[
https://issues.apache.org/jira/browse/DRILL-2245?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14329839#comment-14329839
]
Abhishek Girish commented on DRILL-2245:
----------------------------------------
Hit an issue related to one described here - On a large dataset, canceling a
select * query (without limits) causes subsequent queries to hang. Will need to
verify the same once this has been merged.
> Improve Drillbit resiliency
> ---------------------------
>
> Key: DRILL-2245
> URL: https://issues.apache.org/jira/browse/DRILL-2245
> Project: Apache Drill
> Issue Type: Bug
> Components: Execution - Flow
> Reporter: Chris Westin
> Assignee: Chris Westin
> Attachments: DRILL-2245.1.patch.txt
>
>
> There are paths through WorkManager and Foreman that will allow a Drillbit to
> get wedged if there are failures. This can happen because events delivered
> via other threads will get stuck if Foreman throws an exception before it is
> accepting external events.
> This is a call to do some general cleanup here and make drillbits more robust
> so that they don't hang when there are query failures.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
