I'm building a spark app with maven (with shade plugin) and scp'ing it to a data node for execution with spark-submit --deploy-mode cluster (since launching right from the build system with --deploy-mode client doesn't work because of asymmetric network not under my control).
Here's my launch command
spark-submit
--class Test
--master yarn
--deploy-mode cluster
--supervise
--verbose
jarName.jar
hdfs:///somePath/Test.txt
hdfs:///somePath/Test.out
The job quickly fails with a ClassNotFoundException for Test$1; one of the anonymous classes java creates from my main class
6/03/18 12:59:41 WARN scheduler.TaskSetManager: Lost task 0.0 in stage
0.0 (TID 0, dataNode3): java.lang.ClassNotFoundException: Test$1
I've seen this error mentioned many times (google) and most recommendations boil down to calling conf.setJars(jarPaths) or similar.
I really don't see why this is needed when the missing class is definitely (I've checked) available in jarName.jar , why specifying this at compile time is preferable to doing it at run time with --jar as a spark-submit argument, and in either case, what path I should provide for the jar. I've been copying it to my home directory on the datanode from target/jarName.jar on the build system but it seems spark-submit copies it to hdfs somewhere that's hard to nail down into a hard-coded path name at either compile time or launch time.
And most of all, why isn't spark-submit handling this automatically based on the someJar.jar argument, and if not, what should I do to fix it?
Check the answer from here
spark submit java.lang.ClassNotFoundException
spark-submit --class Test --master yarn --deploy-mode cluster --supervise --verbose jarName.jar hdfs:///somePath/Test.txt hdfs:///somePath/Test.out
Try to use, also you could check the absolute path in your project
--class com.myclass.Test
I had the same issue with my Scala Spark application when I tried to run it in "cluster" mode:
--master yarn --deploy-mode cluster
I found the solution on this page. Basically what I was missing (that is missing also in your command) is the "--jars" parameter that allows you to distribute the application jars to your cluster.
Suggestion: to be able to troubleshooting this kind of error you could use the following command:
yarn logs --applicationId yourApplicationId
where yourApplicationId shoould be in your yarn exception log.
Related
I have just configured spark on my Hadoop cluster and i want to run the spark sample job.
before that I want to understand what, this below job code stands for.
spark-submit --deploy-mode client --class org.apache.spark.examples.SparkPi $SPARK_HOME/examples/jars/spark-examples_2.11-2.4.0.jar 10
You can see all possible parameters for submitting a spark job on here. I summarized the ones in your submit script as below:
spark-submit
--deploy-mode client # client/cluster. default value client. Whether to deploy your driver on the worker nodes or locally
--class org.apache.spark.examples.SparkPi # The entry point for your application
$SPARK_HOME/examples/jars/spark-examples_2.11-2.4.0.jar 10 #jar file path and expected arguments
--master is another parameter usually defined in submit scripts. For my HDP cluster default value of master is yarn. You can see all possible values for master in spark documentation again.
I am new to Spark.
I want to run a Spark Structured Streaming application on cluster.
Master and workers has same configuration.
I have few queries for submitting app on cluster using spark-submit:
You may find them comical or strange.
How can I give path for 3rd party jars like lib/*? ( Application has 30+ jars)
Will Spark automatically distribute application and required jars to workers?
Does it require to host application on all the workers?
How can i know status of my application as I am working on console.
I am using following script for Spark-submit.
spark-submit
--class <class-name>
--master spark://master:7077
--deploy-mode cluster
--supervise
--conf spark.driver.extraClassPath <jar1, jar2..jarn>
--executor-memory 4G
--total-executor-cores 8
<running-jar-file>
But code is not running as per expectation.
Am i missing something?
To pass multiple jar file to Spark-submit you can set the following attributes in file SPARK_HOME_PATH/conf/spark-defaults.conf (create if not exists):
Don't forget to use * at the end of the paths
spark.driver.extraClassPath /fullpath/to/jar/folder/*
spark.executor.extraClassPath /fullpathto/jar/folder/*
Spark will set the attributes in the file spark-defaults.conf when you use the spark-submit command.
Copy your jar file on that directory and when you submit your Spark App on the cluster, the jar files in the specified paths will be loaded, too.
spark.driver.extraClassPath: Extra classpath entries to prepend
to the classpath of the driver. Note: In client mode, this config
must not be set through the SparkConf directly in your application,
because the driver JVM has already started at that point. Instead,
please set this through the --driver-class-path command line option or
in your default properties file.
