UpgradeIndexMergePolicy rewrites the old segments its delegate declines with its own plain OneMerge, which a wrapping delegate never sees, so IndexUpgrader (or UpgradeIndexMergePolicy) with a SoftDeletesRetentionMergePolicy inside silently purges the soft-deleted documents the retention policy must keep

Finding LFV-617Area: core index / UpgradeIndexMergePolicy and IndexUpgrader with OneMergeWrappingMergePolicy delegates (SoftDeletesRetentionMergePolicy)Priority: lowImpact: silent wrong resultsNovelty: new

Finding page for LFV-617

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

What it shows

You configure soft deletes with SoftDeletesRetentionMergePolicy("soft", MatchAllDocsQuery::new, new TieredMergePolicy()), which must keep every soft-deleted document, and upgrade a one-segment index written by Lucene 10.5.2 (10 visible plus 3 soft-deleted documents) with new IndexUpgrader(dir, thatConfig, false).upgrade(). You expect the segment to be rewritten in the current format with all 13 documents, exactly as happens when the same index has two old segments. Instead TieredMergePolicy treats the lone segment as already merged, UpgradeIndexMergePolicy rewrites it with its own 'merge remaining ones into new segment' OneMerge that the retention policy never wraps, and the 3 retained documents are purged with no error (maxDoc 13 -> 10; partial retention queries lose their documents the same way).

The test shows that UpgradeIndexMergePolicy around a SoftDeletesRetentionMergePolicy (the nesting IndexUpgrader builds) rewrites a lone old segment with its own unwrapped merge, so the soft-deleted documents the retention policy must keep are purged.

Expected vs actual

Expected maxDoc 2 after forceMerge(1) (one live plus one retained soft-deleted doc); actual 1 (expected:<2> but was:<1>), with -ea and -da.

Files

• src/org/apache/lucene/index/TestUpgradeIndexMergePolicyRetainsSoftDeletes.java (JUnit 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.
• 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=$CP:randomizedtesting-runner-2.9.2.jar:junit-4.13.2.jar:hamcrest-3.0.jar   # test-framework dependencies (Maven Central)
CP=$CP:junit-platform-console-standalone-6.1.3.jar   # JUnit Platform console launcher (Maven Central)

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:$CP" org.junit.platform.console.ConsoleLauncher \
     execute --select-class org.apache.lucene.index.TestUpgradeIndexMergePolicyRetainsSoftDeletes

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.