Amazonhascreatedasingle services usingouraccessandsecretkeys keypairbyfollowingtheinstructionsalthoughakeypairisnotstrictlynecessarytorunanemrcluster capabilitytoremotelylogintothemasternodeandgainlow-levelaccesstothecluster theresultingconfigurationinthe createandqueryans bucketusesomethinglikethefollowingcommand bucketsneedtobegloballyuniqueacrossallawsaccounts

The�AWS�command-line�interface

Each�AWS�service�historically�had�its�own�set�of�command-line�tools.�Recently�though, Amazon�has�created�a�single,�unified�command-line�tool�that�allows�access�to�most services.�The�Amazon�CLI�can�be�found�at�.

This�is�also�a�good�moment�to�set�up�an�EC2�key�pair�by�following�the�instructions

Although�a�key�pair�is�not�strictly�necessary�to�run�an�EMR�cluster,�it�will�give�us�the capability�to�remotely�log�in�to�the�master�node�and�gain�low-level�access�to�the�cluster.

We�can�provision�an�EMR�cluster�with�five�m1.xlarge�nodes�using�the�following commands:

$�aws�emr�create-cluster�--name�"EMR�cluster"�\--ami-version�3.2.0�\
--instance-type�m1.xlarge��\
--instance-count�5�\
--log-uri�s3://learninghadoop2/emr-logs

We�can�submit�workflows�by�adding�steps�to�a�running�cluster�using�the�following command:

$�aws�emr�add-steps�--cluster-id�<cluster>�--steps�<steps>