[
https://issues.apache.org/jira/browse/TEZ-120?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14254360#comment-14254360
]
Siddharth Seth commented on TEZ-120:
------------------------------------
Yes, there's another related jira floating around somewhere on skipping
payloads sent over the wire.
This becomes more relevant if we start localizing large payloads.
> Optimize YarnTezDagChild to be more efficient in case of re-use
> ---------------------------------------------------------------
>
> Key: TEZ-120
> URL: https://issues.apache.org/jira/browse/TEZ-120
> Project: Apache Tez
> Issue Type: Sub-task
> Reporter: Siddharth Seth
>
> We end up re-reading tokens multiple times, creating and localizing the same
> config file. Typically, such information is common across multiple tasks in a
> job and does not need to be re-initialized for each task.
> Based on the task being allocated to a container, the AM should be able to
> indicate whether the Child needs to do a full re-initialization.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
