You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by "Matt McCline (Jira)" <ji...@apache.org> on 2021/06/09 19:26:00 UTC

[jira] [Created] (HIVE-25228) Thrift CLI Service Protocol: Watch for lack of interest by client and kill queries faster

Matt McCline created HIVE-25228:
-----------------------------------

             Summary: Thrift CLI Service Protocol: Watch for lack of interest by client and kill queries faster
                 Key: HIVE-25228
                 URL: https://issues.apache.org/jira/browse/HIVE-25228
             Project: Hive
          Issue Type: Improvement
            Reporter: Matt McCline
            Assignee: Matt McCline


CONSIDER: Have Hive Server 2 monitor operations (queries) for continuing client interest. If a client does not ask for status every 15 seconds, then automatically kill a query and release its txn locks and job resources.

 

Users will experience queries cleaning up much faster (15 to 30 seconds instead of minutes and possibly many minutes) when client communication is lost. Cleaning up those queries prevents other queries from being blocked on EXCLUSIVE txn locks and blocking of scheduling of their queries including retries of the original query. Today, users can get timeouts when they retry a query that got a connection error causing understandably upset users.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)