The use of kerberos is transparent to job development, unless you have long-running application needs.
When properly configured to use security (via the XML files), the JobClient/etc. will auto-discover a local kinit-done login to make use of it for authentication. Counter question therefore is, what is your specific error (or need) that leads you to look for a different approach? That said, if you are looking at performing a keytab login (rather than using the local kinit cache) from within your application, you could use JAAS or rely on SecurityUtil#login(…) API: https://hadoop.apache.org/docs/stable1/api/org/apache/hadoop/security/SecurityUtil.html#login(org.apache.hadoop.conf.Configuration, java.lang.String, java.lang.String) On Tue, Feb 24, 2015 at 12:22 AM, Rahul Shrivastava <rhshr...@gmail.com> wrote: > Hi All, > > I am new to Hadoop Authentication ( Kerberos, Simple Authentication). I > have been doing reading for security aspects of Hadoop at > http://blog.godatadriven.com/kerberos_kdc_install.html and > http://blog.cloudera.com/blog/2012/03/authorization-and-authentication-in-hadoop/. > Can someone provide me a Java code example of a map -reduce program that > uses Kerberos. A writeup that does a walk through of how to setup a > map-reduce program using kerberos would be helpful. Here I would not need > how to do setup of kerberos but rather how to setup a map-reduce program > that uses kerberos. > > > Sorry if the question is previously been answered. > > > Thanks > > Rahul -- Harsh J
