Re: NewSparkInterpreter fails on yarn-cluster
I specifically mean visualisation via ZeppelinContext inside a Spark interpreter. (e.g. "z.show(...)") The visualisation of SparkSQL results inside a SparkSQLInterpreter work fine, also in yarn-cluster mode. Am Do., 7. Juni 2018 um 14:30 Uhr schrieb Thomas Bünger < thom.bu...@googlemail.com>: > Hey Jeff, > > I tried your changes and now it works nicely. Thank you very much! > > But I still can't use any of the forms and visualizations in yarn-cluster? > I was hoping that this got resolved with the new SparkInterpreter so that > I can switch from yarn-client to yarn-cluster mode in 0.8, but I'm still > getting errors like > "error: not found: value z" > > Was this not in scope of that change? Is this a bug? Or is it known > limitation and also not supported in 0.8? > > Best regards, > Thomas > > Am Mi., 6. Juni 2018 um 03:28 Uhr schrieb Jeff Zhang : > >> >> I can confirm that this is a bug, and created >> https://issues.apache.org/jira/browse/ZEPPELIN-3531 >> >> Will fix it soon >> >> Jeff Zhang 于2018年6月5日周二 下午9:01写道: >> >>> >>> hmm, it looks like a bug. I will check it tomorrow. >>> >>> >>> Thomas Bünger 于2018年6月5日周二 下午8:56写道: >>> >>>> $ ls /usr/lib/spark/python/lib >>>> py4j-0.10.6-src.zip PY4J_LICENSE.txt pyspark.zip >>>> >>>> So folder exists and contains both necessary zips. Please note, that in >>>> local or yarn-client mode the files are properly picked up from that very >>>> same location. >>>> >>>> How does yarn-cluster work under the hood? Could it be that environment >>>> variables (like SPARK_HOME) are lost, because they are only available in my >>>> local shell + zeppelin daemon process? Do I need to tell YARN somehow about >>>> SPARK_HOME? >>>> >>>> Am Di., 5. Juni 2018 um 14:48 Uhr schrieb Jeff Zhang >>> >: >>>> >>>>> >>>>> Could you check whether there's folder /usr/lib/spark/python/lib ? >>>>> >>>>> >>>>> Thomas Bünger 于2018年6月5日周二 下午8:45写道: >>>>> >>>>>> >>>>>> sys.env >>>>>> java.lang.NullPointerException at >>>>>> org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) >>>>>> at >>>>>> org.apache.zeppelin.spark.NewSparkInterpreter.open(NewSparkInterpreter.java:90) >>>>>> at >>>>>> org.apache.zeppelin.spark.SparkInterpreter.open(SparkInterpreter.java:62) >>>>>> at >>>>>> org.apache.zeppelin.interpreter.LazyOpenInterpreter.open(LazyOpenInterpreter.java:69) >>>>>> at >>>>>> org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer$InterpretJob.jobRun(RemoteInterpreterServer.java:617) >>>>>> at org.apache.zeppelin.scheduler.Job.run(Job.java:188) at >>>>>> org.apache.zeppelin.scheduler.FIFOScheduler$1.run(FIFOScheduler.java:140) >>>>>> at >>>>>> java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) >>>>>> at java.util.concurrent.FutureTask.run(FutureTask.java:266) at >>>>>> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180) >>>>>> at >>>>>> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) >>>>>> at >>>>>> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) >>>>>> at >>>>>> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) >>>>>> at java.lang.Thread.run(Thread.java:748) >>>>>> >>>>>> >>>>>> Am Di., 5. Juni 2018 um 14:41 Uhr schrieb Jeff Zhang < >>>>>> zjf...@gmail.com>: >>>>>> >>>>>>> Could you paste the full stracktrace ? >>>>>>> >>>>>>> >>>>>>> Thomas Bünger 于2018年6月5日周二 下午8:21写道: >>>>>>> >>>>>>>> I've tried the 0.8.0-rc4 on my EMR cluster using the preinstalled >>>>>>>> version of spark under /usr/lib/spark. >>>>>>>> >>>>>>>> This works fine in local or yarn-client mode, but in yarn-cluster >>>>>>>> mode i just get a >>>>>>>> >>>>>>>> java.lang.NullPointerException at >>>>>>>> org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) >>>>>>>> >>>>>>>> Seems to be caused by an unsuccessful search for the py4j libraries. >>>>>>>> I've made sure that SPARK_HOME is actually set in .bash_rc, in >>>>>>>> zeppelin-env.sh and via the new %spark.conf, but somehow in the remote >>>>>>>> interpreter, something odd is going on. >>>>>>>> >>>>>>>> Best regards, >>>>>>>> Thomas >>>>>>>> >>>>>>>
Re: NewSparkInterpreter fails on yarn-cluster
