DanielLeens commented on issue #12267: URL: https://github.com/apache/seatunnel/issues/12267#issuecomment-5714102880
Thank you for the full rerun trace. **Classification: E / Postgres-CDC HighGo source initialization.** The new trace changes the immediate diagnosis. The job reaches the standard PostgreSQL JDBC connection and table discovery, then fails before replication starts in `PostgresIncrementalSource.tableChanges()`: two `CREATE` records for the same `highgo.testdb.test_a` are collected into a map and `Collectors.toMap` throws `Duplicate key`. `REPLICA IDENTITY FULL`, the CDC port, and a driver substitution cannot affect that initialization failure, so please do not keep changing those variables for this particular run. The next step is to isolate why the same table is emitted twice. Please provide a minimal sanitized source block containing `database-names`, `schema-names`, `table-names`, `schema-changes.enabled`, startup mode, and any Debezium properties, together with the smallest HighGo database that reproduces the duplicate `CREATE` records. A correct fix needs a regression that feeds duplicate table-change records through this path: identical definitions may be deduplicated deterministically, but divergent definitions must fail with a descriptive conflict rather than silently choosing one. After source initialization succeeds, we can separately validate replica identity and snapshot/CDC behavior. No connector patch is proposed yet. -- 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]
