That is great! Thank you!
On Thu, Mar 14, 2019 at 9:04 PM Jeff Zhang wrote:
> If you use yarn-client mode, the python process is till in the same host
> as Zeppelin server. You can run it under yarn-cluster mode, so that python
> process would run in yarn container.
>
> Xun Liu 于2019年3月15日周五 上午
Ok, this had more information:
> INFO [2019-03-15 02:00:46,364] ({pool-2-thread-3}
> Logging.scala[logInfo]:54) - Logging events to
> hdfs:///var/log/spark/applicationHistory/application_1551287663522_0145
> ERROR [2019-03-15 02:00:46,366] ({SparkListenerBus}
> Logging.scala[logError]:91) - unc
This log is zeppelin server log, the root log should be in the spark
interpreter log. The file name is something like this :
zeppelin-interpreter-spark*.log
Dave Boyd 于2019年3月15日周五 上午9:31写道：
> Jeff:
>
> Running a simple spark.version paragraph I sometimes get this:
>
> INFO [2019-03-15 01:12
Jeff:
Running a simple spark.version paragraph I sometimes get this:
INFO [2019-03-15 01:12:18,720] ({pool-2-thread-49}
RemoteInterpreter.java[call]:142) - Open RemoteInterpreter
org.apache.zeppelin.spark.SparkInterpreter
INFO [2019-03-15 01:12:18,721] ({pool-2-thread-49}
RemoteInterpreter
If you use yarn-client mode, the python process is till in the same host as
Zeppelin server. You can run it under yarn-cluster mode, so that python
process would run in yarn container.
Xun Liu 于2019年3月15日周五 上午8:13写道：
> The latest version of zeppelin, PySpark driver run in yarn.
> You can use the
Hi Dave,
Could you paste the full stacktrace ? You can find it in the spark
interpreter log file which is located in ZEPPELIN_HOME/logs
Xun Liu 于2019年3月15日周五 上午8:21写道：
> Hi
>
> You can first execute a simple statement in spark, through sparksql, to
> see if it can run normally in YARN.
> If spa
Hi
You can first execute a simple statement in spark, through sparksql, to see if
it can run normally in YARN.
If sparksql is running without problems, check the zeppelin and spark on yarn
issues.
Also, what do you use for zeppelin-0.7.4? zeppelin-0.8.2? Is it a branch that
you maintain yourse
The latest version of zeppelin, PySpark driver run in yarn.
You can use the zeppelin 0.8.1 release version.
> 在 2019年3月15日，上午6:42，Roshan Punnoose 写道：
>
> Hey guys,
>
> In the latest release does the pyspark driver run in yarn? In 0.7 I know it
> ran on the same box as the Zeppelin installation
Hey guys,
In the latest release does the pyspark driver run in yarn? In 0.7 I know it
ran on the same box as the Zeppelin installation.
All:
I have some code that worked fine in Zeppelin 0.7.4 but I am having issues
in 0.8.2 when going from spark master of local to yarn-client. Yarn client
worked in 0.7.4.
When my master is set to local[*] it runs just fine. However, as soon as I
switch to yarn-client I get the Cannot cal
10 matches
Mail list logo
