[
https://issues.apache.org/jira/browse/HAMA-990?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15292566#comment-15292566
]
Behroz Sikander commented on HAMA-990:
--------------------------------------
Ok. So, I am listing basic steps (requirements) that the script would do and I
will refine them over time as I have more information.
0- MRQL needs HAMA/FLINK/SPARK already installed, so we need to assume that
they are configured.
1- Download and extract the Apache MRQL latest stable release.
2- If the default configurations in mrql-env.sh [1] are alright then do nothing
otherwise update the mrql-env.sh file for the correct version.
3- Read the command input (all | kmeans ...) and prepare the input data for the
algorithm(s) and place them in HDFS (e.g. mrql.bsp -dist -nodes 50 RMAT.mrql
100000 1000000)
4- Execute the algorithm (mrql.bsp -dist -nodes 50 pagerank.mrql)
5- Dump the output
6- Repeat the algorithm for other platforms
[1] https://github.com/apache/incubator-mrql/blob/master/conf/mrql-env.sh
[2] https://mrql.incubator.apache.org/getting_started.html
> GSoC'16: Apache Hama benchmark against Spark and Flink
> ------------------------------------------------------
>
> Key: HAMA-990
> URL: https://issues.apache.org/jira/browse/HAMA-990
> Project: Hama
> Issue Type: Documentation
> Reporter: Behroz Sikander
> Priority: Minor
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
