@Manjunath P N You will need to Install the KDC server on the KDC host manually on your own as described in https://docs.hortonworks.com/HDPDocuments/HDP2/HDP-2.6.5/bk_security/content/_optional_install_a_new_mit_kdc.html . What Ambari will do ? >>> Once the KDC server is setup completed as per the above doc and it is running fine then you can simply enable the Kerberos from Ambari UI to your cluster. Ambari will be able to install the Kerberos Client packages on all the cluster host (using yum) and it will also setup the "/etc/krb5.conf" on all cluster machines where the kerberos clients need to be setup and will create the keytabs there. https://docs.hortonworks.com/HDPDocuments/Ambari-2.6.2.2/bk_ambari-security/content/launching_the_kerberos_wizard_automated_setup.html The only thing you will need to manually do it to Install and Setup the KDC server .. so that you can tell ambari where your KDC server is running and what are the Kadmin principals, So that ambari can use that to setup kerberos to your cluster. ... View more

@Michael Bronson I do not think it is possible. This is because you are talking about two different File Systems (HDFS and Local FileSystem). If you want to keep syncing your Local Data Directory to HDFS directory then you can make use of some tools like Apache Flume. ... View more

@Sami Ahmad You are adding "hadoop classpath" however you will also need to add "hbase classpath" something like following: # javap -cp `hadoop classpath`:`hbase classpath`:.:  TestHbaseTable.java . ... View more

@Giridharan C 1. On the problematic host are you able to start those service components manually? As described in the following doc? https://docs.hortonworks.com/HDPDocuments/HDP3/HDP-3.1.0/administration/content/starting_hdp_services.html 2. Have you restarted ambari-agent on that host and after restart do you see any Error in the "/var/log/ambari-agent/ambari-agent.log" 3. Do you see the new command-xxx.json / output-xxx.txt and errors-xxx.txt kind of files are getting created in the Problematic host in the following location as soon as you trigger the stale config refresh command form ambari UI ? Just check the newly created file as soon as you triger stal config refresh command from the UI. # ls -lart /var/lib/ambari-agent/data/command*.json # ls -lart /var/lib/ambari-agent/data/output-*.txt # ls -lart /var/lib/ambari-agent/data/error*.txt . 4. When the Stale Config update fails on the problematic node then in the Ambarti UI operational log do you see any error. Can you please share that complete message which you see in the ambari UI when the stale config refresh was going on? . ... View more

@Ilia K Do you want to control the number of attempts? If yes then you might be interested in the following property "spark.yarn.maxAppAttempts" Example: --conf spark.yarn.maxAppAttempts=1 https://spark.apache.org/docs/latest/running-on-yarn.html spark.yarn.maxAppAttempts yarn.resourcemanager.am.max-attempts in YARN The maximum number of attempts that will be made to submit the application. It should be no larger than the global number of max attempts in the YARN configuration. ... View more