I am a newbie to SolrCloud. I have setup a SolrCloud of n leaders, n replicas and a zookeeper ensemble. I have a client that uses SolrJ and has access to millions of docs. This client program runs on a separate machine. Since I want these docs to be indexed as fast as possible, I would like to spawn multiple threads - each adding a set of docs to solr cloud. In this scenario, should each thread be using CloudSolrServer? Since, CloudSolrServer is not thread safe, should each thread maintain its own instance of CloudSolrServer or they be creating a new instance of CloudSolrServer for each doc being submitted? Should I be using ConcurrentUpdateSolrServer instead? But ConcurrentUpdateSolrServer is attached to a single URL. Should I be passing load balancer URL to ConcurrentUpdateSolrServer then?
I would like to hear from the Solr gurus out there as how they would design the indexer/submitter client program for optimal throughput. Thanks in advance. Happy thanks giving! -- View this message in context: http://lucene.472066.n3.nabble.com/How-to-use-CloudSolrServer-in-multi-threaded-indexing-program-tp4021783.html Sent from the Solr - User mailing list archive at Nabble.com.
