[
https://issues.apache.org/jira/browse/DRILL-3448?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14623074#comment-14623074
]
Khurram Faraaz commented on DRILL-3448:
---------------------------------------
When I kill a Drillbit where no fragments were assigned to it for a long
running query, I see the below exception message. Can you please fix the error
message ?
{code}
java.lang.RuntimeException: java.sql.SQLException: SYSTEM ERROR:
ForemanException: One more more nodes lost connectivity during query.
Identified nodes were [].
{code}
> typo in QueryManager.DrillbitStatusListener will cause the Foreman to hang if
> a Drillbit dies
> ---------------------------------------------------------------------------------------------
>
> Key: DRILL-3448
> URL: https://issues.apache.org/jira/browse/DRILL-3448
> Project: Apache Drill
> Issue Type: Bug
> Components: Execution - Flow
> Affects Versions: 1.0.0
> Reporter: Deneche A. Hakim
> Assignee: Sudheesh Katkam
> Priority: Critical
> Fix For: 1.2.0
>
> Attachments: DRILL-3448.1.patch.txt
>
>
> at the end of DrillbitStatusListener.drillbitUnregistered() there is if block:
> {code}
> if (!atLeastOneFailure) {
> logger.warn("...");
> stateListener.moveToState(QueryState.FAILED,
> new ForemanException(...));
> }
> {code}
> this will basically fail the query if the drillbit DIDN'T contain any
> fragment for this Foreman, which in fact should be the inverse.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
