[jira] Commented: (HADOOP-406) Tasks launched by tasktracker in separate JVM can't generate log output

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

[jira] Commented: (HADOOP-406) Tasks launched by tasktracker in separate JVM can't generate log output

JIRA jira@apache.org

    [ https://issues.apache.org/jira/browse/HADOOP-406?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12461953 ]

Dominik Friedrich commented on HADOOP-406:

Please ignore my previous comment. It worked after the upgrade only because there were no mapred errors that had to be logged. I should have read the all comments to figure out that the log appender causes this problem.

> Tasks launched by tasktracker in separate JVM can't generate log output
> -----------------------------------------------------------------------
>                 Key: HADOOP-406
>                 URL: https://issues.apache.org/jira/browse/HADOOP-406
>             Project: Hadoop
>          Issue Type: Bug
>          Components: mapred
>    Affects Versions: 0.4.0
>            Reporter: Chris Schneider
>         Assigned To: Owen O'Malley
>         Attachments: HADOOP-406.patch
> Child JVM's don't have access to logging config system properties. When the child JVM gets launched, it doesn't inherit the Java system properties hadoop.log.dir and hadoop.log.file (which are actually based on the Bash environment variables $HADOOP_LOG_DIR and $HADOOP_LOGFILE). This means that you get no log messages from the actual map/reduce tasks that are executing.
> Stefan Groschupf reported this problem a while back:
> -------------------------------------------------------------------------
> To: [hidden email]
> From: Stefan Groschupf <[hidden email]>
> Subject: tasks can't log bug?
> Date: Tue, 25 Jul 2006 19:26:17 -0700
> X-Virus-Checked: Checked by ClamAV on apache.org
> Hi Hadoop developers,
> I'm confused about the way logging works within map or reduce tasks.
> Since tasks are launched in a new JVM  the java system properties "hadoop.log.dir" and "hadoop.log.file" are not passed to the new JVM.
> This prevents the child process from logging properly. In fact you get:
>  java.io.FileNotFoundException: / (Is a directory)
>   at java.io.FileOutputStream.openAppend(Native Method)
>   at java.io.FileOutputStream.<init>(FileOutputStream.java:177)
>   at java.io.FileOutputStream.<init>(FileOutputStream.java:102)
>   at org.apache.log4j.FileAppender.setFile(FileAppender.java:289)
>   at org.apache.log4j.RollingFileAppender.setFile(RollingFileAppender.java:165)
>   at org.apache.log4j.FileAppender.activateOptions(FileAppender.java:163)
>   at org.apache.log4j.config.PropertySetter.activate(PropertySetter.java:256)
>   at org.apache.log4j.config.PropertySetter.setProperties(PropertySetter.java:132)
>   at org.apache.log4j.config.PropertySetter.setProperties(PropertySetter.java:96)
>   at org.apache.log4j.PropertyConfigurator.parseAppender(PropertyConfigurator.java:654)
>   at org.apache.log4j.PropertyConfigurator.parseCategory(PropertyConfigurator.java:612)
>   at org.apache.log4j.PropertyConfigurator.configureRootCategory(PropertyConfigurator.j
> 2006-07-25 15:59:07,553 INFO  mapred.TaskTracker (TaskTracker.java:main(993)) - Child
>   at org.apache.log4j.PropertyConfigurator.doConfigure(PropertyConfigurator.java:415)
>   at org.apache.log4j.PropertyConfigurator.doConfigure(PropertyConfigurator.java:441)
>   at org.apache.log4j.helpers.OptionConverter.selectAndConfigure(OptionConverter.java:4
>   at org.apache.log4j.LogManager.<clinit>(LogManager.java:122)
>   at org.apache.log4j.Logger.getLogger(Logger.java:104)
>   at org.apache.commons.logging.impl.Log4JLogger.getLogger(Log4JLogger.java:229)
>   at org.apache.commons.logging.impl.Log4JLogger.<init>(Log4JLogger.java:65)
>   at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
>   at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImp
>   at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAcc
>   at java.lang.reflect.Constructor.newInstance(Constructor.java:494)
>   at org.apache.commons.logging.impl.LogFactoryImpl.newInstance(LogFactoryImpl.java:529
>   at org.apache.commons.logging.impl.LogFactoryImpl.getInstance(LogFactoryImpl.java:235
>   at org.apache.commons.logging.LogFactory.getLog(LogFactory.java:370)
>   at org.apache.hadoop.mapred.TaskTracker.<clinit>(TaskTracker.java:44)
>   at org.apache.hadoop.mapred.TaskTracker$Child.main(TaskTracker.java:993)
> We see several ways to solve this problem. First retrieve the properties "hadoop.log.dir" and "hadoop.log.file" from the mother JVM and then pass them to the child JVM as within the args parameter.
> Second would be to  access the environment variables "$HADOOP_LOG_DIR" and "$HADOOP_LOGFILE" using System.getEnv (java 1.5).
> Third there would be a more general solution. Taskrunner would resolve any environment variables it found in "mapred.child.java.opts" by lookup the value using System.getEnv().
> Eg:
> unix:
> export MAX_MEMORY = 200
> hadoop-site.xml:
> <name>mapred.child.java.opts</name>
> <value>-Xmx${MAX_MEMORY}</value>

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators: https://issues.apache.org/jira/secure/Administrators.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira