Hi Gerald, If you don't export SPARK_HOME, could you try add jar using
export ZEPPELIN_CLASSPATH=/path/to/your.jar that supposed to add your jar in the classpath of spark interpreter process. Hope this helps. Best, moon On Fri, Jun 3, 2016 at 1:02 AM Gerald Loeffler < gerald.loeff...@googlemail.com> wrote: > JL, > > would it be possible to show the config settings for the 2nd option, i.e. > without a separate SPARK_HOME (and therefore using Zeppelin's > built-in Spark distribution)? That would be very helpful! > > thank you in advance > gerald > > > On Wednesday, 1 June 2016, Jongyoul Lee <jongy...@gmail.com> wrote: > >> Hi, >> >> Generally, if you have your own "SPARK_HOME" settings, it's OK to set >> SPARK_HOME in zeppelin-env.sh and set your arguments into interpreter tab. >> Otherwise, you should add Zeppelin classpath and set arguments in the >> interpreter tab with "--jars". >> >> Hope this help, >> JL >> >> On Wed, Jun 1, 2016 at 5:55 PM, kishore r <vkishore...@gmail.com> wrote: >> >>> Hi >>> >>> We have built custom data provider from which we read the data and >>> eventually create a data frame. Is it possible to add the custom data >>> providers in to the Zeppelin classpath. If so can you share your thoughts >>> on how we can make it. >>> >>> Thanks >>> K >>> >> >> >> >> -- >> 이종열, Jongyoul Lee, 李宗烈 >> http://madeng.net >> > > > -- > -- > Gerald Loeffler > mailto:gerald.loeff...@googlemail.com > http://www.gerald-loeffler.net > >
