[
https://issues.apache.org/jira/browse/IGNITE-4526?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15802126#comment-15802126
]
Denis Magda commented on IGNITE-4526:
-------------------------------------
[~manish__mishra], great thanks!
All Apache Ignite examples are located under "ignite/examples/src/main"
directory. You need to create a package for Spark examples over there. As the
description of this ticket says we need to have two Spark workers that will be
able to share the state over a shared RDD. There has to be a way to run these
workers directly from an IDE. It's possible to run all Ignite examples from an
IDE like IntellijIdea or Eclipse and the Spark example shouldn't be an
exception.
Ignite related configuration needs to be in a form of a Spring XML file located
in "ignite/examples/config".
Once the example is ready it has to be tested. Add a test to
"ignite/examples/src/test" directory referring to existing tests that are
already there.
When everything is ready you need to do a final testing on projects TeamCity
and send a pull-request as described here
https://cwiki.apache.org/confluence/display/IGNITE/How+to+Contribute#HowtoContribute-1.CreateGitHubpull-request
I'm going to add you to the contributors list. Please assign the ticket on
yourself when this happens.
If you have any other questions or suggestions please go ahead sharing them.
> Add Spark Shared RDD examples
> -----------------------------
>
> Key: IGNITE-4526
> URL: https://issues.apache.org/jira/browse/IGNITE-4526
> Project: Ignite
> Issue Type: Task
> Reporter: Denis Magda
> Fix For: 2.0
>
>
> Spark Shared RDD functionality doesn't have its own examples. We need to add
> an example that will do the following:
> - First Spark Worker: creation of a shared RDD and filling it in with data.
> - First Spark Worker: performing some native spark transformation with the
> RDD.
> - Second Spark Worker: connecting to the same shared RDD.
> - Second Spark Worker: execution of SQL query using Spark API and Ignite API.
> Show that Ignite's query executes faster.
> The reason why the example should consist of two workers is to showcase one
> of the main benefits of Ignite's RDDs - ability to share the state (RDD) amid
> different Spark workers and processes.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
