You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by Hitendra Molleti <hi...@itp.com> on 2010/08/16 15:34:22 UTC

Solr for multiple websites

Hi

We have a news portal built on a CMS that heavily uses solr for indexing.
Going ahead we will be migrating all our other portals to the same platform
and are not sure how do we work with Solr for multiple websites.

The options are:

1) Using multiple publications/indexes within solr for each site
2) Run one instance/schema of solr for all websites

The advantage of (1) is all websites will not be affected if there is
something wrong with one solr publication, the downside being sharing of
content between multiple websites.
While the (2) does provide you with flexibility of sharing/searching content
across multiple websites but also comes with the risk of adding all websites
to a single schema (which may cause performance issues).

Would appreciate if someone can guide us in the right direction and share
their experience with us.

Hitendra 




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


RE: Solr for multiple websites

Posted by Hitendra Molleti <hi...@itp.com>.
Hi

Thanks for your reply.

Well, that is the dilemma currently we have, should we go for multiple Solr
indexes for each site or separate schema for each.

Not sure which is the best practice and more reliable solution. 

With Distributed indexing I believe there are few limitations:

> Distributed Searching Limitations
> 
>    * Documents must have a unique key and the unique key must be stored 
>      (stored="true" in schema.xml)
>     
>    * The unique key field must be unique across all shards. If docs with 
>      duplicate unique keys are encountered, Solr will make an attempt to 
>      return valid results, but the behavior may be non-deterministic.
>     
>    * No distributed idf (see
>
http://wunderwood.org/most_casual_observer/2007/04/progressive_reranking.htm
l
> ) (Also see https://issues.apache.org/jira/browse/SOLR-1632 for some new
work
> on this feature.)
>     
>    * Doesn't support QueryElevationComponent
>     
>    * The index could change between stages, e.g. a document that matched a

>      query and was subsequently changed may no longer match but will still
be

Would help to learn if someone has implemented something similar in their
environment and how it perfomed.

Hitendra 

-----Original Message-----
From: Gora Mohanty [mailto:gora@srijan.in] 
Sent: Monday, August 16, 2010 8:32 PM
To: dev@lucene.apache.org
Subject: Re: Solr for multiple websites

On Mon, 16 Aug 2010 17:34:22 +0400
"Hitendra Molleti" <hi...@itp.com> wrote:

> Hi
> 
> We have a news portal built on a CMS that heavily uses solr for
> indexing. Going ahead we will be migrating all our other portals
> to the same platform and are not sure how do we work with Solr
> for multiple websites.
> 
> The options are:
> 
> 1) Using multiple publications/indexes within solr for each site
> 2) Run one instance/schema of solr for all websites
[...]

Sorry, are you talking about using one Solr index vs. several, or
one Solr schema vs. several? You might well do the latter, but I
think that separate Solr indexes will be a significant improvement,
though one might use a separate Solr core for each index, rather
than a separate Solr instance.

We will soon be faced with a similar situation, and at the moment
are leaning towards separate indexes for separate sites, probably
with separate Solr cores. We will be doing some benchmarking on
this, but probably only in a few weeks time.

Regards,
Gora

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




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


Re: Solr for multiple websites

Posted by Gora Mohanty <go...@srijan.in>.
On Mon, 16 Aug 2010 17:34:22 +0400
"Hitendra Molleti" <hi...@itp.com> wrote:

> Hi
> 
> We have a news portal built on a CMS that heavily uses solr for
> indexing. Going ahead we will be migrating all our other portals
> to the same platform and are not sure how do we work with Solr
> for multiple websites.
> 
> The options are:
> 
> 1) Using multiple publications/indexes within solr for each site
> 2) Run one instance/schema of solr for all websites
[...]

Sorry, are you talking about using one Solr index vs. several, or
one Solr schema vs. several? You might well do the latter, but I
think that separate Solr indexes will be a significant improvement,
though one might use a separate Solr core for each index, rather
than a separate Solr instance.

We will soon be faced with a similar situation, and at the moment
are leaning towards separate indexes for separate sites, probably
with separate Solr cores. We will be doing some benchmarking on
this, but probably only in a few weeks time.

Regards,
Gora

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