You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@lucene.apache.org by "Robert Muir (Jira)" <ji...@apache.org> on 2022/05/16 09:50:00 UTC
[jira] [Commented] (LUCENE-10573) Improve stored fields bulk merge for degenerate O(n^2) merges
[ https://issues.apache.org/jira/browse/LUCENE-10573?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17537437#comment-17537437 ]
Robert Muir commented on LUCENE-10573:
--------------------------------------
Fix TieredMergePolicy, don't hack around its shortcomings in the stored fields, thats awful.
> Improve stored fields bulk merge for degenerate O(n^2) merges
> -------------------------------------------------------------
>
> Key: LUCENE-10573
> URL: https://issues.apache.org/jira/browse/LUCENE-10573
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Adrien Grand
> Priority: Minor
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Spin-off from LUCENE-10556.
> For small merges that are below the floor segment size, TieredMergePolicy may merge segments that have vastly different sizes, e.g. one 10k-docs segment with 9 100-docs segments.
> While we might be able to improve TieredMergePolicy (LUCENE-10569), there are also improvements we could make to stored fields, such as bulk-copying chunks of the first segment until the first dirty chunk. In this scenario where segments keep being rewritten, this would help significantly.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org