1. Stop zeppelin 2. Add a dependency in interpreter.json "dependencies": [ { "groupArtifactVersion": "/........./spark-cassandra-connector-1.5.0_2.10.jar", "local": false, "exclusions": [] } ] 3. Start zeppelin 4. Run a paragraph with import com.datastax.spark.connector._ => error: object datastax is not a member of package com 5. Edit Spark interpreter settings, save it and restart it 6. Re run the paragraph => no error
2016-02-23 19:25 GMT+01:00 moon soo Lee <m...@apache.org>: > interpreter.json supposed to be loaded on launch. > Could double check that interpreter.json is not read at zeppelin launch? > Or if it keep happening, could you let me know how to reproduce? > > Thanks, > moon > > > On Tue, Feb 23, 2016 at 8:22 AM vincent gromakowski < > vincent.gromakow...@gmail.com> wrote: > >> What is the best way to configure spark interpreter ? >> >> Should I use zeppelin-env.sh and a very long line of "export >> SPARK_SUBMIT_OPTIONS" >> >> or configure interpreter.json before launching Zeppelin daemon >> >> It seems interpreter.json is not read at zeppelin launch,I need to >> manually go to settings web UI,edit the spark interpreter and restart it... >> >> 2016-02-23 15:15 GMT+01:00 vincent gromakowski < >> vincent.gromakow...@gmail.com>: >> >>> Hi, >>> I am trying to automatcally add jars to spark interpreter with several >>> methods but I cannot achieve it. >>> I am currently generating an interpreter.json file from ansible >>> templates before launching Zeppelin in Marathon. >>> 1. spark.jars >>> 2. spark.driver.extraClassPath >>> 3. groupArtifactVersion (dependency loading) >>> >>> In all case I get a class not found exception for the spark cassandra >>> connector. The only way to make it works is to go to interpreter settings, >>> edit spark settings, then save and restart the interpreter but it's not >>> automatic at all as we need to do it each time Zeppelin is started. >>> >>> Is the interpreter.json file automatically loaded at the start of >>> Zeppelin ? >>> >> >>