@Rocky Apache fluss catalog support ?

As of May 2026, StarRocks provides support for Apache Fluss primarily through two methods: a native Fluss Catalog (recently introduced/Beta) and an indirect but highly efficient integration via Tiered Storage (Iceberg/Paimon). 1. Native Fluss Catalog (Recent/Beta) StarRocks has introduced a native Fluss Catalog as part of its 2026 roadmap to provide direct integration with the Apache Fluss streaming storage layer. * Purpose: Automatically synchronizes database and table metadata from a Fluss cluster. * Direct Querying: It allows StarRocks to perform high-performance queries directly on Fluss tables by leveraging Fluss's native columnar storage and Arrow-based format. * Optimization: Supports predicate pushdown to the Fluss storage layer to minimize data transfer and improve sub-second query performance. 2. Indirect Integration via Data Lake Tiering (Recommended) This is currently the most mature way to analyze Fluss data using StarRocks. Apache Fluss is designed to automatically tier its data into a Data Lakehouse format like Apache Iceberg or Apache Paimon. * How it works: In Fluss, you can enable data lake tiering (e.g., setting

table.datalake.enabled = 'true'

). This causes "hot" real-time data to stay in Fluss while older "warm" data is automatically compacted and moved to Iceberg/Paimon. * StarRocks Role: You can then use StarRocks' high-performance Iceberg Catalog or Paimon Catalog to query the tiered data directly. * Union Reads: While the native Fluss catalog handles the "hot" data, StarRocks can join or union data across the Fluss Catalog and the Lakehouse Catalog (Iceberg/Paimon) to provide a unified view of both real-time and historical data. Summary of Support Feature | Status (as of May 2026) Fluss Catalog | Supported (Beta/Latest Releases) Metadata Sync | Automatic via Catalog Data Lake Tiering | Supported via Iceberg/Paimon Catalogs Query Engine | Native C++ execution for both Fluss and Lakehouse formats How to use it (Native Catalog Example):

sql
CREATE EXTERNAL CATALOG fluss_catalog
PROPERTIES (
    "type" = "fluss",
    "fluss.bootstrap.servers" = "coordinator-server:9123"
);

-- Query real-time data directly
SELECT * FROM fluss_catalog.db_name.table_name LIMIT 10;

If you are using a version of StarRocks that does not yet have the native Fluss catalog, the recommended architecture is to tier Fluss data to Iceberg and use the StarRocks Iceberg Catalog for analysis.