Falcon replication and mirroring between two Kerberized clusters?

[#1]

Does Falcon support replication and mirroring for HDFS and Hive between two Kerberized clusters (Prod and DR)?

[#2]

Can you tell if both these env using same KDC server? DistCP (which is what is used by Falcon for data replication) supports copying data between two kerberos environments so its a matter of making sure that the clusters can communicate correctly.

[#3]

Both clusters will be using either local KDC or AD Kerberos

[#4 · Accepted solution]

Set

<property>
    <name>oozie.launcher.mapreduce.job.hdfs-servers</name>
    <value>${nameNode1},${nameNode2}</value>
</property>

As this documentation suggests: http://oozie.apache.org/docs/4.1.0/DG_DistCpActionExtension.html

[#5]

Falcon will automatically add this property to oozie jobs in secure kerberized clusters. User does not have to add the property separately.

When defining a cluster entity, please make sure to add the following cluster property.

    <property name="dfs.namenode.kerberos.principal" value="nn/[email protected]"/> 

http://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.3.0/bk_installing_manually_book/content/configur...

[#6]

Replication works between clusters using Kerberos since Dal (and some additional scenarios using HCat works from Dal-M10