MR-Runner - The MapReduce Runner

Synopsis

mrrunner COMMAND [OPTIONS]

Commands

start -l <LEVEL> -f <jdf> : configure and deploy an MR cluster 
list : list the active MR clusters
master <clusterID> : get the master node of an MR cluster 
stop <clusterID> : stop the MR cluster with the given clusterID
LEVEL = FATAL|ERROR|WARN|DEBUG|INFO

Description

The MR-Runner deploys MapReduce (MR) clusters on demand over the DAS-4 system. The MR-Runner is implemented in Java and currently configures Hadoop-1.0.0 clusters. KOALA is responsible for scheduling jobs, which in this case are complete MR clusters, received from the MR-runners. Based on the desired size (number of nodes) of the MR cluster, KOALA schedules the job on the adequate physical cluster by applying one of its placement policies.

The Koala MR-Runner System Architecture

Setup

1. Create your own Job Description File:

+(&
( count = "40") 
( maxWallTime = "30" )
( resourcemanagercontact = "fs3.das4.tudelft.nl" )
)
count = the size of the MapReduce cluster (number of machines x 8 processors)
maxWallTime = the duration of the SGE reservation in minutes
resourcemanagercontact = preferred execution site

2. Configuration and log files:

Hadoop configuration files path: ~/.mrcluster/<clusterID>
Hadoop log files path: ~/var/scratch/$USER/logs/

3. Executing Hadoop commands:

All Hadoop commands are executed on the master node of the MR cluster:

e.g. ssh <masterNode> /home/koala/hadoop/bin/hadoop --config ~/.mrcluster/<clusterID> dfs -ls /data

Big Data Processing System

 The MR-Runner enables access to the following stack of frameworks for big data processing:

  1. Storage layer: Hadoop Distributed File System
  2. Execution engine: Hadoop MapReduce
  3. Query language: Pig
Koala-based Big Data Processing System

KOALA News

  • January 2013: MR-Runner upgraded! Now the MR-Runner deploys Hadoop-1.0.0 clusters, compatible with Pig-0.10.0. 
  • December 2012KOALA 2.1 released! Deploy MapReduce clusters on DAS-4 with the Koala MR-Runner
      • November 2012:  Best Paper Award at MTAGS12 workshop (co-located with SC12) with work on MapReduce!
      • November 2009KOALA 2.0 released! You can now run Parameter sweep applications (PSAs) with KOALA CSRunner
      • April 2008: New KOALA runner! The OMRunner enables DRMAA and OpenMPI job submissions. 

      • July 2007: Paper accepted at Grid07 conference with work on scheduling malleable jobs in KOALA.

      • May 2007: KOALA has now been ported successfully to DAS-3. All the KOALA runners are operational apart from the DRunner.

      • April 2007: The KOALA IRunner has been updated to include recommendations made by the Ibis group