You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@drill.apache.org by "Sudheesh Katkam (JIRA)" <ji...@apache.org> on 2016/07/11 20:45:11 UTC

[jira] [Created] (DRILL-4774) Replace NonCloseableHiveClientWithCaching

Sudheesh Katkam created DRILL-4774:
--------------------------------------

             Summary: Replace NonCloseableHiveClientWithCaching
                 Key: DRILL-4774
                 URL: https://issues.apache.org/jira/browse/DRILL-4774
             Project: Apache Drill
          Issue Type: Bug
            Reporter: Sudheesh Katkam


Likely the intention was that a NonCloseableHiveClientWithCaching object will be used as long as the Drill process lives.

(1) Given how StoragePluginRegistryImpl works (see createOrUpdate), HiveSchemaFactory leaks a NonCloseableHiveClientWithCaching every time the hive storage plugin is updated (because close is not invoked).

(2) I am not sure if the HiveMetastore and HiveMetaStoreClient communication path has a heartbeat mechanism. Otherwise when Drill is shutdown, from metastore point of view, at least one connection to the metastore is not closed.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)