--jars will transfer your jar files to worker nodes, and become available in both driver and executors' classpaths.
Please refer below link to see more details.
http://spark.apache.org/docs/latest/submitting-applications.html#advanced-dependency-management
You can make a fat jar containing all dependencies. Below link helps you understand that.
https://community.hortonworks.com/articles/43886/creating-fat-jars-for-spark-kafka-streaming-using.html
When running a Spark Shell query using something like this:
spark-shell yarn --name myQuery -i ./my-query.scala
Inside my query is simple Spark SQL query where I read parquet files and run simple queries and write out parquet files. When running these queries I get a nice progress bar like this:
[Stage7:===========> (14174 + 5) / 62500]
When I create a jar using the exact same query and run it with the following command-line:
spark-submit \
--master yarn-cluster \
--driver-memory 16G \
--queue default \
--num-executors 5 \
--executor-cores 4 \
--executor-memory 32G \
--name MyQuery \
--class com.data.MyQuery \
target/uber-my-query-0.1-SNAPSHOT.jar
I don't get any such progress bar. The command simply says repeatedly
17/10/20 17:52:25 INFO yarn.Client: Application report for application_1507058523816_0443 (state: RUNNING)
The query works fine and the results are fine. But I just need to have feedback when the process will finish. I have tried the following.
The web page of RUNNING Hadoop Applications does have a progress bar but it basically never moves. Even in the case of the spark-shell query that progress bar is useless.
I have tried get the progress bar through the YARN logs but they are not aggregated until the job is complete. Even then there is no progress bar in the logs.
Is there is a way to launch a spark query in jar on a cluster and have a progressbar?
When I create a jar using the exact same query and run it with the following command-line (...) I don't get any such progress bar.
The difference between these two seemingly similar Spark executions is the master URL.
In the former Spark execution with spark-shell yarn, the master is YARN in client deploy mode, i.e. the driver runs on the machine where you start spark-shell from.
In the latter Spark execution with spark-submit --master yarn-cluster, the master is YARN in cluster deploy mode (which is actually equivalent to --master yarn --deploy-mode cluster), i.e. the driver runs on a YARN node.
With that said, you won't get the nice progress bar (which is actually called ConsoleProgressBar) on the local machine but on the machine where the driver runs.
A simple solution is to replace yarn-cluster with yarn.
ConsoleProgressBar shows the progress of active stages to standard error, i.e. stderr.
The progress includes the stage id, the number of completed, active, and total tasks.
ConsoleProgressBar is created when spark.ui.showConsoleProgress Spark property is turned on and the logging level of org.apache.spark.SparkContext logger is WARN or higher (i.e. less messages are printed out and so there is a "space" for ConsoleProgressBar).
You can find more information in Mastering Apache Spark 2's ConsoleProgressBar.
I'm trying to submit an application to my spark cluster (standalone mode) through the spark-submit command. I'm following the
official spark documentation, as well as relying on this other one. Now the problem is that I get strange behaviors. My setup is the following:
I have a directory where all the dependency jars for my application are located, that is /home/myuser/jars
The jar of my application is in the same directory (/home/myuser/jars), and is called dat-test.jar
The entry point class in dat-test.jar is at the package path my.package.path.Test
Spark master is at spark://master:7077
Now, I submit the application directly on the master node, thus using the client deploy mode, running the command
./spark-submit --class my.package.path.Test --master spark://master:7077 --executor-memory 5G --total-executor-cores 10 /home/myuser/jars/*
and I received an error as
java.lang.ClassNotFoundException: my.package.path.Test
If I activate the verbose mode, what I see is that the primaryResource selected as jar containing the entry point is the first jar by alphabetical order in /home/myuser/jars/ (that is not dat-test.jar), leading (I supppose) to the ClassNotFoundException. All the jars in the same directory are anyway loaded as arguments.
Of course if I run
./spark-submit --class my.package.path.Test --master spark://master:7077 --executor-memory 5G --total-executor-cores 10 /home/myuser/jars/dat-test.jar
it finds the Test class, but it doesn't find other classes contained in other jars. Finally, if I use the --jars flag and run
./spark-submit --class my.package.path.Test --master spark://master:7077 --executor-memory 5G --total-executor-cores 10 --jars /home/myuser/jars/* /home/myuser/jars/dat-test.jar
I obtain the same result as the first option. First jar in /home/myuser/jars/ is loaded as primaryResource, leading to ClassNotFoundException for my.package.path.Test. Same if I add --jars /home/myuser/jars/*.jar.
