lucene-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "David Smiley (Jira)" <j...@apache.org>
Subject [jira] [Commented] (SOLR-13875) Check why LBSolrClient thinks it has to sort on _docid_
Date Fri, 01 Nov 2019 13:41:00 GMT

    [ https://issues.apache.org/jira/browse/SOLR-13875?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16964831#comment-16964831
] 

David Smiley commented on SOLR-13875:
-------------------------------------

It's awfully suspicious to me that {{_docid_}} sort is slow no matter how many docs.  In
theory, it's the fastest sort possible!  It's basically no-sorting.

> Check why LBSolrClient thinks it has to sort on _docid_
> -------------------------------------------------------
>
>                 Key: SOLR-13875
>                 URL: https://issues.apache.org/jira/browse/SOLR-13875
>             Project: Solr
>          Issue Type: Improvement
>      Security Level: Public(Default Security Level. Issues are Public) 
>            Reporter: Erick Erickson
>            Priority: Major
>
> Related to SOLR-13874. LBSolrClient issues a query that sorts on \_docid\_. This can
take over 12 seconds on a 1B doc replica. Is the sorting necessary at all? If it's removed,
what are the effects on performance? Of we change SOLR-13874, we can close this.
> Frankly I suspect that no matter what, there have to be over 1B comparisons done and
that's what's taking the time, but I wanted to call this out explicitly.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@lucene.apache.org
For additional commands, e-mail: issues-help@lucene.apache.org


Mime
View raw message