You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@hive.apache.org by "Sergey Shelukhin (JIRA)" <ji...@apache.org> on 2016/09/12 23:23:20 UTC

[jira] [Commented] (HIVE-14740) LLAP DynamicServiceInstasnce.isAlive does not seem to work as it should

    [ https://issues.apache.org/jira/browse/HIVE-14740?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15485608#comment-15485608 ] 

Sergey Shelukhin commented on HIVE-14740:
-----------------------------------------

There seem to be 3 things... node set from which dead nodes should be removed but aren't. Active instance set, that was only used in specific-node scheduling, but not in random scheduling, until recently. And then there's isAlive.
I wonder if so many ways to find if node is alive are needed.

> LLAP DynamicServiceInstasnce.isAlive does not seem to work as it should
> -----------------------------------------------------------------------
>
>                 Key: HIVE-14740
>                 URL: https://issues.apache.org/jira/browse/HIVE-14740
>             Project: Hive
>          Issue Type: Bug
>            Reporter: Siddharth Seth
>            Priority: Critical
>
> Follow up from HIVE-14608. Is used in multiple places in LlapTaskScheduler. Critical to validate the broken functionality does not cause problems in the scheduler.



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