[
https://issues.apache.org/jira/browse/MAPREDUCE-6261?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14320744#comment-14320744
]
Jason Lowe commented on MAPREDUCE-6261:
---------------------------------------
The flush method can be invoked twice if we get an exception during flush/close
of the map output collector due to the following code in MapTask (this is for
new API but the same setup occurs for the old API as well):
{code}
try {
[...]
output.close(mapperContext);
output = null;
} finally {
closeQuietly(input);
closeQuietly(output, mapperContext);
}
{code}
The close method invokes the flush method, and if we throw during the close we
will try to close (and flush) again. If we throw trying to close *after* the
flush method has already set kvbuffer to null we will turn around and try to
close and flush again due to the finally clause and encounter the NPE.
Here's a sample stacktrace. Note that the exceptions are logged in the reverse
order of when they occurred due to the finally clause:
{noformat}
2015-02-12 07:42:01,482 INFO [main] org.apache.hadoop.mapred.MapTask: Ignoring
exception during close for
org.apache.hadoop.mapred.MapTask$NewOutputCollector@1234
java.lang.NullPointerException
at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.flush(MapTask.java:1447)
at
org.apache.hadoop.mapred.MapTask$NewOutputCollector.close(MapTask.java:700)
at org.apache.hadoop.mapred.MapTask.closeQuietly(MapTask.java:1990)
at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:774)
at org.apache.hadoop.mapred.MapTask.run(MapTask.java:340)
at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:168)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:415)
at
org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1637)
at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:163)
2015-02-12 07:42:01,484 WARN [main] org.apache.hadoop.mapred.YarnChild:
Exception running child :
org.apache.hadoop.util.DiskChecker$DiskErrorException: Could not find any valid
local directory for output/attempt_1416815736267_5558092_m_000000_0/file.out
at
org.apache.hadoop.fs.LocalDirAllocator$AllocatorPerContext.getLocalPathForWrite(LocalDirAllocator.java:402)
at
org.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:150)
at
org.apache.hadoop.fs.LocalDirAllocator.getLocalPathForWrite(LocalDirAllocator.java:131)
at
org.apache.hadoop.mapred.YarnOutputFiles.getOutputFileForWrite(YarnOutputFiles.java:84)
at
org.apache.hadoop.mapred.MapTask$MapOutputBuffer.mergeParts(MapTask.java:1813)
at org.apache.hadoop.mapred.MapTask$MapOutputBuffer.flush(MapTask.java:1484)
at
org.apache.hadoop.mapred.MapTask$NewOutputCollector.close(MapTask.java:700)
at org.apache.hadoop.mapred.MapTask.runNewMapper(MapTask.java:770)
at org.apache.hadoop.mapred.MapTask.run(MapTask.java:340)
at org.apache.hadoop.mapred.YarnChild$2.run(YarnChild.java:168)
at java.security.AccessController.doPrivileged(Native Method)
at javax.security.auth.Subject.doAs(Subject.java:415)
at
org.apache.hadoop.security.UserGroupInformation.doAs(UserGroupInformation.java:1637)
at org.apache.hadoop.mapred.YarnChild.main(YarnChild.java:163)
{noformat}
> NullPointerException if MapOutputBuffer.flush invoked twice
> -----------------------------------------------------------
>
> Key: MAPREDUCE-6261
> URL: https://issues.apache.org/jira/browse/MAPREDUCE-6261
> Project: Hadoop Map/Reduce
> Issue Type: Bug
> Components: mrv2
> Affects Versions: 2.5.0
> Reporter: Jason Lowe
>
> MapOutputBuffer.flush will throw an NPE if it is invoked twice, since it
> blindly assumes kvbuffer is not null yet sets kvbuffer to null towards the
> end of the method.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
