HDInsight Zeppelin Notebook Not Working

Christoph Kiefer 141 Reputation points
2020-08-12T11:46:21.087+00:00

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

Azure HDInsight
Azure HDInsight
An Azure managed cluster service for open-source analytics.
199 questions
{count} votes

Accepted answer
  1. Christoph Kiefer 141 Reputation points
    2020-08-24T09:29:20.047+00:00

    Hi @KranthiPakala-MSFT

    Thx for the reply and sorry for the long delay. It acutally worked when I implemented the changes mentioned here: https://community.cloudera.com/t5/Support-Questions/livy2-zepplin-issue/td-p/220980

    However, I am not sure if turning of impersonation is a good idea ...

    1 person found this answer helpful.

0 additional answers

Sort by: Most helpful