Solr Search Performance Recommendations

Question

We have a solr instance with 86,315,770 documents. It uses up to 4 GB of memory, and we need this for cutting in a tokenized field called content. The index size on the disk is 23 GB.

Why are we faceted on a tokenized field? Because we want to query the top "n" most used terms in this field. The problem is that there is too much time to complete such requests. Is there a way to improve the time when you draw such a line? Any recommendations?

Thanks in advance.

Answer 1

Solr - , , , . ( getCounts UnInvertedField ).

. , - , Solr ?.

, , , , count (term, query). , , , . , . , , , . (, ) , Solr.

Answer 2

topTerms LukeRequestHandler.