[
https://issues.apache.org/jira/browse/LUCENE-10088?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17411354#comment-17411354
]
Adrien Grand commented on LUCENE-10088:
---------------------------------------
Here are some more details on what I'm seeing:
- The test starts stressing concurrent updates.
- At some point RandomIndexWriter sets maxBufferedDocs to 15, forcing frequent
flushes.
- One thread calls IndexWriter#getReader, a full flush starts, which sets
{{DocumentsWriterFlushControl#fullFlush}} as true as a side-effect.
- While the full flush is still in progress, the indexing threads keep adding
new documents.
- Given the low maxBufferedDocs, DWPTs are regularly marked for flush. However
since {{DocumentsWriterFlushControl#fullFlush}} is true, indexing threads don't
help flushing these DWPTs, they keep accumulating in
{{DocumentsWriterFlushControl#blockedFlushes}}.
- We hit the limit on the number of open files before
{{DocumentsWriterFlushControl#finishFullFlush}} had a chance to prune all the
blocked flushes.
> Too many open files in
> TestIndexWriterMergePolicy.testStressUpdateSameDocumentWithMergeOnGetReader
> --------------------------------------------------------------------------------------------------
>
> Key: LUCENE-10088
> URL: https://issues.apache.org/jira/browse/LUCENE-10088
> Project: Lucene - Core
> Issue Type: Improvement
> Reporter: Michael McCandless
> Priority: Major
>
> [This build
> failure|https://ci-builds.apache.org/job/Lucene/job/Lucene-NightlyTests-main/386/]
> reproduces for me. I'll try to dig.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
