[ 
https://issues.apache.org/jira/browse/TEZ-902?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13921675#comment-13921675
 ] 

Siddharth Seth commented on TEZ-902:
------------------------------------

Removal from 'remaining' will likely cause a ConcurrentModificationException or 
incorrect iteration in it's current form (from the middle of a for (Input input 
: remaining) loop).

On the de-duplication, could you please include the jria number in which this 
will be removed, as well as change to a warn message - since hitting this case 
is a bug in the InputFailedEvent generation / processing.

Looking at teh Fetcher code, it tires to re-register a lot of Inputs which it 
has also reported as Failed - I believe some of that code was meant to be for 
intermediate errors - but isn't working the way it should anyway. The patch 
changes this to not retry such cases, which in the context I think is correct. 
Will open a follow up jira to get the error handling under control.


> Fetch failure issues in shuffle Input
> -------------------------------------
>
>                 Key: TEZ-902
>                 URL: https://issues.apache.org/jira/browse/TEZ-902
>             Project: Apache Tez
>          Issue Type: Bug
>            Reporter: Bikas Saha
>            Assignee: Bikas Saha
>         Attachments: TEZ-902.1.patch, TEZ-902.2.patch
>
>




--
This message was sent by Atlassian JIRA
(v6.2#6252)

Reply via email to