[jira] (SPARK-19371) Cannot spread cached partitions evenly across executors

[Thunder Stumpges (JIRA)]

Title: Message Title

Thunder Stumpges commented on SPARK-19371   Re: Cannot spread cached partitions evenly across executors ... one execution slot is as good as another, because it's a core operating on local memory. I do believe Spark already prefers spreading tasks across executors, all else equal. So, this does not seem to be what we're seeing. Here's output from our executors tab after running a while. Notice the executors with more RDD partitions have proportionally more tasks executed. Add Comment   This message was sent by Atlassian JIRA (v6.3.15#6346-sha1:dbc023d)