Solr Streams Mechanics

Question

With SOLR streaming , is the entire pipeline of activities mentioned in the streaming expression reduce(merge(search())) executed on SOLR server or on the client side ?

Answer 1

All operations are performed on available nodes in your Solr cluster (for example by using the parallel function).

The parallel function is the only oen that can spread work across nodes that isn't used when retrieving the actual data set, so in most cases the work is performed by nodes when retrieving the data set from the node.

A worker collection can be any SolrCloud collection that has the /stream handler configured. Unlike normal SolrCloud collections, worker collections don't have to hold any data. Worker collections can be empty collections that exist only to execute streaming expressions.

You can invoke Streaming Expression through cURL, which shows that the feature doesn't depend on any client side functionality.