Manticore Search 15.1.0 has been released. This maintenance release includes bug fixes and minor improvements.

Breaking Changes

MCL Update to 9.0.0

Version 15.0.0 updated the MCL (Manticore Columnar Library) requirement to 9.0.0, bringing significant improvements:

• Uncompressed float-vector storage
• Block-size change for KNN vectors
• Non-buffered reads

⚠️ Important: This update changes the columnar storage data format. Older columnar library versions won't be able to read data created with MCL 9.0.0, but the new version can still read your existing columnar tables without any issues. The Manticore Search / MCL API version has been increased.

If you follow the official installation guide , you don't need to worry about this as the correct MCL version will be installed automatically.

Recommended Versions

• MCL: 9.0.0
• Manticore Buddy: 3.40.2

New Features and Improvements

This release includes the following improvements:

• Improved disk chunk flushing log records by breaking the total time into clearer parts
• Support for arbitrary filter expressions in JOIN ON clauses (not just equality comparisons)
• Improved handling of LOCK TABLES / UNLOCK TABLES: write locks now return warnings instead of errors, read locks are correctly shown in SHOW LOCKS
• Trimmed extra whitespace and new lines in JSON payloads when logging query requests
• Updated Manticore Buddy requirement to 3.40.2, which includes fuzzy-search support for the /sql endpoint and various fixes

Bug Fixes

This release includes bug fixes from previous versions, including:

• Fixed regression where the native Windows service failed to start when installed with a custom configuration path
• Fixed regression where HTTP /sql responses incorrectly used Content-Type: text/html instead of application/json
• Fixed crash in KNN vector search by skipping searches on empty HNSW indexes
• Fixed KNN oversampling behavior: no longer compute rescored KNN distance when rescoring is not requested
• Fixed issue where grouping via GROUP BY / FACET on attributes generated with JSON-to-attribute mapping failed to work
• Fixed MATCH() parsing so empty groups like () no longer trigger an error
• Fixed JSON-attribute faceting: aliasing a JSON array in a FACET clause now behaves the same as faceting on the array directly
• Fixed handling of "joined fields" in SQL-based sources so that an "end marker" gets set correctly when iterating joined hits
• Fixed multi-query requests via /cli_json that include semicolons (e.g., combining SQL statements)
• Enhanced update validation to check that updated attributes do not conflict with full-text fields
• Fixed missing error messages for HTTP UPDATE and DELETE requests on distributed tables handled via agents
• Fixed crash when using knn_dist() in a custom ranker expression with KNN+MATCH queries
• Fixed table-status counters misleadingly labeled with "_sec" (e.g., query_time_1min) to correctly report seconds instead of milliseconds
• Fixed SHOW INDEX <name> STATUS search timing stats to match the values calculated from query logs
• Fixed internal error when using distributed tables with persistent_connections_limit
• Fixed bug where some commands executed via MySQL client 9 caused a "unexpected $undefined near '$$'" error in the query log
• Updated backup documentation for mysqldump replication mode and fixed several broken manual links
• Updated documentation and tests related to supporting Logstash 9.2
• Fixed tests on Windows so "ubertests" no longer leave minidumps after completion
• Fixed Windows build by correcting a type mismatch in binlog.cpp
• Fixed autocomplete parsing so queries with escaped quotes are handled correctly
• Documented the internal 32-bit mask used when counting ranking factors for Phrase/Proximity/NEAR operators

Compatibility

Manticore Search 15.1.0 maintains strong backward compatibility with important considerations:

• Fully compatible with existing data and queries
• New MCL 9.0.0 can read existing columnar tables without issues
• MCL 9.0.0 updates the columnar storage data format. The new version can still read your existing data, but older MCL versions cannot read columnar data created with MCL 9.0.0. If you are not using columnar storage, there is nothing to worry about.

To upgrade, follow the installation guide .

Need help or want to connect?

• Join our Slack
• Visit the Forum
• Report issues or suggest features on GitHub
• Email us at [email protected]

For full details, see the Changelog .