Hi Saif, HDP works well for me. What version of HDP and zeppelin do you use. And could you paste the full interpreter log ?
Hyung Sung Shim <hss...@nflabs.com>于2016年12月14日周三 上午8:36写道： > Hello. Saif. > > I believe pre-build Spark binary includes YARN. > Please refer to [0] for yarn setting, if doesn't work please share your > *conf/zeppelin-env.sh* and logs. > > [0] > > http://zeppelin.apache.org/docs/0.7.0-SNAPSHOT/install/spark_cluster_mode.html#4-configure-spark-interpreter-in-zeppelin > > 2016-12-14 7:32 GMT+09:00 <saif.a.ell...@wellsfargo.com>: > > Hello all, > > I am trying to setup a Spark interpreter who hits YARN on Zeppelin. We > have both HDP and Standalone Spark running. > > Stand alone Spark works fine but fails if interpreter is set to yarn > master (YARN log): > > Error: Could not find or load main class > org.apache.spark.deploy.yarn.ExecutorLauncher > > HDP Spark does not work at all when used within Zeppelin (zeppelin > interpreter log) : > > ERROR [2016-12-13 14:25:54,971] ({pool-2-thread-2} > Utils.java[instantiateClass]:80) - > java.lang.reflect.InvocationTargetException > at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native > Method) > > One of my questions is If pre-built Spark binary distribution includes > YARN? To confirm whether the first error is caused by the lack of YARN > libraries in SPARK > > Other than that, I am quite at loss in this one, any help appreciated > > Saif > > > > >
