Configure Kerberos authentication

To configure Kerberos authentication, add the following lines to the relevant provider stanza in indexes.conf:

vix.hadoop.security.authentication = kerberos
vix.java.security.krb5.kdc = <kerberos server name>
vix.java.security.krb5.realm = <kerberos default realm>
vix.kerberos.principal = <kerberos principal name of the user you want Hunk to interact with Hadoop, for example: hunk@YOUR-REALM.COM>
vix.kerberos.keytab = <kerberos keytab path, i.e., /path/hunk.keytab>
vix.hadoop.security.authorization = <hadoop security authorization true/false>
vix.dfs.namenode.kerberos.principal = <hadoop namenode kerberos principal name, i.e., hdfs/_HOST@YOUR-REALM.COM>
vix.mapreduce.jobtracker.kerberos.principal = <the hadoop jobtracker kerberos principal name, i.e., mapred/_HOST@YOUR-REALM.COM>
vix.hadoop.security.auth_to_local = <the mapping from Kerberos principals to short names (optional)>
vix.mapred.job.reuse.jvm.num.tasks = 1 

Note: Setting vix.mapred.job.reuse.jvm.num.tasks = 1 lets you avoid ENOENT task failures (detailed here: https://issues.apache.org/jira/browse/MAPREDUCE-4490).

If you are using YARN, you must also add the following property to the provider stanza:

vix.yarn.resourcemanager.principal = yarn/_HOST@YOUR-REALM.COM
vix.yarn.nodemanager.principal = yarn/_HOST@YOUR-REALM.COM

# kerberos with Hive
vix.hive.metastore.sasl.enabled = <true|false> 
vix.hive.metastore.kerberos.principal = <service principal for the metastore thrift server>

If you are preprocessing Hive via a metastore, vix.hive.metastore.sasl.enabled must be set to "true".