pan3793 opened a new pull request, #58573: URL: https://github.com/apache/spark/pull/58573
### What changes were proposed in this pull request? Flip the default of `spark.sql.thriftServer.catalogMetadata.enabled` from `true` to `false`. Tests in `SparkMetadataOperationSuite` are updated accordingly: pre-existing tests assert the legacy behavior again, and SPARK-57518 tests that exercise the catalog-aware path now enable the conf explicitly. ### Why are the changes needed? The catalog-aware metadata operations shipped in #56627 are incomplete: getTables/getColumns do not route through DSv2 catalogs, and getCatalogs only lists already-loaded catalogs. There are also known correctness issues with the getSchemas implementation, documented in #57468, which has seen no active engagement from the author. Since 4.3.0 is not released yet, it is safer to keep the legacy behavior by default and let users opt in until the follow-up work lands. ### Does this PR introduce _any_ user-facing change? Within unreleased branches only: JDBC metadata operations (getCatalogs, getSchemas, getTables, getColumns) fall back to the legacy behavior unless the conf is explicitly enabled. ### How was this patch tested? Updated `SparkMetadataOperationSuite`; all 25 tests pass (`build/sbt -Phive-thriftserver 'hive-thriftserver/testOnly *SparkMetadataOperationSuite'`). ### Was this patch authored or co-authored using generative AI tooling? Generated-by: GLM 5.3 -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
