Solr 3.5.0 - different behaviour on rows?

[Frederico Azeiteiro <Fr...@cision.com>]

Hi all,

 

Just testing SOLR 3.5.0. and notice a different behavior on this new
version:

select?rows=1000000000&q=sig%3a("54ba3e8fd3d5d8371f0e01c403085a0c")&?

 

this query returns no results on my indexes, but works for SOLR 1.4.0
and returns "Java heap space java.lang.OutOfMemoryError: Java heap
space" on SOLR 3.5.0

 

Is this normal? As there are no results, why the OutOfMemoryError?

Is it some memory allocated based on the rows number?

 

Regards,

Frederico

 

Re: Solr 3.5.0 - different behaviour on rows?

[Erick Erickson <er...@gmail.com>]

Well, a lot depends upon the query analysis. Are you using
the *exact* same analysis chains in both? Look at the admin/analysis
page and see how your term evaluates. I'm guessing that
WordDelimiterFilterFactory is being used in the 3.5 case and not
in the 1.4.1 case so the 3.5 case is matching everything in your
index in 3.5 and trying to re turn a huge number of rows.

How many documents are found in the 1.4.1 case as opposed to the 3.5 case?

Best
Erick

On Thu, Mar 15, 2012 at 12:22 PM, Frederico Azeiteiro
<Fr...@cision.com> wrote:
> Hi all,
>
>
>
> Just testing SOLR 3.5.0. and notice a different behavior on this new
> version:
>
> select?rows=1000000000&q=sig%3a("54ba3e8fd3d5d8371f0e01c403085a0c")&?
>
>
>
> this query returns no results on my indexes, but works for SOLR 1.4.0
> and returns "Java heap space java.lang.OutOfMemoryError: Java heap
> space" on SOLR 3.5.0
>
>
>
> Is this normal? As there are no results, why the OutOfMemoryError?
>
> Is it some memory allocated based on the rows number?
>
>
>
> Regards,
>
> Frederico
>
>
>