Corona Single Node Setup

Single Node Corona Cluster

This wiki provides instructions for starting a pseudo-distributed Corona Cluster on a single machine.

Building

Start in the hadoop-20 parent path and compile the code with:

ant -Dversion=0.20 clean jar bin-package

Starting up the cluster

First, set up the required environment and useful aliases.

source singleNodeHadoop/singleNodeSwitch.sh corona

If this is the first time, format the namespace in /tmp with

hadoop namenode -format

Start HDFS:

start-dfs

Start Corona:

start-corona

There are corresponding commands to stop the clusters - stop-dfs, and stop-corona

At this point you will be able to look at the local Corona Cluster Manager UI

Example Job

export HADOOP_CLASSPATH=$HADOOP_HOME/build/contrib/corona/lib/libthrift-0.7.0.jar hadoop jar build/hadoop-0.20-examples.jar sleep -Dmapred.fairscheduler.pool=group_a.pool_sla -m 10 -r 2