Welcome to the Centrify Express forums.
Moderation Notice: When posting to the forums always make sure you include the type and version of your UNIX, Linux or Mac Platform as well as the version of adclient you're using (adinfo -v).
Several things to note:
- Each Hadoop distribution has its own implementation path. For information, please review the Centrify integration documentation for Cloudera, Hortonworks or MapR
- You're using Centrify Express; this freemium version does not support AD one-way trusts.
Now to your question.
Note the Kerberos ticket cache file name: /tmp/krb5cc_cdc201327698_saYNYF (the cdc means Centrify DirectControl); this means that your system has been automatically configured to work with your AD Kerberos realm. In order for you to work with multiple configurations, you need to follow the guidance from this post:
Hopefully you'll understand that there's an assumption that your realm will be AD, but that you can use both configuration (joined to AD and using a MIT Kerberos as well) by redirecting the location of the krb5.conf file to an alternative location; this way the system can be joined to both realms.
HOWEVER; why would you stand-up an independent MIT Kerberos if with Centrify software you can make your Hadoop deployment work with AD? Less complexity, easier path from test to production.
This is a non-trivial task and my advice is that you use Centrify Standard Edition + our Award-winning PS if this a commercial organization.
R.P