@prsingh When RM UI is Kerberized, only the admin / job user can view the startedTime, finishedTime, elapsedTime and amContainerLogs on using the RM web service. To enable access for other users, we can either set yarn.admin.acl = false or yarn.scheduler.capacity.root.acl_administer_jobs=* yarn.scheduler.capacity.root.acl_administer_queue=* ... View more

@Pranav Manwatkar Can you please verify what is "hadoop.http.staticuser.user" property set in yarn-site.xml. I'm wondering it might be set as "dr.who", can you please remove it and set it as yarn user, and try accessing the yarn logs from UI? ... View more

@prsingh That seems to be a bug, let's disable the ACL for ZK, by adding below in zookeeper-env.sh -Dzookeeper.skipACL=yes 1) Stop HBase 2) Add -Dzookeeper.skipACL=yes in zookeeper-env.sh export SERVER_JVMFLAGS="$SERVER_JVMFLAGS -Dzookeeper.skipACL=yes -Djava.security.auth.login.config={{zk_server_jaas_file}}" 3) rmr /hbase-secure/table/hbase:acl 4)revert back the change in step 2 -Dzookeeper.skipACL=yes 5) Restart HBase again 6) Try grant it should work now ... View more

Hi @prsingh, Spark currently doesn't support "insert into feature", you may create dataframe and append to the table. scala> var data = sqlContext.createDataFrame(Seq(("ZZ", "m:x", 34.0))).toDF("pv", "metric", "value") scala> data.show() scala> data.write.mode("append").saveAsTable("results_test_hive") scala> println(sqlContext.sql("select * from results_test_hive").count()) ... View more