Here's a good explanation of replicationFactor: http://wiki.apache.org/solr/SolrCloud You don't want to define this statically, it's about the number of nodes not _how_ they replicate.
This will explain the indexing process: https://cwiki.apache.org/confluence/display/solr/Shards+and+Indexing+Data+in+SolrCloud. There really isn't one unless a replica goes off-line, all files are indexed to all replicas at the same time. Multiple collections work just fine, give them separate names is all and address them as you would cores for requests, i.e. ...solr/collection/select? For your specific question, set up 4 shards with a replicationFactor of at least 2. Best, Erick On Tue, Mar 11, 2014 at 4:40 AM, Gastone Penzo <gastone.pe...@gmail.com> wrote: > Hello, > i i'm testing new solr 4.7 with solr cloud and solr replication. > I can't find any documentation on replicationFactor parameter. > It seems it can be passed only by api on the creation of new collection. > How does this parameter works? > It there a way to specify it statically on solrconfig.xml? > > Another question: > How does replication (not the standard master/slave, but the shard one) > works? > > I explain my situation. > > i'd like to setup 4 nodes with 1 collection and 4 shards and obtain some > type of replication such that if one of four nodes goes down, all data are > still available. > And if i have multiple collections? > > Thank you > > *Gastone Penzo*
