lasdf1234 opened a new issue, #11900: URL: https://github.com/apache/gravitino/issues/11900
### Describe the feature When Gravitino Iceberg REST Catalog (IRC) uses **JDBC** as the catalog backend and the warehouse is on **HDFS secured with Kerberos**, IRC should authenticate to HDFS the same way the Hive backend already does. Today, `IcebergCatalogUtil.loadJdbcCatalog` creates a plain `JdbcCatalog` without Kerberos login or HDFS security settings, so table metadata and data file operations fail on Kerberized clusters. ### Motivation Users deploying IRC with a JDBC catalog backend (e.g. MySQL/PostgreSQL metadata store + HDFS warehouse) on secured Hadoop clusters need Kerberos support to access warehouse paths. The Hive catalog backend already supports this via `ClosableHiveCatalog`; JDBC backend should have equivalent support. ### Describe the solution 1. Introduce `ClosableJdbcCatalog` wrapping Iceberg `JdbcCatalog`, implementing `SupportsKerberos` and `Closeable` (mirroring `ClosableHiveCatalog`). 2. Update `IcebergCatalogUtil.loadJdbcCatalog` to use `ClosableJdbcCatalog`, configure HDFS security (`hadoop.security.authentication=kerberos`), and initialize Kerberos login from catalog properties. 3. Make `JdbcCatalogWithMetadataLocationSupport` extend `ClosableJdbcCatalog`. Example IRC Kerberos configuration: ``` gravitino.iceberg-rest.authentication.type=kerberos gravitino.iceberg-rest.authentication.kerberos.principal=xxx@REALM gravitino.iceberg-rest.authentication.kerberos.keytab-uri=hdfs:///path/to/xxx.keytab ``` ### Additional context Related prior work on `branch-1.1`: https://github.com/apache/gravitino/pull/9900 -- 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]
