DanielLeens commented on issue #8918: URL: https://github.com/apache/seatunnel/issues/8918#issuecomment-4420093311
Thanks, this extra clue is useful. Based on the stack in this issue, the failure happens before CDC event processing or sink writing starts. The current path is: - `SqlServerIncrementalSourceFactory` -> `CatalogTableUtil.getCatalogTables(...)` - `AbstractJdbcCatalog.getTable(...)` - SQL Server JDBC driver throws `The "variant" data type is not supported` So this is not a Doris-specific sink path. Both `JDBC SQLServer -> Doris` and `SqlServer-CDC -> ...` can fail the same way if the runtime classpath is loading an older or conflicting SQL Server JDBC driver during schema discovery. If `datasource-hive-1.0.3-SNAPSHOT.jar` really carries an older `mssql-jdbc`, that is consistent with what you observed: SeaTunnel is still resolving source metadata when the driver blows up, so the sink choice does not matter yet. Because this thread was auto-closed long ago and the reported versions are quite old, the most helpful next step is to reopen this only if you can still reproduce it on a current maintained version, or open a fresh issue with: 1. exact SeaTunnel and SeaTunnel-Web versions 2. the full list of `mssql-jdbc` copies on the runtime classpath 3. which jar actually provides `com.microsoft.sqlserver.jdbc.SQLServerDriver` 4. whether the problem disappears after removing the conflicting jar or keeping only one SQL Server driver version That will let us distinguish a driver/classpath conflict from a core `SqlServer-CDC` bug much more reliably. If possible, please keep the follow-up in English so more maintainers can help. -- 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]
