drill-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Deneche A. Hakim (JIRA)" <j...@apache.org>
Subject [jira] [Created] (DRILL-3052) canceling a fragment executor before it starts running will cause the Foreman to wait indefinitely for a terminal message from that fragment
Date Wed, 13 May 2015 03:57:59 GMT
Deneche A. Hakim created DRILL-3052:
---------------------------------------

             Summary: canceling a fragment executor before it starts running will cause the
Foreman to wait indefinitely for a terminal message from that fragment
                 Key: DRILL-3052
                 URL: https://issues.apache.org/jira/browse/DRILL-3052
             Project: Apache Drill
          Issue Type: Bug
            Reporter: Deneche A. Hakim
            Assignee: Deneche A. Hakim
             Fix For: 1.0.0


When a Foreman receives a cancellation from the client it will cancel all it's fragments and
go into a CANCELLATION_REQUEST state. Each fragment will cancel it's work, close it's resources
and report a terminal state to the Foreman (either CANCELLED or FAILED). The Foreman will
wait for all fragments to report before sending a terminal message to the client.
If a fragment is cancelled before it started running (similar to DRILL-2878) it will close
it's resources but it will never send a terminal state to the foreman which will cause the
foreman to wait indefinitely.

TestDrillbitResilience.cancelBeforeAnyResultsArrive() exposes this problem.



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

Mime
View raw message