diqiu50 opened a new issue, #13844: URL: https://github.com/apache/gravitino/issues/13844
### What would you like to be improved? When the Gravitino Spark connector is configured at the cluster level (for example, `spark.plugins=org.apache.gravitino.spark.connector.plugin.GravitinoSparkPlugin` in `spark-defaults.conf`), every Spark job loads it, including jobs that never use a Gravitino catalog. `GravitinoDriverPlugin#init` currently fails fast whenever: - `spark.sql.gravitino.uri` or `spark.sql.gravitino.metalake` is missing or blank, - the Gravitino server can't be reached, or loading catalogs fails, - the client configuration (such as auth settings) is invalid. Any of these makes the `SparkContext` fail to start, so the job fails even though it doesn't depend on Gravitino at all. ### How should we improve? When the plugin can't initialize, it should log a clear warning and skip registering Gravitino catalogs. The Spark job should keep running. A job only needs to fail if it actually accesses a Gravitino catalog. If some users want the current fail-fast behavior, we could add an option to keep it. -- 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]
