[
https://issues.apache.org/jira/browse/MRQL-65?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14305236#comment-14305236
]
Robert Metzger commented on MRQL-65:
------------------------------------
I'm currently very overloaded with work.
We had a similar situation a while back in the Flink community. There was some
effort in our project to give users a nice "the first 45 minutes with the
project" experience. So every minute counts ;)
What we did in the Flink project is just calling "java" by default, instead of
expecting the user to set the JAVA_HOME path.
If the user sets it, we'll also use it.
The same applies to the "jar" command in your case.
I would also recommend to "re-use" defined environment variables (in particular
hadoop configuration directories etc.)
I can highly recommend using Google Compute Engine's "bdutil" hadoop2
deployment. Its very well set up and has all the important environment
variables there.
> Improve the environment configuration file
> ------------------------------------------
>
> Key: MRQL-65
> URL: https://issues.apache.org/jira/browse/MRQL-65
> Project: MRQL
> Issue Type: Improvement
> Reporter: Leonidas Fegaras
> Priority: Critical
> Fix For: 0.9.6
>
>
> This was reported by Robert Metzger. We need to improve the conf/mrql-env.sh
> file so that it works out of the box for certain environments, such Google
> cloud and Amazon EC2. Maybe also create scripts in bin/ to deploy MRQL on the
> cloud.
> Any volunteers?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
