Open shuttie opened 4 months ago
So you can have cluster-global cache of matching documents. So no need to prune it if segment was not touched.
Redis is a main candidate. Open question: who should fill this cache? We have multiple searchers, how should they decide?
So you can have cluster-global cache of matching documents. So no need to prune it if segment was not touched.
Redis is a main candidate. Open question: who should fill this cache? We have multiple searchers, how should they decide?