[jira] [Commented] (LUCENE-10272) checkindex: cross-check norms against docCount

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

    [ https://issues.apache.org/jira/browse/LUCENE-10272?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17451125#comment-17451125 ] 

Robert Muir commented on LUCENE-10272:
--------------------------------------

See attached patch, I haven't addressed the non-aborting exception case yet, but it reproduces the problem every time:

{noformat}
  2> NOTE: reproduce with: gradlew test --tests TestIndexWriterExceptions.testUpdateDocsNonAbortingException -Dtests.seed=F9962994C8B838B4 -Dtests.slow=true -Dtests.badapples=true -Dtests.locale=bs-Cyrl-BA -Dtests.timezone=Asia/Oral -Dtests.asserts=true -Dtests.file.encoding=UTF-8
   >     org.apache.lucene.index.CheckIndex$CheckIndexException: actual norm count: 0 but expected: 1
{noformat}


> checkindex: cross-check norms against docCount
> ----------------------------------------------
>
>                 Key: LUCENE-10272
>                 URL: https://issues.apache.org/jira/browse/LUCENE-10272
>             Project: Lucene - Core
>          Issue Type: Bug
>            Reporter: Robert Muir
>            Priority: Major
>         Attachments: LUCENE-10272.patch
>
>
> Spinoff from LUCENE-10271
> Despite the current crazy ghost values / elasticsearch abuse of norms, we should still cross-check the data in checkindex.
> Terms.getDocCount should equal the number of norms (we just ignore the bogus ghost values for now).
> Currently checkindex is missing really any check of norms: and it turns out that non-aborting exceptions create inconsistencies.



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

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