application is currently working. Or of course, you don't do updates.
Thus, why don't use an ES search server? You can start it on the same machine as your search application (or, if you want to, even on a different machine, your flexible, then) and use it via the Rest API, i.e. through the Client Java API. This should work very fine you woudn't have to bother with version conflicts.
On Thursday, 13 June 2013 21:35:29 UTC+2, Zoran Jeremic wrote:
Maybe i didn't understand your idea very well. I'm trying to use java api to integrate elasticsearch with my application. At the moment I'm using local node during development as there are other guys who are working other things, so I don't want to bother them with configuring elasticsearch, but later I want to be able to switch to separate Amazon instance. So, I didn't even run elasticsearch server. The conflict is between elasticsearch client dependencies and hibernate search dependencies.
On Thursday, 13 June 2013 12:23:27 UTC-7, AlexR wrote:
Why not run elastic in a separate jvm?
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to email@example.com.
For more options, visit https://groups.google.com/groups/opt_out.