You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@tajo.apache.org by "Jihoon Son (JIRA)" <ji...@apache.org> on 2014/12/02 02:46:12 UTC

[jira] [Commented] (TAJO-1218) Implement straggler detector and the block list

    [ https://issues.apache.org/jira/browse/TAJO-1218?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14230822#comment-14230822 ] 

Jihoon Son commented on TAJO-1218:
----------------------------------

Hi [~hyunsik].
It is a good idea.
I think that we also need to recover blacklisted nodes when they become normal.

> Implement straggler detector and the block list
> -----------------------------------------------
>
>                 Key: TAJO-1218
>                 URL: https://issues.apache.org/jira/browse/TAJO-1218
>             Project: Tajo
>          Issue Type: Improvement
>            Reporter: Hyunsik Choi
>              Labels: failure-handling, fault-tolerance
>
> A straggler is a machine that takes an unusually longer time to complete tasks than other machines do. Straggler is usual in large-scale distributed systems. Earlier straggler detection and handling can significantly reduce query response times. So, we need to add the straggler detector. Detected stragglers should be added to the black list, and they should not used for processing.



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