How to troubleshoot creation of default service instance in Ranger?

[#1]

I have a cluster that did not create a default service instance for HDFS in Ranger. On others, a service instance was created when I enabled the HDFS plugin and restarted services in Ambari.

I can't find anything in the Ambari Server logs or Ranger logs hinting that communication between them failed.

Where should this be logged? Do I need to turn on debug logging for that service?

(I'm using Ambari 2.4 / HDP 2.5 / Ranger 0.6)

[#2 · Accepted solution]

Usually there will be some output in Ambari UI console. Check in the output of HDFS service start.

[#3]

Thank you, I was able to find it on a system with successful creation. This helped me very much. In the "Restart NameNode" in restarting HDFS services after changing the plugin setting to Enabled.

2016-10-07 08:38:32,458 - Ambari admin user creation successful.
2016-10-07 08:38:33,148 - Repository created Successfully
2016-10-07 08:38:33,148 - Hdfs Repository created in Ranger admin
2016-10-07 08:38:33,149 - File['/usr/hdp/current/hadoop-client/conf/ranger-security.xml'] {'content': InlineTemplate(...), 'owner': 'hdfs', 'group': 'hadoop', 'mode': 0644}

[#4]

Turn on logging.

See: http://hortonworks.com/blog/best-practices-in-hdfs-authorization-with-apache-ranger/

https://hortonworks.com/hadoop-tutorial/manage-security-policy-hive-hbase-knox-ranger/

[#5]

Thanks for pointing to more information. I'm going to accept vperiasamy's answer.

[#6]

Please check the following property on HDFS Config to see if the Ranger Hook is enabled.

dfs.namenode.inode.attributes.provider.class=org.apache.ranger.authorization.hadoop.RangerHdfsAuthorizer