You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@tez.apache.org by "Bikas Saha (JIRA)" <ji...@apache.org> on 2013/08/07 19:40:47 UTC
[jira] [Commented] (TEZ-350) Allow multiple vertices to be run
within a task if #tasks in each stage is 1
[ https://issues.apache.org/jira/browse/TEZ-350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13732201#comment-13732201 ]
Bikas Saha commented on TEZ-350:
--------------------------------
This is a corollary extension of container reuse. What we may need are input output format that write to a piece of shared memory exposed by yarntezdagchild. or some user land ramfs that can be exposed via filesystem to regular output formats.
> Allow multiple vertices to be run within a task if #tasks in each stage is 1
> ----------------------------------------------------------------------------
>
> Key: TEZ-350
> URL: https://issues.apache.org/jira/browse/TEZ-350
> Project: Apache Tez
> Issue Type: Improvement
> Reporter: Siddharth Seth
>
> Depending on the nature of data being processed, a dag with multiple dependent vertices can end up running a single task for each of the last few vertexes. This information may only be available at execution time.
> In such cases, Tez should be able to run these tasks within the same process - without having to shuffle data between the tasks. Also, as a follow on step - we could avoid serialization / deserialization of the data with custom Inputs / Outputs.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira