Question about solution, using couchbase , or couchbase+elasticsearch

Hi. I'm web developer.
I'm working now with Solr .
It's search engine for our forums (I indexed all the vbulletin forums into solr).
Because is different site, and perspective... Only search engine to bring solutions.

But we have new goal, new site, It will have a lot of sites indexed, like www.peeplo.com .

We go to index some sites (with this fields: title, url, description, language, country, etc), and develop site to search different words/keywords and get results...

I'm looking to use again Solr, or couchbase+elasticsearch or elasticsearch or ...

What do you think about that ?

How do you plan on storing your json documents in CB?
EX.
key=>document

12345=>{"title":"some title","content":[{"post":"1st post content here","user":"Fred"},{"post":"2nd post content here","user":"Doug"}], "tags":["relevant tag","relevant tag word"," relevant tag keyword"]}

or

Are you just using CB as a Memcached replacement with a TTL or expiration?

1 Answer

« Back to question.

Hello,


Couchbase Server has a plugin for Elasticsearch that allows you to index and query the JSON document you store in Couchbase with Elasticsearch. You can find some information here:


- Webinar : http://info.couchbase.com/webinar-using-elasticsearch-in-couchbase-for-f...
- Documentation : http://www.couchbase.com/docs/couchbase-elasticsearch/index.html



It is really straightforward, you use the XDCR (Cross Data Center Replication), to push the data into Elasticsearch to index them, then you query Elasticsearch to query them. You can look at this code snippet to see how to call elasticsearch and get the document from Couchbase:
https://github.com/tgrall/beersample-java/blob/99-full-text-integration/...



Regards
Tug
@tgrall

thanks, i'm trying with this guides

I have a follow up question after reviewing the example code.

Is there a way to fetch the docs directly from ES and not go through the extra get operations from CB?

in the current implementation - if the ES serch result yields 50 docs, the code will do 50 additional roundtrips to CB... very bad idea...

can ES be configured to store entire docs and not just the index?