spark.radl

network publica (outbound = 'yes')

system spark (
cpu.count>=1 and
memory.size>=1g and
net_interface.0.connection = 'publica' and
disk.0.os.name='linux' and
disk.0.os.flavour='ubuntu'
)

configure spark (
@begin
 - vars:
    VERSION: 1.6.3
   tasks:
    - unarchive: src=http://apache.rediris.es/spark/spark-{{VERSION}}/spark-{{VERSION}}-bin-hadoop2.6.tgz dest=/opt/ copy=no creates=/opt/spark-{{VERSION}}-bin-hadoop2.6 copy=no
    - file: src=/opt/spark-{{VERSION}}-bin-hadoop2.6 dest=/opt/spark state=link
    - copy: dest=/etc/profile.d/spark.sh content="export PATH=$PATH:/opt/spark/bin\nexport SPARK_HOME=/opt/spark"
@end
)

deploy spark 1