[
https://issues.apache.org/jira/browse/HIVE-13858?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15304294#comment-15304294
]
Siddharth Seth edited comment on HIVE-13858 at 5/27/16 9:08 PM:
----------------------------------------------------------------
Updated patch with comments addressed.
Throwing an InterruptedException does not clear the InterruptStatus (-catching
it probably does-. Don't think catching it does anything either.). Going by
most recommendations - I've removed the code to clear the interrupt status.
Also, propagating InterruptedException all the way out of the Hive processor.
was (Author: sseth):
Updated patch with comments addressed.
Throwing an InterruptedException does not clear the InterruptStatus (catching
it probably does). Going by most recommendations - I've removed the code to
clear the interrupt status. Also, propagating InterruptedException all the way
out of the Hive processor.
> LLAP: A preempted task can end up waiting on completeInitialization if some
> part of the executing code suppressed the interrupt
> -------------------------------------------------------------------------------------------------------------------------------
>
> Key: HIVE-13858
> URL: https://issues.apache.org/jira/browse/HIVE-13858
> Project: Hive
> Issue Type: Bug
> Affects Versions: 2.0.0
> Reporter: Siddharth Seth
> Assignee: Siddharth Seth
> Priority: Critical
> Labels: llap
> Attachments: HIVE-13858.01.patch, HIVE-13858.02.patch,
> HIVE-13858.03.patch
>
>
> An interrupt along with a HiveProcessor.abort call is made when attempting to
> preempt a task.
> In this specific case, the task was in the middle of HDFS IO - which
> 'handled' the interrupt by retrying. As a result the interrupt status on the
> thread was reset - so instead of skipping the future.get in
> completeInitialization - the task ended up blocking there.
> End result - a single executor slot permanently blocked in LLAP. Depending on
> what else is running - this can cause a cluster level deadlock.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
