[
https://issues.apache.org/jira/browse/TEZ-1119?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14000255#comment-14000255
]
Hitesh Shah edited comment on TEZ-1119 at 5/16/14 8:27 PM:
-----------------------------------------------------------
[~rohini] The concern which Bikas has is that, in most cases, with HDFS inputs
and splits being calculated in the AM, each task has a different payload.
Furthermore, vertex managers have the capability to change the payload for any
task. Given that each task may have its own payload that is different from the
initial configured payload, this information is lost in ATS unless all tasks
generate their own information to ATS.
was (Author: hitesh):
[~rohini] The concern which Bikas has is that in most cases, with HDFS inputs
and splits being calculated in the AM, each task has a different payload.
Furthermore, vertex managers have the capability to change the payload for any
task. Given that each task may have its own payload that is different from the
initial configured payload, this information is lost in ATS unless all tasks
generate their own information to ATS.
> Support display of user payloads in Tez UI
> ------------------------------------------
>
> Key: TEZ-1119
> URL: https://issues.apache.org/jira/browse/TEZ-1119
> Project: Apache Tez
> Issue Type: Sub-task
> Reporter: Hitesh Shah
> Assignee: Hitesh Shah
>
> Applications like Pig rely heavily on configurations bundled in as part of
> the user payload for debugging. This requires each Input/Output/Processor's
> user payload to be converted into something that easily consumable by the UI
> layer to be displayed back to the user as meaningful data.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
