Your problem seems a lot like an issue I see with Near Real Time (NRT)
replicas. I posted about it in this forum. I was told that a possible
solution was to use the Global Stats feature. I am looking at testing that
now.
Have you tried using Tlog replicas? That fixed my issues with relevancy
diffe
Correction, SOLR-10952 is actually the real case (i misconfigured a test
previously, p).
Well, what's the solution around this problem then? I would assume the cache
stores the results of a query that was already executed with ExactStats, but
that doesn't seem to be the case.
So, what to
