Hi Tim, Rack awareness is related to how blocks are replicated amongst a cluster, so you won't loose any data (and improve locality) by setting it; It's not related to how masters are configured, but you can ensure by configuring them accordingly that you won't get your cluster down. Note that you possibly install more than 3 zk or journal nodes (has pros and cons though) There's a nice 2-parts article on HCC: https://community.hortonworks.com/articles/43057/rack-awareness-1.html ... View more

Tez, Pig, Sqoop... are clients so they won't be "green" like HDFS or YARN ... View more

@bigdata.neophyte kerberizing the sandbox is not as tough as you may think, you can find some guidelines here, or even simplier with freeIPA : here or here ... View more

@Sunile Manjee HA is only related to Oozie server, which is stateless (rely on its backend database). Since, failover can only be seen from a job submission perspective so there's no impact on current running jobs. ... View more

@Abhishek Srivastava You can find hive-site.xml in /etc/hive/conf/, but if you're using Ambari to manage your cluster then do the modification in Ambari so it can deploy it to every host. ... View more