Also, if you are using maven 3.3 you can add -pl '!flink' or any other interpreter you don't need. It will reduce both the nuild time and size.
בתאריך יום ד׳, 2 בספט׳ 2015, 18:48 מאת Alexander Bezzubov <b...@apache.org>: > Hi, > > thank you for you interest in Zeppelin project! > > Yes, by default a build that you did includes a lot of different > interpreters like Spark, Flink, Lens, etc so that is why the size is quite > substantial. > > In case you are about to use existing Spark\Hadoop - as of > https://issues.apache.org/jira/browse/ZEPPELIN-160 now, there is an > option now to build Zeppelin with those dependencies in a provided scope > (so they are not included in the final archive). > Then you just need to set SPARK_HOME and HADOOP_HOME to be able to use > existing Spark\Hadoop. > > Please, let me know if that helps! > > On Thu, Sep 3, 2015 at 12:38 AM, MrAsanjar . <afsan...@gmail.com> wrote: > >> I build zeppelin with following options as it was documented: >> *mvn clean package -Pspark-1.4 -Dspark.version=1.4.1 >> -Dhadoop.version=2.4.0 -Phadoop-2.4 -Pyarn -DskipTests -P build-distr* >> >> However the generated tarfile in >> zeppelin-distribution/target/zeppelin-0.6.0-incubating-SNAPSHOT.tar.gz is >> *414 >> Meg*, is that correct? >> I also noticed it does include spark, hadoop, and other tarfiles, do I >> need them if I am using existing hadoop & spark client configured and >> functioning? >> > >
