I installed Spark using the AWS EC2 guide and I can launch the program fine using the bin/pyspark script to get to the spark prompt and can also do the Quick Start quide successfully.
However, I cannot for the life of me figure out how to stop all of the verbose INFO logging after each command.
I have tried nearly every possible scenario in the below code (commenting out, setting to OFF) within my log4j.properties file in the conf folder in where I launch the application from as well as on each node and nothing is doing anything. I still get the logging INFO statements printing after executing each statement.
I am very confused with how this is supposed to work.
#Set everything to be logged to the console log4j.rootCategory=INFO, console
log4j.appender.console=org.apache.log4j.ConsoleAppender
log4j.appender.console.target=System.err
log4j.appender.console.layout=org.apache.log4j.PatternLayout
log4j.appender.console.layout.ConversionPattern=%d{yy/MM/dd HH:mm:ss} %p %c{1}: %m%n
# Settings to quiet third party logs that are too verbose
log4j.logger.org.eclipse.jetty=WARN
log4j.logger.org.apache.spark.repl.SparkIMain$exprTyper=INFO
log4j.logger.org.apache.spark.repl.SparkILoop$SparkILoopInterpreter=INFO
Here is my full classpath when I use SPARK_PRINT_LAUNCH_COMMAND:
Spark Command: /Library/Java/JavaVirtualMachines/jdk1.8.0_05.jdk/Contents/Home/bin/java -cp :/root/spark-1.0.1-bin-hadoop2/conf:/root/spark-1.0.1-bin-hadoop2/conf:/root/spark-1.0.1-bin-hadoop2/lib/spark-assembly-1.0.1-hadoop2.2.0.jar:/root/spark-1.0.1-bin-hadoop2/lib/datanucleus-api-jdo-3.2.1.jar:/root/spark-1.0.1-bin-hadoop2/lib/datanucleus-core-3.2.2.jar:/root/spark-1.0.1-bin-hadoop2/lib/datanucleus-rdbms-3.2.1.jar -XX:MaxPermSize=128m -Djava.library.path= -Xms512m -Xmx512m org.apache.spark.deploy.SparkSubmit spark-shell --class org.apache.spark.repl.Main
contents of spark-env.sh:
#!/usr/bin/env bash
# This file is sourced when running various Spark programs.
# Copy it as spark-env.sh and edit that to configure Spark for your site.
# Options read when launching programs locally with
# ./bin/run-example or ./bin/spark-submit
# - HADOOP_CONF_DIR, to point Spark towards Hadoop configuration files
# - SPARK_LOCAL_IP, to set the IP address Spark binds to on this node
# - SPARK_PUBLIC_DNS, to set the public dns name of the driver program
# - SPARK_CLASSPATH=/root/spark-1.0.1-bin-hadoop2/conf/
# Options read by executors and drivers running inside the cluster
# - SPARK_LOCAL_IP, to set the IP address Spark binds to on this node
# - SPARK_PUBLIC_DNS, to set the public DNS name of the driver program
# - SPARK_CLASSPATH, default classpath entries to append
# - SPARK_LOCAL_DIRS, storage directories to use on this node for shuffle and RDD data
# - MESOS_NATIVE_LIBRARY, to point to your libmesos.so if you use Mesos
# Options read in YARN client mode
# - HADOOP_CONF_DIR, to point Spark towards Hadoop configuration files
# - SPARK_EXECUTOR_INSTANCES, Number of workers to start (Default: 2)
# - SPARK_EXECUTOR_CORES, Number of cores for the workers (Default: 1).
# - SPARK_EXECUTOR_MEMORY, Memory per Worker (e.g. 1000M, 2G) (Default: 1G)
# - SPARK_DRIVER_MEMORY, Memory for Master (e.g. 1000M, 2G) (Default: 512 Mb)
# - SPARK_YARN_APP_NAME, The name of your application (Default: Spark)
# - SPARK_YARN_QUEUE, The hadoop queue to use for allocation requests (Default: ‘default’)
# - SPARK_YARN_DIST_FILES, Comma separated list of files to be distributed with the job.
# - SPARK_YARN_DIST_ARCHIVES, Comma separated list of archives to be distributed with the job.
# Options for the daemons used in the standalone deploy mode:
# - SPARK_MASTER_IP, to bind the master to a different IP address or hostname
# - SPARK_MASTER_PORT / SPARK_MASTER_WEBUI_PORT, to use non-default ports for the master
# - SPARK_MASTER_OPTS, to set config properties only for the master (e.g. "-Dx=y")
# - SPARK_WORKER_CORES, to set the number of cores to use on this machine
# - SPARK_WORKER_MEMORY, to set how much total memory workers have to give executors (e.g. 1000m, 2g)
# - SPARK_WORKER_PORT / SPARK_WORKER_WEBUI_PORT, to use non-default ports for the worker
# - SPARK_WORKER_INSTANCES, to set the number of worker processes per node
# - SPARK_WORKER_DIR, to set the working directory of worker processes
# - SPARK_WORKER_OPTS, to set config properties only for the worker (e.g. "-Dx=y")
# - SPARK_HISTORY_OPTS, to set config properties only for the history server (e.g. "-Dx=y")
# - SPARK_DAEMON_JAVA_OPTS, to set config properties for all daemons (e.g. "-Dx=y")
# - SPARK_PUBLIC_DNS, to set the public dns name of the master or workers
export SPARK_SUBMIT_CLASSPATH="$FWDIR/conf"
This question is related to
python
scala
apache-spark
hadoop
pyspark
I you want to keep using the logging (Logging facility for Python) you can try splitting configurations for your application and for Spark:
LoggerManager()
logger = logging.getLogger(__name__)
loggerSpark = logging.getLogger('py4j')
loggerSpark.setLevel('WARNING')
This below code snippet for scala users :
Option 1 :
Below snippet you can add at the file level
import org.apache.log4j.{Level, Logger}
Logger.getLogger("org").setLevel(Level.WARN)
Option 2 :
Note : which will be applicable for all the application which is using spark session.
import org.apache.spark.sql.SparkSession
private[this] implicit val spark = SparkSession.builder().master("local[*]").getOrCreate()
spark.sparkContext.setLogLevel("WARN")
Option 3 :
Note : This configuration should be added to your log4j.properties.. (could be like /etc/spark/conf/log4j.properties (where the spark installation is there) or your project folder level log4j.properties) since you are changing at module level. This will be applicable for all the application.
log4j.rootCategory=ERROR, console
IMHO, Option 1 is wise way since it can be switched off at file level.
