First of all, make sure you use docvalues for facet fields with many unique values.
If that still does not help you can try the following. My kollega Toke Eskildsen has made a huge improvement when faceting IF the number of results in the facets are less than 8% of the total number of documents. In this case we get a substantial improvement in both memory use and query time: See: https://plus.google.com/+TokeEskildsen/posts/7oGxWZRKJEs We have tested it for index with 300M documents. From, Thomas Egense On Wed, Jun 11, 2014 at 5:36 PM, marcos palacios <mpcmar...@gmail.com> wrote: > Hello everyone. > > > > I’m having problems with the performance of queries with facets, the temp > expend to resolve a query is very high. > > > > The index has 10Millions of documents, each one with 100 fields. > > The server has 8 cores and 56 Gb of ram, running with jetty with this > memory configuration: -Xms24096m -Xmx44576m > > > > When I do a query, with 20 facets, the time expended is 4 – 5 seconds. If > the same request is did another time, the > > > > Debug query first execution: > > <double name="time">6037.0</double><lst name="query"><double > name="time">265.0</double></lst><lst name="facet"><double > name="time">5772.0</double></lst> > > > > Debug query seconds executions: > > <double name="time">6037.0</double><lst name="query"><double > name="time">1.0</double></lst><lst name="facet"><double > name="time">4872.0</double></lst> > > > > > > What can I do? Why the facets are not cached? > > > > > > Thank you, Marcos >
