Hi Thomas,
I try to the latest branch-0.8, it works for me. Could you try again to
verify it ?
Thomas Bünger 于2018年6月7日周四 下午8:34写道：
> I specifically mean visualisation via ZeppelinContext inside a Spark
> interpreter. (e.g. "z.show(...)")
> The visualisation of SparkSQL results inside a
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
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
>>
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
$ 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
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
>
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)
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
>
>
