You are viewing a plain text version of this content. The canonical link for it is here.
Posted to java-user@lucene.apache.org by jens bertheau <be...@booklooker.de> on 2006/03/17 13:54:37 UTC

Best practice setup in multi server environment?

Hi,

I am currently working on switching from MySQL fulltext search to Lucene.
The indexing and searching already works pretty well. I have the following
environment:
1 web server running PHP
1 MySQL server (which will still be used, but not for fulltext queries)
1 server running Lucene
The Lucene index will be created out of the MySQL data.

My question: How can I send a query from the webserver using PHP to the
Lucene server and get my list of result IDs back? 
I suppose one needs some kind of daemon?
Or would if be feasible to setup a Tomcat webserver on the Lucene server
that handles the queries?

Best regards,
Jens
--
View this message in context: http://www.nabble.com/Best-practice-setup-in-multi-server-environment--t1297618.html#a3455237
Sent from the Lucene - Java Users forum at Nabble.com.


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Re: Best practice setup in multi server environment?

Posted by Chris Hostetter <ho...@fucit.org>.
: > 1 web server running PHP
: > 1 MySQL server (which will still be used, but not for fulltext
: > queries)
: > 1 server running Lucene
: > The Lucene index will be created out of the MySQL data.
: >
: > My question: How can I send a query from the webserver using PHP to
: > the
: > Lucene server and get my list of result IDs back?
: > I suppose one needs some kind of daemon?

: Have a look at Solr, which sounds like the ideal tool for the job:
:
: 	<http://incubator.apache.org/solr/>

I agree, Solr sounds perfect for your needs (and i'm not just saying that
because I'm a Solr committer)

 * you update a Solr index by HTTP POSTing simple xml "documents" - so
   your indexing code can be written in any language you want, and run on
   whatever machinemakes sense for your installation.
 * you query a Solr index by HTTP GETing simple xml resuults - so your
   query code client can be a PHP app, or anything else thta can
   fetch/parse xml documents over HTTP.
 * if your documents have a unique identifier, Solr will make sure that
   any "adds" you send it with a document id it already knows about get
   replaced (no duplicates) so you can just POST data to it anytime you
   modify your MySQL DB - no need to worry baout wether it's an add or an
   update.
 * Solr's schema will let you tell it which fields are
   integers, floats, strings, dates, etc... and will make sure range
   queries or sorting on those fields works properly without you having to
   worry about special formatting.  the Schema also lets you say which
   fields you wnat to be searchable, which ones should be stored, which
   analyzer to use for which fields, etc... so you don't have to write any
   javacode to maintain your index -- just configuration.


(okay .. i think that's enough pimping for one email ... head on over to
the "solr-user <a> lucene.apache.org" mailing list if you have any
questions)


-Hoss


---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org


Re: Best practice setup in multi server environment?

Posted by Erik Hatcher <er...@ehatchersolutions.com>.
On Mar 17, 2006, at 7:54 AM, jens bertheau wrote:
> I am currently working on switching from MySQL fulltext search to  
> Lucene.
> The indexing and searching already works pretty well. I have the  
> following
> environment:
> 1 web server running PHP
> 1 MySQL server (which will still be used, but not for fulltext  
> queries)
> 1 server running Lucene
> The Lucene index will be created out of the MySQL data.
>
> My question: How can I send a query from the webserver using PHP to  
> the
> Lucene server and get my list of result IDs back?
> I suppose one needs some kind of daemon?
> Or would if be feasible to setup a Tomcat webserver on the Lucene  
> server
> that handles the queries?

Have a look at Solr, which sounds like the ideal tool for the job:

	<http://incubator.apache.org/solr/>



---------------------------------------------------------------------
To unsubscribe, e-mail: java-user-unsubscribe@lucene.apache.org
For additional commands, e-mail: java-user-help@lucene.apache.org