How to persist flink catalogs configuration?

[cutiechi]

How to persist flink catalogs configuration?

[cutiechi]

My sql-client-defaults.yaml:

catalogs:
  - name: hive_catalog
    type: iceberg
    catalog-type: hive
    uri: thrift://192.168.9.70:9083
    clients: 5
    property-version: 1
    warehouse: s3a://hadoop/user/hive/warehouse

[pan3793]

This feature have not implemented yet, and as far as I know, the Flink deprecated the init yaml conf and removed it totally in recent release, cc @SteNicholas @deadwind4, would you please provide more information about the next plan for init SQL?

[pan3793]

The beeline supports init sql script, does it make sense?

➜  apache-kyuubi git:(master) beeline --help
Usage: java org.apache.hive.cli.beeline.BeeLine
   ...
   -i <init file>                  script file for initialization
   ...

[a49a]

Thanks, I will research how to achieve init SQL.

[a49a]

I write CREATE CATALOG in init.sql and use initFile JDBC option. The catalog create success.

bin/beeline -u 'jdbc:hive2://foo:10009/default;initFile=init.sql'

The following is init.sql content.

CREATE CATALOG myhive WITH (
    'type' = 'hive',
    'default-database' = 'fooo',
    'hive-conf-dir' = '/etc/hive3-conf'
);

USE CATALOG myhive;

[waywtdcc]

Can Kyuubi create a global catalog on the server side? You can see the catalog when all clients open it @deadwind4