You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@drill.apache.org by "Ramana Inukonda Nagaraj (JIRA)" <ji...@apache.org> on 2015/03/24 23:04:52 UTC

[jira] [Commented] (DRILL-2542) Killing a drillbit when a query is running results in hanging query

    [ https://issues.apache.org/jira/browse/DRILL-2542?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14378748#comment-14378748 ] 

Ramana Inukonda Nagaraj commented on DRILL-2542:
------------------------------------------------

Looks like the behavior is different depending on which drillbits are killed- Foreman or otherwise.
In case of non foreman node the query hangs around with CANCELLATION_REQUESTED message.

In case of the foreman node drillbit process being killed there seems to be no change in the state of the query.  From the JDBC client or from the profiles page where its listed as RUNNING.

> Killing a drillbit when a query is running results in hanging query
> -------------------------------------------------------------------
>
>                 Key: DRILL-2542
>                 URL: https://issues.apache.org/jira/browse/DRILL-2542
>             Project: Apache Drill
>          Issue Type: Bug
>          Components: Execution - Flow
>    Affects Versions: 0.8.0
>            Reporter: Ramana Inukonda Nagaraj
>            Assignee: Chris Westin
>
> git.commit.id.abbrev=1b9be30
> Tried the following test- Have a 8 node cluster running TPCH 10 on SF 100 data. Killed one of the drillbits- Query does not fail and hangs without progress until manually killed.



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