[
https://issues.apache.org/jira/browse/TEZ-2778?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Rajesh Balamohan updated TEZ-2778:
----------------------------------
Attachment: dag_1439860407967_0579_1.svg
DAG.png
lgtm. +1. Very minor comment
- Should TaskAttemptInfo getLastDataEventTime()/getLastDataEventSourceTA() can
be clubbed to be a single method which would return DataDependencyEvent to
avoid duplication?
Created TEZ-2782 which is not directly related to this patch, but can throw NP
when trying to compute average execution time.
Also tried it on hive tpcds query_95 by manually introducing errors in a
specific node (to simulate source failure). Attaching the graphs as well, and
it works fine. "Reducer 5" had 4 attempts for a specific task and all of them
were on the critical path.
> Improvements to handle read errors - part 2
> -------------------------------------------
>
> Key: TEZ-2778
> URL: https://issues.apache.org/jira/browse/TEZ-2778
> Project: Apache Tez
> Issue Type: Sub-task
> Reporter: Bikas Saha
> Assignee: Bikas Saha
> Attachments: DAG.png, TEZ-2778.1.patch, TEZ-2778.2.patch,
> TEZ-2778.3.patch, cp-complex.JPG, dag_1439860407967_0579_1.svg
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
