caching - mongodb: force in-memory -
after using myisam years 3 indexes + around 500 columns mio of rows, wonder how "force" mongodb store indexes in memory fast-read performance.
in general, structured table , queries index1=.. or index2=... or index3=.. (myisam) , pretty simple in mongodb well. it's nice if mongodb managing index , ram on own. however, not sure if , way mongodb can speed these queries on indexs-only best.
thanks
it's nice if mongodb managing index , ram on own.
mongodb not manage ram @ all. uses memory-mapped files , "pretends" ram of time.
instead, operating system responsible managing objects kept in ram. typically on lru basis.
you may want check sizes of indexes. if cannot keep of indexes in ram, mongodb perform poorly.
however, not sure if , way mongodb can speed these queries on indexs-only best.
mongodb can use covered indexes retrieve directly db. however, have very specific fields returned. if include fields not part of index, not return "index-only" queries.
the default behavior include all fields, need @ specific queries , make appropriate changes allow "index-only". note these queries not include _id, may cause issues down line.
Comments
Post a Comment