Reordering merges (BPReorderingMergePolicy) silently break document blocks written before Lucene 9.9, because IndexWriter's reorder guard reads the unrecorded LeafMetaData#hasBlocks flag as 'no blocks'

Finding LFV-429Area: index / IndexWriter merge-time reordering (OneMerge#reorder guard; misc BPReorderingMergePolicy)Priority: lowImpact: data corruptionNovelty: new

Finding page for LFV-429

AI-assisted finding, confirmed by automated checks, not yet triaged by a human Lucene committer. This site files nothing upstream.

What it shows

You have an index with document blocks (children followed by their parent, added with IndexWriter#addDocuments) that was written by Lucene 9.8 or earlier, open it with a current Lucene, and force-merge it through BPReorderingMergePolicy. You expect the merge to leave the blocks alone, because since 9.11 IndexWriter lets a merge reorder documents only when there are no blocks or a parent field identifies them, and such an old index cannot have a parent field. Instead the guard trusts LeafMetaData#hasBlocks, which was never recorded before 9.9 and reads false for those segments, so the reorderer renumbers the documents one by one and interleaves the blocks, with no exception and a clean CheckIndex, while the same documents written by 9.9.2 are left intact.

The test shows that a reordering forceMerge (BPReorderingMergePolicy) tears apart document blocks written by Lucene 9.8.0, because IndexWriter's reorder guard reads the unrecorded pre-9.9 LeafMetaData#hasBlocks as "no blocks" (fixture blocks.9.8.0.zip: 3 child+parent blocks in 2 segments).

Expected vs actual

Expected the merged segment's docs, in pairs, to be the blocks [c0p0, c1p1, c2p2]; actual: [c0c1, c2p0, p1p2] (doc order c0 c1 c2 p0 p1 p2), with -ea and -da.

Files

src/org/apache/lucene/misc/index/TestReorderPre99Blocks.java (JUnit test)
fixture-src/MakePre99BlockIndex.java (writes the fixture the test reads, as Run it says; not compiled with the test)

Run it

Tested on Lucene main e1faeba08bde (2026-11-02): still reproduces.
Tested on Lucene main 645d4fe14a20 (2026-10-28): reproduced.
Recorded outcome of run 1 on Lucene main 645d4fe14a20: 1 test(s) run, 1 failed.
Fixture: the test reads blocks.9.8.0.zip, an index of 3 blocks (a child and its parent) in 2 segments written by Lucene 9.8.0, from its package on the class path. This site ships no binary index files, so the recipe first makes it with fixture-src/MakePre99BlockIndex.java and lucene-core-9.8.0.jar (Maven Central) and puts it in src/, which the run puts on the class path; without it the test fails for another reason.
How to read the outcome: a failing test is the finding reproducing (the failure message states expected against actual).
Seeds: a rerun with the same seed replays the same random choices. Assertions off (-da -dsa -Dtests.asserts=false) is the production-like setting.

# Lucene checkout at the tested commit, built with: ./gradlew --no-daemon assemble   (JDK 25)
LUCENE=/path/to/lucene; V=12.0.0-SNAPSHOT
CP=$LUCENE/lucene/core/build/libs/lucene-core-$V.jar
CP=$CP:$LUCENE/lucene/codecs/build/libs/lucene-codecs-$V.jar
CP=$CP:$LUCENE/lucene/test-framework/build/libs/lucene-test-framework-$V.jar
CP=$CP:$LUCENE/lucene/backward-codecs/build/libs/lucene-backward-codecs-$V.jar
CP=$CP:$LUCENE/lucene/misc/build/libs/lucene-misc-$V.jar
CP=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar   # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar   # JUnit Platform console launcher (Maven Central)

# fixture blocks.9.8.0.zip, written by the Lucene 9.8.0 release: lucene-core-9.8.0.jar (Maven Central)
javac -d fixture-classes -cp lucene-core-9.8.0.jar fixture-src/MakePre99BlockIndex.java
java -cp fixture-classes:lucene-core-9.8.0.jar MakePre99BlockIndex fixture-index
(cd fixture-index && rm -f write.lock && zip -q -X ../src/org/apache/lucene/misc/index/blocks.9.8.0.zip *)

javac -d classes -cp "$CP" $(find src -name '*.java')

# run 1: assertions on, seed DEADBEEF
java -ea -esa --add-modules jdk.incubator.vector --enable-native-access=ALL-UNNAMED -Dtests.seed=DEADBEEF \
     -cp "classes:src:$CP" org.junit.platform.console.ConsoleLauncher \
     execute --select-class org.apache.lucene.misc.index.TestReorderPre99Blocks

Gradle alternative: copy the test into the matching module's src/test/ tree of the checkout and run ./gradlew -p lucene/<module> test --tests <class> -Ptests.seed=<seed> (add -Ptests.asserts=false for a run without assertions).

Feedback

Open an issue in lucene-ai-bug-hunt. Please do not open issues or pull requests in apache/lucene about this finding unless a Lucene committer agrees it should be filed.

Snapshot of 2026-11-04. AI-assisted; not yet triaged by a human Lucene committer. Feedback: open an issue in lucene-ai-bug-hunt.