Spark 1.6.2:
log4j = sc._jvm.org.apache.log4j
log4j.LogManager.getRootLogger().setLevel(log4j.Level.ERROR)
Spark 2.x:
spark.sparkContext.setLogLevel('WARN')
(spark being the SparkSession)
Alternatively the old methods,
Rename conf/log4j.properties.template to conf/log4j.properties in Spark Dir.
In the log4j.properties, change log4j.rootCategory=INFO, console to log4j.rootCategory=WARN, console
Different log levels available:
Programmatic way
spark.sparkContext.setLogLevel("WARN")
Available Options
ERROR
WARN
INFO
Simply add below param to your spark-submit command
--conf "spark.driver.extraJavaOptions=-Dlog4jspark.root.logger=WARN,console"
This overrides system value temporarily only for that job. Check exact property name (log4jspark.root.logger here) from log4j.properties file.
Hope this helps, cheers!
You can use setLogLevel
val spark = SparkSession
.builder()
.config("spark.master", "local[1]")
.appName("TestLog")
.getOrCreate()
spark.sparkContext.setLogLevel("WARN")
This may be due to how Spark computes its classpath. My hunch is that Hadoop's log4j.properties file is appearing ahead of Spark's on the classpath, preventing your changes from taking effect.
If you run
SPARK_PRINT_LAUNCH_COMMAND=1 bin/spark-shell
then Spark will print the full classpath used to launch the shell; in my case, I see
Spark Command: /usr/lib/jvm/java/bin/java -cp :::/root/ephemeral-hdfs/conf:/root/spark/conf:/root/spark/lib/spark-assembly-1.0.0-hadoop1.0.4.jar:/root/spark/lib/datanucleus-api-jdo-3.2.1.jar:/root/spark/lib/datanucleus-core-3.2.2.jar:/root/spark/lib/datanucleus-rdbms-3.2.1.jar -XX:MaxPermSize=128m -Djava.library.path=:/root/ephemeral-hdfs/lib/native/ -Xms512m -Xmx512m org.apache.spark.deploy.SparkSubmit spark-shell --class org.apache.spark.repl.Main
where /root/ephemeral-hdfs/conf is at the head of the classpath.
I've opened an issue [SPARK-2913] to fix this in the next release (I should have a patch out soon).
In the meantime, here's a couple of workarounds:
export SPARK_SUBMIT_CLASSPATH="$FWDIR/conf" to spark-env.sh./root/ephemeral-hdfs/conf/log4j.properties.For PySpark, you can also set the log level in your scripts with sc.setLogLevel("FATAL"). From the docs:
Control our logLevel. This overrides any user-defined log settings. Valid log levels include: ALL, DEBUG, ERROR, FATAL, INFO, OFF, TRACE, WARN
The way I do it is:
in the location I run the spark-submit script do
$ cp /etc/spark/conf/log4j.properties .
$ nano log4j.properties
change INFO to what ever level of logging you want and then run your spark-submit
In Spark 2.0 you can also configure it dynamically for your application using setLogLevel:
from pyspark.sql import SparkSession
spark = SparkSession.builder.\
master('local').\
appName('foo').\
getOrCreate()
spark.sparkContext.setLogLevel('WARN')
In the pyspark console, a default spark session will already be available.
Inspired by the pyspark/tests.py I did
def quiet_logs(sc):
logger = sc._jvm.org.apache.log4j
logger.LogManager.getLogger("org"). setLevel( logger.Level.ERROR )
logger.LogManager.getLogger("akka").setLevel( logger.Level.ERROR )
Calling this just after creating SparkContext reduced stderr lines logged for my test from 2647 to 163. However creating the SparkContext itself logs 163, up to
15/08/25 10:14:16 INFO SparkDeploySchedulerBackend: SchedulerBackend is ready for scheduling beginning after reached minRegisteredResourcesRatio: 0.0
and it's not clear to me how to adjust those programmatically.
Edit your conf/log4j.properties file and Change the following line:
log4j.rootCategory=INFO, console
to
log4j.rootCategory=ERROR, console
Another approach would be to :
Fireup spark-shell and type in the following:
import org.apache.log4j.Logger
import org.apache.log4j.Level
Logger.getLogger("org").setLevel(Level.OFF)
Logger.getLogger("akka").setLevel(Level.OFF)
You won't see any logs after that.
>>> log4j = sc._jvm.org.apache.log4j
>>> log4j.LogManager.getRootLogger().setLevel(log4j.Level.ERROR)
I used this with Amazon EC2 with 1 master and 2 slaves and Spark 1.2.1.
# Step 1. Change config file on the master node
nano /root/ephemeral-hdfs/conf/log4j.properties
# Before
hadoop.root.logger=INFO,console
# After
hadoop.root.logger=WARN,console
# Step 2. Replicate this change to slaves
~/spark-ec2/copy-dir /root/ephemeral-hdfs/conf/
Source: Stackoverflow.com