DanielLeens commented on issue #11386: URL: https://github.com/apache/seatunnel/issues/11386#issuecomment-4931023415
Thanks for writing this up so clearly. I checked the current Hudi sink path, and this does not look like a runtime bug in Hive Metastore itself. The current SeaTunnel Hudi sink does not wire `hoodie.datasource.hive_sync.*` into the write path today. The key point is: - the documented Hudi sink options only cover the SeaTunnel-owned table and write settings; - `HudiTableConfig` ignores unknown fields, so those `hoodie.datasource.hive_sync.*` keys are not rejected today; - the current Hudi write-client creation path does not pass Hive sync config or trigger a metastore sync step. So the behavior you saw is expected with the current implementation, although I agree the current docs and validation make this too easy to misunderstand. For production today, the safer approach is exactly what you suggested: let SeaTunnel write the Hudi table, then run Hudi's HiveSyncTool (or another metastore-registration step) separately. This issue is still useful because the support boundary should be made much clearer. The two reasonable follow-up directions are: 1. document explicitly that Hive Metastore sync is not supported by the current Hudi sink, or 2. add validation / warning for unsupported `hoodie.datasource.hive_sync.*` options instead of silently accepting them. If the community wants native HMS sync inside the sink itself, that should be tracked as a separate enhancement rather than treated as an already supported capability. -- 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]
