I'm confused. Are you sharding or not? Sharding is used when your index is too big to fit on one Solr, so your docs go to separate Solr nodes. That is, if shard1 contains the doc with id=12, shard2 will NOT have that doc.
If you're not sharding (i.e. each slave has all the docs in your collection), then the shards parameter is totally unnecessary, just put the slaves behind a load balancer as you've indicated and you're done. If you _are_ sharding, then I _strongly_ recommend you go to SolrCloud where you don't have to pay attention to these kinds of details. Otherwise you're trying to re-invent the wheel as the fault tolerance & etc is what SolrCloud is _about_. Best, Erick On Fri, Apr 22, 2016 at 11:21 AM, sangeetha.subraman...@gtnexus.com <sangeetha.subraman...@gtnexus.com> wrote: > Hey guys, > > I am trying to implement Distributed search with Master Slave server. Search > requests goes to Slave Servers. I am planning to have a load balancer before > the Slave servers. So here is the custom search handler which is defined. > > <requestHandler name="/newsearch" class="solr.SearchHandler"> > <lst name="defaults"> > <str name="q.alt">*:*</str> > <str name="shards"> host address of the slaves</str> > </lst> > </requestHandler> > > I believe if more than one slave servers are provided in the shards > parameter, it will not be fault tolerant. So in that case I came across > something like shard.tolenance = true parameter. > But I am not sure on where we can define this ? > https://cwiki.apache.org/confluence/display/solr/Read+and+Write+Side+Fault+Tolerance > Can we set this up with Solr Master Slave architecture. > Could someone please tell me if this is possible to set up at Solr Server > level ? > > Thanks > Sangeetha
