Optimizing SOLR

[gauravmunjal <ma...@gmail.com>]

Hi,

I used the example from the tutorial of Apache SOLR 4.0 and implemented the
same on around 8 million documents. The document has three fields namely
title, id and features. There 8 million docs are basically html files
converted to XML files where title contains the title of the html page and
features contain the text in the body part of the html file.

Now, searching is easy but is taking time. If I search for some term, search
is fast but if the query is something like "burgers in america and canada"
then the query takes a lot of time ~ 5 seconds.

Please let me know ways to optimize the same.

Thank you!




--
View this message in context: http://lucene.472066.n3.nabble.com/Optimizing-SOLR-tp4021261.html
Sent from the Solr - User mailing list archive at Nabble.com.

Re: Optimizing SOLR

[Gora Mohanty <go...@mimirtech.com>]

On 20 November 2012 14:59, gauravmunjal <ma...@gmail.com> wrote:
> Hi,
>
> I used the example from the tutorial of Apache SOLR 4.0 and implemented
the
> same on around 8 million documents. The document has three fields namely
> title, id and features. There 8 million docs are basically html files
> converted to XML files where title contains the title of the html page and
> features contain the text in the body part of the html file.
>
> Now, searching is easy but is taking time. If I search for some term,
search
> is fast but if the query is something like "burgers in america and canada"
> then the query takes a lot of time ~ 5 seconds.

(a) Where are you checking these times? At the Solr web interface
     or through some front-end of your own? I.e., are you sure that Solr
     is the bottle-neck?
(b) If Solr is the bottle-neck, you would need to provide details on what
     else you are doing besides a simple search, e.g., are there facets,
     etc., being used?
(c) What is the size of your data, and have you checked things like
     memory allocated to Solr, etc. One possibility is that Solr is being
     starved of RAM. Also, check memory/CPU usage on your server.

Regards,
Gora