hello. Is there zeppelin and spark and yarn in the same machine? If is not you should open port they(zeppelin,spark,yarn,hdfs) are using.
2015년 11월 25일 수요일, Guillermo Ortiz<konstt2...@gmail.com>님이 작성한 메시지: > The spark's version is 1.3.1 in this CDH. > > 2015-11-24 17:38 GMT+01:00 Guillermo Ortiz <konstt2...@gmail.com > <javascript:_e(%7B%7D,'cvml','konstt2...@gmail.com');>>: > >> I'm using zeppelin 0.5.5 (binary installation) >> >> SPARK_HOME is correct /opt/cloudera/parcels/CDH/lib/spark/ >> and I checked /opt/cloudera/parcels/CDH/lib/spark/bin/spark-submit >> >> I don't know about checking connection between zeppelin and spark. Could >> you be more specific? thanks. >> >> 2015-11-24 17:19 GMT+01:00 Hyung Sung Shim <hss...@nflabs.com >> <javascript:_e(%7B%7D,'cvml','hss...@nflabs.com');>>: >> >>> Hello. >>> >>> It seems like no problem with your configuration. >>> >>> Please check the below things. >>> - Check connection(port) between zeppelin and spark. >>> >>> - If you set the SPARK_HOME in conf/zeppelin-env.sh, zeppelin use the >>> SPARK_HOME/bin/spark-submit command. >>> So check spark-submit command exist. >>> >>> Let me know your check result and if you still have the problem, >>> tell me the zeppelin version you are using then I will test your problem. >>> >>> Thanks. >>> >>> 2015-11-24 23:57 GMT+09:00 Guillermo Ortiz <konstt2...@gmail.com >>> <javascript:_e(%7B%7D,'cvml','konstt2...@gmail.com');>>: >>> >>>> I config spark.home to >>>> /opt/cloudera/parcels/CDH-5.4.8-1.cdh5.4.8.p0.4/lib/spark >>>> I'm not sure about the configuration of spark.yarn.jar >>>> >>>> I only have this configuration in zeppelin-env.sh >>>> export MASTER=yarn-client >>>> export >>>> SPARK_HOME=/opt/cloudera/parcels/CDH-5.4.8-1.cdh5.4.8.p0.4/lib/spark >>>> export HADOOP_CONF_DIR=/etc/hadoop/conf >>>> >>>> >>>> 2015-11-24 15:17 GMT+01:00 Guillermo Ortiz <konstt2...@gmail.com >>>> <javascript:_e(%7B%7D,'cvml','konstt2...@gmail.com');>>: >>>> >>>>> I'm trying to install Zeppelin in CDH 5.4.8 and I don't get to >>>>> connect. It starts correctly and I don't see any error in the logs but it >>>>> has a "disconnect" state in the web. >>>>> >>>>> I have tried with export MASTER=yarn-client and local[*] >>>>> I have edited the zeppelin-env.sh adding the export >>>>> HADOOP_CONF_DIR=/etc/hadoop/conf >>>>> >>>>> Spark works fine with the spark-shell. >>>>> >>>>> >>>> >>> >> >
