lucene-solr-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Shawn Heisey <s...@elyograg.org>
Subject Re: using SolrJ with SolrCloud, searching multiple indexes.
Date Sat, 22 Mar 2014 21:05:56 GMT
On 3/22/2014 7:34 AM, Russell Taylor wrote:
> Yeah sorry didn't explain myself there, one of the three zookeepers will return me one
of the solrcloud machines for me to access the index. I either need to know which machine
it returned(is this feasible I can't seem to find a way to access information in SolrCloudServer)
 and then add the extra indexes as shards 
> String shards = solrCloudMachine+":8080/indexB,"+solrCloudMachine+":8080/indexC"
> (solrQuery.add("shards",shards);) 
> 
> or do it in a new way within solrcloud.
> 
> FYI
> My returned index is one of seven indexes under one webapp (solr_search) I want to stitch
on the other six indexes so I can search all of the data (each index is updated from separate
feeds).

SolrCloud eliminates the need to use the shards parameter, so
CloudSolrServer does not expose the actual Solr instances.  You *can*
use the shards parameter, but typically it is done differently than
traditional Solr.

CloudSolrServer thinks mostly in terms of collections, not cores.  There
is a setDefaultCollection method on the server object, or you can do
solrQuery.set("collection","name").

You can query certain shards of a collection or multiple collections,
without ever knowing the host/port/core combinations:

http://wiki.apache.org/solr/SolrCloud#Distributed_Requests

There are also collection aliases on the server side, which let you
access one or more real collections with a "virtual" collection name.

Thanks,
Shawn


Mime
View raw message