Skip to content

CNDB-18275: Port CNDB-18238: Update JVector to hotfix regression in recall. The pruning option is ignored and no matter what a user sets pruning will be always pointing to FALSE (#2479)#2485

Open
ekaterinadimitrova2 wants to merge 1 commit into
mainfrom
c18275-main

Conversation

@ekaterinadimitrova2

Copy link
Copy Markdown

...
4.0.0-rc.8 has a regression in recall when we run queries with pruning set to true. This is the default behavior in our vector queries - we use pruning.
...
This hotfix changes the defaults of pruning being set to False. JVector testing shows that no pruning produces similar latencies but better recall, so this was the simplest way to fix recall in a hotfix release in JVector.

The pruning option is ignored and no matter what a user sets pruning will be always pointing to FALSE.

NOTE: It is a known issue that currently VectorCompactionTest and VectorSiftSmallTest are timing out in CI so I reran them locally and they still pass.

What is the issue

...

What does this PR fix and why was it fixed

...

…ecall. The pruning option is ignored and no matter what a user sets pruning will be always pointing to FALSE (#2479)

...
4.0.0-rc.8 has a regression in recall when we run queries with pruning
set to true. This is the default behavior in our vector queries - we use
pruning.

...

This hotfix changes the defaults of pruning being set to False. JVector
testing shows that no pruning produces similar latencies but better
recall, so this was the simplest way to fix recall in a hotfix release
in JVector.

The pruning option is ignored and no matter
what a user sets pruning will be always pointing to FALSE.

NOTE: It is a known issue that currently `VectorCompactionTest` and
`VectorSiftSmallTest` are timing out in CI so I reran them locally and
they still pass.
@github-actions

github-actions Bot commented Jun 23, 2026

Copy link
Copy Markdown

Checklist before you submit for review

  • This PR adheres to the Definition of Done
  • Make sure there is a PR and ticket in the CNDB project updating the Converged Cassandra version
  • Use NoSpamLogger for log lines that may appear frequently in the logs
  • Verify test results on Butler
  • Test coverage for new/modified code is > 80%
  • Proper code formatting
  • Proper title for each commit staring with the project-issue number, like CNDB-1234
  • Each commit has a meaningful description
  • Each commit is not very long and contains related changes
  • Renames, moves and reformatting are in distinct commits
  • All new files should contain the IBM copyright header instead of the Apache License one (no DataStax copyright any longer)

@plpesvc-ds

Copy link
Copy Markdown

❌ Build ds-cassandra-pr-gate/PR-2485 rejected by Butler


3 regressions found
See build details here


Found 3 new test failures

Test Explanation Runs Upstream
o.a.c.distributed.test.AbortedQueryLoggerTest.testLogsReadMetrics REGRESSION 🔴 0 / 30
o.a.c.index.sai.cql.VectorCompaction100dTest.testCompactionWithEnoughRowsForPQAndDeleteARow[version=ec enableNVQ=true] REGRESSION 🔴 0 / 30
o.a.c.index.sai.cql.VectorSiftSmallTest.testRerankKZeroOrderMatchesFullPrecisionSimilarity[ed false] REGRESSION 🔴 0 / 30

Found 2 known test failures

@ekaterinadimitrova2

Copy link
Copy Markdown
Author

All tests marked as regressions are already known timeouts in CI and the tests pass locally.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants