I assume this type of node setting would benefit querying the same data over and over.
But is there any down side to this in term of GC?
Has anybody deploy nodes with much more RAM for FS than JVM?
Anybody sharing his/her real experience would be highly appreciated.
Thanks.