whirr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Eli Reisman (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (WHIRR-698) Implement the installation and configuration of Spark and its dependencies as a Whirr service
Date Fri, 08 Feb 2013 01:31:14 GMT

    [ https://issues.apache.org/jira/browse/WHIRR-698?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13574147#comment-13574147
] 

Eli Reisman commented on WHIRR-698:
-----------------------------------

Is there a timeout setting I can alter to make it more forgiving to installs that take a while?
Its not an undue amount of time, but it seems to be enough to upset Whirr/jclouds. I have
heard this sort of thing can be a problem for Whirr with many cluster nodes being launched
all at once too. Any fixes/ideas about this? Otherwise the service seems to be doing what
I want it to.

Any unrelated ideas about "Whirr best practices" that I'm not using? For instance, why am
I having trouble accessing the "global" scripts in my service scripts, such as install_tarball?
Also, I am still finding out from the Spark people whether there's any point to delivering
a spark conf file directly to the user's local .whirr dir or not, their YARN support is of
a different nature than their Mesos support, and there may not be any point to it.

Thanks!
                
> Implement the installation and configuration of Spark and its dependencies as a Whirr
service
> ---------------------------------------------------------------------------------------------
>
>                 Key: WHIRR-698
>                 URL: https://issues.apache.org/jira/browse/WHIRR-698
>             Project: Whirr
>          Issue Type: New Feature
>    Affects Versions: 0.8.1
>            Reporter: Eli Reisman
>            Priority: Minor
>         Attachments: WHIRR-698-1.patch, WHIRR-698-2.patch, WHIRR-698-3.patch
>
>
> I would like to Launch the Spark Project on a Hadoop YARN cluster on EC2 (or similar
cloud service.) and this service + recipe file make that happen. This is my first service,
and Spark is still learning to play nice with YARN and Whirr, so suggestions are welcome.
> Thanks!

--
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

Mime
View raw message