You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@tez.apache.org by "Siddharth Seth (JIRA)" <ji...@apache.org> on 2014/11/17 21:28:35 UTC

[jira] [Updated] (TEZ-1779) Handle race condition for failure after success of a leaf vertex task

     [ https://issues.apache.org/jira/browse/TEZ-1779?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Siddharth Seth updated TEZ-1779:
--------------------------------
    Target Version/s: 0.6.0
              Labels: speculation  (was: )

> Handle race condition for failure after success of a leaf vertex task
> ---------------------------------------------------------------------
>
>                 Key: TEZ-1779
>                 URL: https://issues.apache.org/jira/browse/TEZ-1779
>             Project: Apache Tez
>          Issue Type: Bug
>            Reporter: Bikas Saha
>              Labels: speculation
>
> Currently, a leaf task is not expected to fail after success because tasks fail after success only when outputs are reported failed and noone can report a leaf vertex output as failed because noone reads it. However, as discussed in TEZ-14, there can be a case where 2 attempts are running and 1 succeeds. Then the second one is killed but it fails before getting killed. This failure would break the cannot fail after success assumption that currently is present in the code. Does the leaf vertex need special handling anyways?



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