Scan scroll query - performance impact while not indexing


(Roman Margolis) #1

According to the guide, scan scroll requests take additional toll on the cluster since they have to keep older segments alive for the snapshot behavior they support.

Normally, the background merge process optimizes the index by merging together smaller segments to create new bigger segments, at which time the smaller segments are deleted. This process continues during scrolling, but an open search context prevents the old segments from being deleted while they are still in use. This is how Elasticsearch is able to return the results of the initial search request, regardless of subsequent changes to documents.

I'm curious, is scan scrolling considered a relatively cheap operation when performed on a sealed index, or index that no longer receives writes? Also, what happens if the index is optimized to a single segment?

Many thanks


(system) #2