[
https://issues.apache.org/jira/browse/FLINK-29549?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17616862#comment-17616862
]
Samrat Deb commented on FLINK-29549:
------------------------------------
*There are two Use Cases:*
# Using Glue as a catalog for Flink HiveSource. Currently the HiveShim
directly create HiveMetastoreClient (HMS).
Instead, it will check the configured factory class (HMS or Glue) introduced
by https://issues.apache.org/jira/browse/HIVE-12679 and based on that
create the IMetaStoreClient.
# Using Glue as an external catalog for Flink. This might require a new
*GlueCatalog* similar to HiveCatalog/JdbcCatalog
([https://nightlies.apache.org/flink/flink-docs-master/docs/dev/table/catalogs/]).
which will directly use Glue APIs without any Hive Dependencies.
> Flink sql to add support of using AWS glue as metastore
> -------------------------------------------------------
>
> Key: FLINK-29549
> URL: https://issues.apache.org/jira/browse/FLINK-29549
> Project: Flink
> Issue Type: Improvement
> Components: Connectors / Common, Connectors / Hive
> Reporter: Samrat Deb
> Priority: Major
>
> Currently , Flink sql hive connector support hive metastore as hardcoded
> metastore-uri.
> It would be good if flink provide feature to have configurable metastore (eg.
> AWS glue).
> This would help many Users of flink who uses AWS
> Glue([https://docs.aws.amazon.com/glue/latest/dg/start-data-catalog.html]) as
> their common(unified) catalog and process data.
> cc [~prabhujoseph]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
