Until you nail down what the user did, this is may cause problems. A sharded system assumes that the unique IDs <uniqueKey> in your schema exists on one and only one shard, otherwise you'll be getting multiple copies of the docs.
And you've only shown a multi-core setup, NOT a sharded setup. You need to define a searchhandler in solrconfig.xml similar to a requestHandler and provide the shards as defaults. I don't have the reference close to hand, but you should be able to find it with some searching. Beware the "recursion" problem that you'll see referenced. Last I knew you can't configure your shards in the default search handler, since that's the one that gets the sub-requests for all your nodes.... Best Erick On Tue, Oct 30, 2012 at 5:01 AM, ravi.n <rav...@ornext.com> wrote: > Erick, > > Thanks for your response. > All the 7 folders are of same schema, i mean document structure is same. I > am not very sure how did customer get this data dump into different folders. > Now we have configured Solr with multicore, each core pointing to each > directory and using shards to get a single search response. Please suggest > is this right approach. > > <cores adminPath="/admin/cores" sharedLib="lib" defaultCoreName="coll1"> > <core name="coll1" instanceDir="1" /> > <core name="coll2" instanceDir="2" /> > <core name="coll3" instanceDir="3" /> > <core name="coll4" instanceDir="4" /> > <core name="coll5" instanceDir="5" /> > <core name="coll6" instanceDir="6" /> > <core name="coll7" instanceDir="7" /> > </cores> > </solr> > > And now we should also configure solr for indexing new data from CSV file, i > am not sure how to configure this? > > Regards, > Ravi > > > > -- > View this message in context: > http://lucene.472066.n3.nabble.com/SOLR-To-point-multiple-indexes-in-different-folder-tp4016640p4016946.html > Sent from the Solr - User mailing list archive at Nabble.com.
