[ https://issues.apache.org/jira/browse/DRILL-4768?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15371499#comment-15371499 ]
ASF GitHub Bot commented on DRILL-4768: --------------------------------------- Github user vkorukanti commented on the issue: https://github.com/apache/drill/pull/543 Change seems reasonable to me. One thing: is the reconnect() call not closing the existing connection? If yes why are we leaking the connections? I think there is one more place we need to do the same thing [1]. May be it is better move this to DrillHiveMetastoreClient. [1] https://github.com/apache/drill/blob/master/contrib/storage-hive/core/src/main/java/org/apache/drill/exec/store/hive/schema/HiveDatabaseSchema.java#L94 > Drill may leak hive meta store connection if hive meta store client call hits > error > ----------------------------------------------------------------------------------- > > Key: DRILL-4768 > URL: https://issues.apache.org/jira/browse/DRILL-4768 > Project: Apache Drill > Issue Type: Bug > Components: Storage - Hive > Reporter: Jinfeng Ni > Assignee: Jinfeng Ni > > We are seeing one drillbit creates hundreds of connections to hive meta > store. This indicates that drill is leaking those connection, and did not > close those connections properly. When such leaking happens, it may prevent > other applications from connecting to hive meta store. > It seems one cause of leaking connection happens when hive meta store client > call hits exception. > -- This message was sent by Atlassian JIRA (v6.3.4#6332)