Replica's don't double the amount of field data loaded. A query will only load what it needs from the shards it needs, but it won't load primary and replica, just one or the other.
Ideally you should 1) upgrade to aggs, facets are deprecated and not as performant, and 2) upgrade to doc_values. On 4 May 2015 at 18:50, Junheng Gong <lunadan...@gmail.com> wrote: > we have about 2.2B data in our elasticsearch, and we using facet and > function score query on those data. > > it will load many data to fielddata, and replica double it. > > so my question is "how to set replica to make sure it won't use field data > expect primary shard down"? > > -- > You received this message because you are subscribed to the Google Groups > "elasticsearch" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to elasticsearch+unsubscr...@googlegroups.com. > To view this discussion on the web visit > https://groups.google.com/d/msgid/elasticsearch/83db33a6-0248-42eb-b913-8da49a45006f%40googlegroups.com > <https://groups.google.com/d/msgid/elasticsearch/83db33a6-0248-42eb-b913-8da49a45006f%40googlegroups.com?utm_medium=email&utm_source=footer> > . > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "elasticsearch" group. To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/CAEYi1X9ag56CtN%3D%3DnkMdNqfUfeOxgNEDTnKQ1LXtDZw%3DRuhzhw%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.