Member since
06-28-2017
2
Posts
0
Kudos Received
0
Solutions
07-13-2017
03:06 PM
Hi Abraham, Spark interpreter is not impersonated. Uncheck <User Impersonate>, restart the interpreter and have another try.
... View more
12-12-2017
12:08 PM
@Abraham Johnson @vpoornalingam There is still another reason and cure for this scenario (HDP-2.6.2.0-205). It can also happen that Ambari is looking for the pid files in the wrong place In my case the pid files were actually located at: /var/run/hadoop/hdfs-<clustername>/hadoop-hdfs-<clustername>-namenode.pid while ambari-agent would look at : /var/run/hadoop/hdfs/hadoop-hdfs-hdfs-namenode.pid In this state, with both the dir and the pid file name wrong, Ambari does not detect a running HDFS service, and you would also not be able to (re)start it. The pid file location is deduted from this snippet in hadoop-env.sh: export HADOOP_PID_DIR={{hadoop_pid_dir_prefix}}/$USER I have yet to find out why Ambari decided to change the value of $USER all of a sudden.
... View more