Important points are:
I do not want to have a single jar with all the dependencies for development reasons
The jars in /home/myuser/jars/ are many. I'd like to know if there's a way to include them all instead of using the comma separated syntax
If I try to run the same commands with --deploy-cluster on the master node, I don't get the error, but the computation fails for some other reasons (but this is another problem).
Which is then the correct way of running a spark-submit in client mode?
Thanks
There is no way to include all jars using the --jars option, you will have to create a small script to enumerate them. This part is a bit sub-optimal.
I am currently working on logger analyse by using apache spark. I am new for Apache Spark. I have tried to use apache spark standalone mode. I can run my code by submitting jar with deploy-mode on the client. But I can not run with multi node cluster. I have used worker nodes are different machine.
sh spark-submit --class Spark.LogAnalyzer.App --deploy-mode cluster --master spark://rishon.server21:7077 /home/rishon/loganalyzer.jar "/home/rishon/apache-tomcat-7.0.63/LogAnalysisBackup/"
when i Run this command, it shows following error
15/10/20 18:04:23 ERROR ClientEndpoint: Exception from cluster was: java.io.FileNotFoundException: /home/rishon/loganalyzer.jar (No such file or directory)
java.io.FileNotFoundException: /home/rishon/loganalyzer.jar (No such file or directory)
at java.io.FileInputStream.open(Native Method)
at java.io.FileInputStream.<init>(FileInputStream.java:146)
at org.spark-project.guava.io.Files$FileByteSource.openStream(Files.java:124)
at org.spark-project.guava.io.Files$FileByteSource.openStream(Files.java:114)
at org.spark-project.guava.io.ByteSource.copyTo(ByteSource.java:202)
at org.spark-project.guava.io.Files.copy(Files.java:436)
at org.apache.spark.util.Utils$.org$apache$spark$util$Utils$$copyRecursive(Utils.scala:514)
at org.apache.spark.util.Utils$.copyFile(Utils.scala:485)
at org.apache.spark.util.Utils$.doFetchFile(Utils.scala:562)
at org.apache.spark.util.Utils$.fetchFile(Utils.scala:369)
at org.apache.spark.deploy.worker.DriverRunner.org$apache$spark$deploy$worker$DriverRunner$$downloadUserJar(DriverRunner.scala:150)
at org.apache.spark.deploy.worker.DriverRunner$$anon$1.run(DriverRunner.scala:79)
As my understanding, The driver program sends the data and application code to worker node. I don't know my understanding is correct or not. So Please help me to run application on a cluster.
I have tried to run jar on cluster and Now there is no exception but why the task is not assigned to worker node?
I have tried without clustering. Its working fine. shown in following figure
Above image shows, Task assigned to worker nodes. But I have one more problem to analyse the log file. Actually, I have log files in master node which is in a folder (ex: '/home/visva/log'). But the worker node searching the file on their own file system.
I met same problem.
My solution was that I uploaded my .jar file on the HDFS.
Enter the command line like this:
spark-submit --class com.example.RunRecommender --master spark://Hadoop-NameNode:7077 --deploy-mode cluster --executor-memory 6g --executor-cores 3 hdfs://Hadoop-NameNode:9000/spark-practise-assembly-1.0.jar
application-jar: Path to a bundled jar including your application and all dependencies. The URL must be globally visible inside of your cluster, for instance, an hdfs:// path or a file:// path that is present on all nodes.
If you use the cluster model in spark-submit , you need use the 6066 port(the default port of rest in spark) :
spark-submit --class Spark.LogAnalyzer.App --deploy-mode cluster --master spark://rishon.server21:6066 /home/rishon/loganalyzer.jar "/home/rishon/apache-tomcat-7.0.63/LogAnalysisBackup/"
In my case, i upload the jar of app to every node in cluster because i do not know how does the spark-submit to transfer the app automatically and i don't know how to specify a node as driver node .
Note: The jar path of app is a path that in the any node of cluster.
There are two deploy modes in Spark to run the script.
1.client (default): In client mode, the driver is launched directly within the spark-submit process which acts as a client to the cluster.(Master node)
2.cluster : If your application is submitted from a machine far from the worker machines, it is common to use cluster mode to minimize network latency between the drivers and the executors.
Reference Spark Documentation For Submitting JAR