Each Hadoop head node is Hadoop component specific. The roles are intended to be run so that they can be layered in a highly-available manner. E.g. multiple BCPC-Hadoop-Head-* machines will correctly build a MySQL, Zookeeper, HDFS JournalNode, etc. cluster and deploy the named component as well. Further, for components which support HA, the intention is one can simply add the role to multiple mach


{{#tags}}- {{label}}
{{/tags}}