Just a guess that this "pollution" occurs during refining counts for distributed facets. Perhaps, there's a way to avoid this "refinement" by tweaking some params.
On Mon, Apr 22, 2019 at 7:48 PM summit <simon.barib...@agilitypr.com> wrote: > Hi, > > I have been digging into our filter cache's low hit rate and i have > discovered that all of its content gets pushed out by some facet results. > This field in particular has a very large number of distinct values. I > have > tried facet.method=fc and facet.method=fcs, but in both cases, it seems > like > an entry gets created in the filter cache for each value of this field, > until the cache is full. Is there a way to "blacklist" a facet field from > the filter cache? We think we'd achieve a much better hit rate with those > fields not interfering with the "good" filters. > > Some specs : > filter cache size : 500 (autowarm 350) > index size : up to 200 million docs over 12 shards > > Thanks > > > > > > -- > Sent from: http://lucene.472066.n3.nabble.com/Solr-User-f472068.html > -- Sincerely yours Mikhail Khludnev
