[jira] [Commented] (LUCENE-7162) geo random tests still have reproducibility issues

["Robert Muir (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/LUCENE-7162?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15221743#comment-15221743 ] 

Robert Muir commented on LUCENE-7162:
-------------------------------------

I wonder if the issue has to do with concurrent merge scheduler? Its just my only thought, as threads could influence the structure of the BKD tree and how it is traversed.

I think its ok to set serial merge scheduler always for geo tests? We can test thread safety of e.g. KDtree indexing elsewhere, with simpler tests.

> geo random tests still have reproducibility issues
> --------------------------------------------------
>
>                 Key: LUCENE-7162
>                 URL: https://issues.apache.org/jira/browse/LUCENE-7162
>             Project: Lucene - Core
>          Issue Type: Bug
>            Reporter: Robert Muir
>         Attachments: mychanges.patch
>
>
> I am trying to remove some tricky test leniency/fix rounding issues in our geo fields, but it is frustrating when failures don't always reproduce.
> Often they now do, but sometimes they still do not.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

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