bvaradar opened a new pull request, #18227: URL: https://github.com/apache/hudi/pull/18227
### Describe the issue this Pull Request addresses When using `mode=jdbc`, HoodieHiveSyncClient performs metadata operations (tableExists, getTableSchema, getAllPartitions, etc.) via the Thrift IMetaStoreClient, which is incompatible with HMS 4.x due to renamed Thrift methods (e.g. get_table → get_table_req). This patch introduces JDBCBasedMetadataOperator, which executes equivalent metadata queries via the existing JDBC connection to HiveServer2 using standard HiveQL (SHOW TABLES, DESCRIBE, ALTER TABLE, etc.). The fallback is automatic and backward compatible: - On first TApplicationException from the Thrift client, the thriftIncompatible flag is set and all subsequent metadata operations are routed through the JDBC operator. - No new configuration required — existing mode=jdbc with a valid jdbcurl is sufficient. - Default behavior (mode=hms) is completely unchanged to avoid any regression completely. Tested against HMS 4.2.0 with HiveServer2, Spark 4.0, Hudi 1.1.1 (patch) in local docker setup. ### Summary and Changelog [HUDI-9064] Support HMS 4.x in JDBC sync mode via automatic Thrift fallback ### Impact Provides a path towards having Hudi Hive Sync support for Hive 4.x ### Risk Level Low ( backwards compatible change - only kicks in when there is failure) ### Documentation Update None ### Contributor's checklist - [ ] Read through [contributor's guide](https://hudi.apache.org/contribute/how-to-contribute) - [ ] Enough context is provided in the sections above - [ ] Adequate tests were added if applicable -- 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]