Hey Jeff, I tried your changes and now it works nicely. Thank you very much! But I still can't use any of the forms and visualizations in yarn-cluster? I was hoping that this got resolved with the new SparkInterpreter so that I can switch from yarn-client to yarn-cluster mode in 0.8, but I'm still getting errors like "error: not found: value z" Was this not in scope of that change? Is this a bug? Or is it known limitation and also not supported in 0.8? Best regards, Thomas Am Mi., 6. Juni 2018 um 03:28 Uhr schrieb Jeff Zhang : > > I can confirm that this is a bug, and created > https://issues.apache.org/jira/browse/ZEPPELIN-3531 > > Will fix it soon > > Jeff Zhang 于2018年6月5日周二 下午9:01写道: > >> >> hmm, it looks like a bug. I will check it tomorrow. >> >> >> Thomas Bünger 于2018年6月5日周二 下午8:56写道: >> >>> $ ls /usr/lib/spark/python/lib >>> py4j-0.10.6-src.zip PY4J_LICENSE.txt pyspark.zip >>> >>> So folder exists and contains both necessary zips. Please note, that in >>> local or yarn-client mode the files are properly picked up from that very >>> same location. >>> >>> How does yarn-cluster work under the hood? Could it be that environment >>> variables (like SPARK_HOME) are lost, because they are only available in my >>> local shell + zeppelin daemon process? Do I need to tell YARN somehow about >>> SPARK_HOME? >>> >>> Am Di., 5. Juni 2018 um 14:48 Uhr schrieb Jeff Zhang : >>> >>>> >>>> Could you check whether there's folder /usr/lib/spark/python/lib ? >>>> >>>> >>>> Thomas Bünger 于2018年6月5日周二 下午8:45写道: >>>> >>>>> >>>>> sys.env >>>>> java.lang.NullPointerException at >>>>> org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) >>>>> at >>>>> org.apache.zeppelin.spark.NewSparkInterpreter.open(NewSparkInterpreter.java:90) >>>>> at >>>>> org.apache.zeppelin.spark.SparkInterpreter.open(SparkInterpreter.java:62) >>>>> at >>>>> org.apache.zeppelin.interpreter.LazyOpenInterpreter.open(LazyOpenInterpreter.java:69) >>>>> at >>>>> org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer$InterpretJob.jobRun(RemoteInterpreterServer.java:617) >>>>> at org.apache.zeppelin.scheduler.Job.run(Job.java:188) at >>>>> org.apache.zeppelin.scheduler.FIFOScheduler$1.run(FIFOScheduler.java:140) >>>>> at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) >>>>> at java.util.concurrent.FutureTask.run(FutureTask.java:266) at >>>>> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180) >>>>> at >>>>> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) >>>>> at >>>>> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) >>>>> at >>>>> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) >>>>> at java.lang.Thread.run(Thread.java:748) >>>>> >>>>> >>>>> Am Di., 5. Juni 2018 um 14:41 Uhr schrieb Jeff Zhang >>>> >: >>>>> >>>>>> Could you paste the full stracktrace ? >>>>>> >>>>>> >>>>>> Thomas Bünger 于2018年6月5日周二 下午8:21写道: >>>>>> >>>>>>> I've tried the 0.8.0-rc4 on my EMR cluster using the preinstalled >>>>>>> version of spark under /usr/lib/spark. >>>>>>> >>>>>>> This works fine in local or yarn-client mode, but in yarn-cluster >>>>>>> mode i just get a >>>>>>> >>>>>>> java.lang.NullPointerException at >>>>>>> org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) >>>>>>> >>>>>>> Seems to be caused by an unsuccessful search for the py4j libraries. >>>>>>> I've made sure that SPARK_HOME is actually set in .bash_rc, in >>>>>>> zeppelin-env.sh and via the new %spark.conf, but somehow in the remote >>>>>>> interpreter, something odd is going on. >>>>>>> >>>>>>> Best regards, >>>>>>> Thomas >>>>>>> >>>>>>
Re: NewSparkInterpreter fails on yarn-cluster
$ ls /usr/lib/spark/python/lib py4j-0.10.6-src.zip PY4J_LICENSE.txt pyspark.zip So folder exists and contains both necessary zips. Please note, that in local or yarn-client mode the files are properly picked up from that very same location. How does yarn-cluster work under the hood? Could it be that environment variables (like SPARK_HOME) are lost, because they are only available in my local shell + zeppelin daemon process? Do I need to tell YARN somehow about SPARK_HOME? Am Di., 5. Juni 2018 um 14:48 Uhr schrieb Jeff Zhang : > > Could you check whether there's folder /usr/lib/spark/python/lib ? > > > Thomas Bünger 于2018年6月5日周二 下午8:45写道: > >> >> sys.env >> java.lang.NullPointerException at >> org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) >> at >> org.apache.zeppelin.spark.NewSparkInterpreter.open(NewSparkInterpreter.java:90) >> at >> org.apache.zeppelin.spark.SparkInterpreter.open(SparkInterpreter.java:62) >> at >> org.apache.zeppelin.interpreter.LazyOpenInterpreter.open(LazyOpenInterpreter.java:69) >> at >> org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer$InterpretJob.jobRun(RemoteInterpreterServer.java:617) >> at org.apache.zeppelin.scheduler.Job.run(Job.java:188) at >> org.apache.zeppelin.scheduler.FIFOScheduler$1.run(FIFOScheduler.java:140) >> at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) >> at java.util.concurrent.FutureTask.run(FutureTask.java:266) at >> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180) >> at >> java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) >> at >> java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) >> at >> java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) >> at java.lang.Thread.run(Thread.java:748) >> >> >> Am Di., 5. Juni 2018 um 14:41 Uhr schrieb Jeff Zhang : >> >>> Could you paste the full stracktrace ? >>> >>> >>> Thomas Bünger 于2018年6月5日周二 下午8:21写道: >>> >>>> I've tried the 0.8.0-rc4 on my EMR cluster using the preinstalled >>>> version of spark under /usr/lib/spark. >>>> >>>> This works fine in local or yarn-client mode, but in yarn-cluster mode >>>> i just get a >>>> >>>> java.lang.NullPointerException at >>>> org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) >>>> >>>> Seems to be caused by an unsuccessful search for the py4j libraries. >>>> I've made sure that SPARK_HOME is actually set in .bash_rc, in >>>> zeppelin-env.sh and via the new %spark.conf, but somehow in the remote >>>> interpreter, something odd is going on. >>>> >>>> Best regards, >>>> Thomas >>>> >>>
Re: NewSparkInterpreter fails on yarn-cluster
sys.env java.lang.NullPointerException at org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) at org.apache.zeppelin.spark.NewSparkInterpreter.open(NewSparkInterpreter.java:90) at org.apache.zeppelin.spark.SparkInterpreter.open(SparkInterpreter.java:62) at org.apache.zeppelin.interpreter.LazyOpenInterpreter.open(LazyOpenInterpreter.java:69) at org.apache.zeppelin.interpreter.remote.RemoteInterpreterServer$InterpretJob.jobRun(RemoteInterpreterServer.java:617) at org.apache.zeppelin.scheduler.Job.run(Job.java:188) at org.apache.zeppelin.scheduler.FIFOScheduler$1.run(FIFOScheduler.java:140) at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511) at java.util.concurrent.FutureTask.run(FutureTask.java:266) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.access$201(ScheduledThreadPoolExecutor.java:180) at java.util.concurrent.ScheduledThreadPoolExecutor$ScheduledFutureTask.run(ScheduledThreadPoolExecutor.java:293) at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149) at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624) at java.lang.Thread.run(Thread.java:748) Am Di., 5. Juni 2018 um 14:41 Uhr schrieb Jeff Zhang : > Could you paste the full stracktrace ? > > > Thomas Bünger 于2018年6月5日周二 下午8:21写道: > >> I've tried the 0.8.0-rc4 on my EMR cluster using the preinstalled version >> of spark under /usr/lib/spark. >> >> This works fine in local or yarn-client mode, but in yarn-cluster mode i >> just get a >> >> java.lang.NullPointerException at >> org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) >> >> Seems to be caused by an unsuccessful search for the py4j libraries. >> I've made sure that SPARK_HOME is actually set in .bash_rc, in >> zeppelin-env.sh and via the new %spark.conf, but somehow in the remote >> interpreter, something odd is going on. >> >> Best regards, >> Thomas >> >
NewSparkInterpreter fails on yarn-cluster
I've tried the 0.8.0-rc4 on my EMR cluster using the preinstalled version of spark under /usr/lib/spark. This works fine in local or yarn-client mode, but in yarn-cluster mode i just get a java.lang.NullPointerException at org.apache.zeppelin.spark.NewSparkInterpreter.setupConfForPySpark(NewSparkInterpreter.java:149) Seems to be caused by an unsuccessful search for the py4j libraries. I've made sure that SPARK_HOME is actually set in .bash_rc, in zeppelin-env.sh and via the new %spark.conf, but somehow in the remote interpreter, something odd is going on. Best regards, Thomas
Re: zeppelin integration with helium not showing added charts
Hey, Usually, the package's json file (ultimate-line-chart.json) has to use the artefact's full, absolute path. Make sure that "artifact" is pointing to the right location where npm installed the package. 2017-10-19 9:24 GMT+02:00 Indtiny S : > Hi, > > I tried to add the helium packages to extend the support of zeppelin vis > to added it , I went to the path ZeppelinHome/local-repo/vis/ > node_modules/zeppelin-vis > and installed npm install ultimate-line-chart . > > and copied the ultimate-line-chart.jsom under ZeppelinHome/helium folder > > then went to home page and enabled that chart(like in the attached > helium.png) > > then wrote some query and tried to see that graph option but this line > chart was not shown , it was showing only attached charts (page.png) > > I did restart and refresh of zeppelin but still added line chart graph was > not showing . > > What could be problem ? > > I tried on zeppelin version 0.7.3 > > > >