iamorchid opened a new issue, #11751: URL: https://github.com/apache/lucene/issues/11751
https://github.com/apache/lucene/blob/84cae4f27cfd3feb3bb42d5a9f7ce034f7a31573/lucene/core/src/java/org/apache/lucene/index/IndexWriter.java#L5075 For merge, I see that the following logic are used to initialize SegmentReader for the merge. Though segments covered in a merge won't be deleted in other concurrent merges, DV updates files could still be overridden by other concurrent flush and then be deleted after the flush performs IndexFileDeleter#checkpoint(). Though for systems like linux, this won't be a big issue (as an opened file could still be read in process even if it's been deleted). However, for windows, the delete operation would be rejected if the file is still open. So I think a more sensible way is increase the refcount in IndexFileDeleter for the files reference by the merge. ``` merge.initMergeReaders( sci -> { final ReadersAndUpdates rld = getPooledInstance(sci, true); rld.setIsMerging(); return rld.getReaderForMerge(context); }); ``` -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org For additional commands, e-mail: issues-h...@lucene.apache.org
