[
https://issues.apache.org/jira/browse/SPARK-7806?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14561657#comment-14561657
]
Matthew Goodman commented on SPARK-7806:
----------------------------------------
There are a couple lingering issues server side, that I am triaging today:
- The config for the http server seems slightly broken, though this may be
unrelated.
- Ganglia has a similar issue, but it may be part of the other config issue
- The per-module init.sh and setup.sh seem to function subtly wrong on the
master node.
Also notably, the macro-expand of the spark config "templates" seem very
brittle and duplicated in a couple of places.
Should I reopen this issue or start a new one?
> spark-ec2 launch script fails for Python3
> -----------------------------------------
>
> Key: SPARK-7806
> URL: https://issues.apache.org/jira/browse/SPARK-7806
> Project: Spark
> Issue Type: Bug
> Components: EC2, PySpark
> Affects Versions: 1.3.1
> Environment: All platforms.
> Reporter: Matthew Goodman
> Assignee: Matthew Goodman
> Priority: Minor
> Fix For: 1.4.0
>
>
> Depending on the options used the spark-ec2 script will terminate
> ungracefully.
> Relevant buglets include:
> - urlopen() returning bytes vs. string
> - floor division change for partition calculation
> - filter() iteration behavior change in module calculation
> I have a fixed version that I wish to contribute.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
