Re: OutOfMemoryError in the job tracker

[Philippe Gassmann]

Koji Noguchi a écrit :
Hi Thomas,

I think we had the same problem before.  Dhruba pointed out,

http://bugs.sun.com/bugdatabase/view_bug.do;:YfiG?bug_id=4705373


I doubt this is the cause of the problem since our jobtracker JVM is 
started with -Xmx1000m and our jars never exceed 25m...

This issue looks like a huge memory leak since everything is ok for a 
certain amount of time, and after a given time the OutOfMemory exception 
is thrown.

 > Did someone have a solution to avoid this ?
 >
Couple of options.
1) Ask the users to submit smaller jar file.
2) Use JDK 1.6
3) Use 64-bit JVM

Hope it works.

Koji


Thomas Friol wrote:
Hi,

I am using the 0.10.1 version of hadoop and after less than 24 hour of
running, the jobtracker send the following error when it tries to run a
job :

2007-02-01 04:10:09,285 ERROR org.apache.hadoop.mapred.JobTracker: Job
initialization failed:
java.lang.OutOfMemoryError
    at java.util.zip.ZipFile.open(Native Method)
    at java.util.zip.ZipFile.<init>(ZipFile.java:203)
    at java.util.jar.JarFile.<init>(JarFile.java:132)
    at java.util.jar.JarFile.<init>(JarFile.java:70)
    at sun.misc.URLClassPath$JarLoader.getJarFile(URLClassPath.java:579)
    at sun.misc.URLClassPath$JarLoader.<init>(URLClassPath.java:546)
    at sun.misc.URLClassPath$3.run(URLClassPath.java:324)
    at java.security.AccessController.doPrivileged(Native Method)
    at sun.misc.URLClassPath.getLoader(URLClassPath.java:313)
    at sun.misc.URLClassPath.getLoader(URLClassPath.java:290)
    at sun.misc.URLClassPath.getResource(URLClassPath.java:160)
    at java.net.URLClassLoader$1.run(URLClassLoader.java:192)
    at java.security.AccessController.doPrivileged(Native Method)
    at java.net.URLClassLoader.findClass(URLClassLoader.java:188)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:306)
    at java.lang.ClassLoader.loadClass(ClassLoader.java:251)
    at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:319)
    at java.lang.Class.forName0(Native Method)
    at java.lang.Class.forName(Class.java:242)
    at
org.apache.hadoop.conf.Configuration.getClassByName(Configuration.java:315) 

    at 
org.apache.hadoop.conf.Configuration.getClass(Configuration.java:326)
    at 
org.apache.hadoop.conf.Configuration.getClass(Configuration.java:339)
    at org.apache.hadoop.mapred.JobConf.getInputFormat(JobConf.java:263)
    at
org.apache.hadoop.mapred.JobInProgress.initTasks(JobInProgress.java:123)
    at
org.apache.hadoop.mapred.JobTracker$JobInitThread.run(JobTracker.java:360) 

    at java.lang.Thread.run(Thread.java:595)

Did soemone notice the same kind of behavior ?
Did someone have a solution to avoid this ?

I just found that restarting the jobtracker solves the problem but
that's it.

Thanks for any help.