Even in the example in that page, the client _is_ just querying a
single Solr instance - it is that Solr instance that is then querying
the shards. Is your interest in moving the shards parameter to the
server-side instead? You can do that with the request handler
mapping configuration in solrconfig.xml, and specify the shards as a
default or invariant parameter.
Or are you looking for something more magical?
Erik
On Sep 12, 2008, at 5:38 AM, Upayavira wrote:
The http://wiki.apache.org/solr/DistributedSearch page implies that
you
must know what shards exist when doing a search across multiple
shards.
A colleague tells me that there is a feature that makes it possible to
hide this knowledge.
That is, a something that you query that then forwards your query
requests on to the backend shards and aggregates the results.
That way, your application doesn't need to know about how you split
the
indexes, it thinks it is just querying a single SOLR instance.
Does anything like this exist, or do I have to write it?
Regards, Upayavira
