Closed harinitya closed 5 years ago
Any feedback please? I am not able to use the profiler correctly.
did you have javaagent
? https://github.com/uber-common/jvm-profiler#example-to-run-with-spark-application
Thanks! I had seen another post here which asked to use library.path, changing it to -javaagent worked.
Hi,
I am trying to run Spark in the client mode, along with the JVM profiler, and I am not able to see any output in the yarn logs. The link below says that the spark.executor.extraJavaOptions does not work in the client mode, even though Spark documentation says its the spark.driver.extraJavaOptions variable that doesn't work.
https://stackoverflow.com/questions/28166667/how-to-pass-d-parameter-or-environment-variable-to-spark-job
I referred to the link below to see how the values need to be specified
https://github.com/uber-common/jvm-profiler/issues/15
My spark-defaults.conf has the following line
spark.jars=hdfs:///user/lib/jvm-profiler-1.0.0.jar spark.executor.extraJavaOptions -Djava.library.path=/home/spark-user/jvm-prof/jvm-profiler/target/jvm-profiler-1.0.0.jar=reporter=com.uber.profiling.reporters.ConsoleOutputReporter,tag=spark-bench,metricInterval=5000,sampleInterval=100
Does this look ok? Please let me know. And where exactly can I see the profiler output.
Thanks