Hello,
I'd like to use zeppelin on my local computer and use it to run spark
executors on a distant yarn cluster since I can't easily install zeppelin
on the cluster gateway.
I installed the correct hadoop version (2.6), and compiled zeppelin (from
the master branch) as following:
*mvn clean package -DskipTests -Phadoop-2.6 -Dhadoop.version=2.6.0-cdh5.5.0
-Pyarn -Pspark-2.0 -Pscala-2.11*
I also set HADOOP_HOME_DIR to /usr/local/lib/hadoop where my hadoop is
installed (I also tried with /usr/local/lib/hadoop/etc/hadoop/ where the
conf files such as yarn-site.xml are). I set yarn.resourcemanager.hostname
to the resource manager of the cluster (I copied the value from the config
file on the cluster) but when I start a spark command it still tries to
connect to 0.0.0.0:8032 as one can see in the logs:
*INFO [2016-11-01 20:48:26,581] ({pool-2-thread-2}
Client.java[handleConnectionFailure]:862) - Retrying connect to server:
0.0.0.0/0.0.0.0:8032 <http://0.0.0.0/0.0.0.0:8032>. Already tried 9
time(s); retry policy is RetryUpToMaximumCountWithFixedSleep(maxRetries=10,
sleepTime=1000 MILLISECONDS)*
Am I missing something something? Is there any additional parameters to set?
Thanks!
Benoit
