On 8/1/2013 8:50 PM, Prasi S wrote: > CAn you pls elaborate on what old-style master/slave mean? Does it mean > that we can have separate indexer and searcher.
With master/slave, you have one Solr instance that's your master. All indexing happens on this server. The other instance(s) replicate from the master. In a typical master/slave setup, queries will be sent to slaves, which all replicate from a master. If there are multiple slaves, there is typically a load balancer in front of them. http://wiki.apache.org/solr/SolrReplication The replication is one way. Unless you do a bunch of manual reconfiguration, you can't change the replication direction. It's difficult to automate. SolrCloud is very different. When you send queries, they are automatically load balanced across the cluster. When you send index updates, they are sent to all replicas automatically, and each replica indexes the data independently. Both queries and updates can be sent to any server in the cluster. http://wiki.apache.org/solr/SolrCloud Thanks, Shawn
