[ 
https://issues.apache.org/jira/browse/SPARK-1239?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14989638#comment-14989638
 ] 

Alex Slusarenko commented on SPARK-1239:
----------------------------------------

> How is this not a major issue for everyone?
Daniel, I had exactly the same thoughts, but I found out that most of Spark 
users have small data. Most users need about 10 servers. I was even more 
surprised when I had to fix SPARK-6246 myself. Spark was not able to launch 
more than 100 instances on Amazon.

> Don't fetch all map output statuses at each reducer during shuffles
> -------------------------------------------------------------------
>
>                 Key: SPARK-1239
>                 URL: https://issues.apache.org/jira/browse/SPARK-1239
>             Project: Spark
>          Issue Type: Improvement
>          Components: Shuffle, Spark Core
>    Affects Versions: 1.0.2, 1.1.0
>            Reporter: Patrick Wendell
>
> Instead we should modify the way we fetch map output statuses to take both a 
> mapper and a reducer - or we should just piggyback the statuses on each task. 



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org

Reply via email to