Hi All
We are running an ESP-enabled HDInsight Spark cluster in Azure. We have no clue why some of our domain users are not able to use Zeppelin notebooks (usint the pyspark interpreter in our case). This is the very simple code that results in errors:
%livy2.pyspark
x = 2
print x
We have followed and tried the advices explained here without success: https://community.cloudera.com/t5/Support-Questions/Error-when-running-a-Spark-job-as-a-different-zeppelin-user/td-p/191846
I have attached the log files from Yarn and Zeppelin below. For some reason whatsoever, no Yarn application is created.
The output from command hdfs dfs -ls /user/ (for the not-working user) on the cluster's head node returns:
hdfs@hn0-upcch:~$ hdfs dfs -ls /user/
drwxrwxrwx+ - sdptest-cloud 00000000-0000-0000-0000-000000000000 0 2020-08-12 08:57 /user/sdptest-cloud
Any help / pointers / references about what else we could try are highly appreciated.
Christoph
17232-yarn-yarn-resourcemanager-hn1-upcchlog.txt
17038-zeppelin-interpreter-livy2-livy-zeppelin-hn0-upcch.txt
17186-zeppelin-zeppelin-hn0-upcchlog.txt